Leveraging native antibody properties to guide bispecific selection

Abstract

Multicomponent polypeptides such as bispecific antibodies comprise multiple unique polypeptide chains, and ensuring their proper quaternary assembly is critical for both efficacy and safety. The present invention relates to purification-based Building Block Compatibility Screening (pBBCS), a new method to select components for the successful integration into multicomponent polypeptides. pBBCS utilizes analysis of physical properties such as the surface charge of those components, and facilitates quickly focusing optimization efforts on only promising candidate components. Therefore, pBBCS can play a critical role in the next-generation workflows for multicomponent polypeptides to enable lower costs, higher speed, and an increase in overall success rates.

Description

LEVERAGING NATIVE ANTIBODY PROPERTIES TO GUIDE BISPECIFIC SELECTIONTECHNICAL FIELD

[0001] The disclosure relates to the generation and purification of multicomponent polypeptides, such as antibodies and particularly bispecific antibodies. In some aspects, the disclosure relates to the use of methods utilizing physical properties of candidate components such as surface charge to select the appropriate components for integration into multicomponent polypeptides.BACKGROUND

[0002] Bispecific antibodies are an exciting generation of biotherapeutics, enabling simultaneous or sequential targeting of two or more unique epitopes located on the same, or distinct, targets. This dual-recognition capability enables diverse applications such as recruiting immune cells to kill tumor cells, crosslinking distinct cell surface receptors, or improving tissue specificity. (Labrijn AF et al., Nat. Rev. Drug Discov. 18:585-608 (2019); Lu RM et al., J. Biomed. Sci. 27: 1 (2020); Fan G et al., J. Hematol. Oncol. 8: 130 (2015)) For example, Amgen’s Bispecific T-cell Engager (BiTE®) creates an artificial immune synapse between cytotoxic T cells and target tumor cells, by simultaneously binding a CD3 epitope on the surface of T cells and a tumor-associated antigen. (Wolf et al., Drug Discov. Today 10:1237-44 (2005);Kantarjian H et al., N. Engl. J. Med. 376:836-47 (2017)) As of 2019, over 100 bispecific formats have been reported, with over 85 in development and three receiving US Food and Drug Administration approval. (Labrijn AF et al., Nat. Rev. Drug Discov. 18:585-608 (2019);Brinkmann U and Kontermann RE, MAbs 9: 182-212 (2017); Wang Q et al., Antibodies (Basel) 8(3):43 (2019))

[0003] There are two major design strategies to generate bispecific molecules. The first approach is to encode two or more unique fragment variable (Fv) sequences on the same polypeptide chain(s), as with formats such as BiTEs®, IgG-scFv, or DVD-Ig. (Wang Q et al., Antibodies (Basel) 8(3):43 (2019); Spiess C et al.. Mol. Immunol. 67:95-106 (2015)) These single-chain formats bypass the challenges associated with assembling multiple polypeptide chains into a single molecule, but they also tend to show poor yields and suboptimal stabi 1 i ty.Alternative strategies utilize biophysics and engineering to "steer’ individual chains to the correct orientation, while simultaneously disfavoring mispaired scenarios. Examples include technologies like Knob-into-Holes (KiH), Charge Pair Mutations (CPMs) and Strand-Exchange Engineered Domains (SEEDbody) and have been utilized in molecules requiring hetero-Fc pairing. (Davis JH et al. , Protein Eng. Des. Sei. 23: 195-202 (2010); Dillon M et al., MAbs 9:213- 30 (2017); Gunasekaran K et al. , J. Biol. Chem. 285: 19637-46 (2010); Ridgway JB et al., Protein Eng. 9:617-21 (1996)). This second approach enables the production of molecules that more closely mimic the native structures of IgG molecules, allowing for increased stability and more diverse format design. (Wang Q et al.. Antibodies (Basel) 8(3):43 (2019)) However, these engineering strategies are often imperfect, and chain-mispairing continues to pose a challenge to hetero-Fc bispecific production. (Ha JH et al., Front. Immunol. 7:394 (2016)) To remove these undesired species, multiple purification steps are typically required.

[0004] The properties of parental antibodies can greatly influence the Ab chain pairing required for assembly of Hetero-IgGs and related bispecific formats, by enhancing or even replacing the requirement for heavy / light chain-pairing engineering. (Gong D et al. , MAbs 13: 1870058 (2021)) In therapeutic development, antibodies are often empirically combined into bispecifics, resulting in hundreds or thousands of potential combinations that translating to extended timelines and increased resource needs to identify the best therapeutic candidates. Chain Selectivity Assessment (CSA) is a high-throughput method to rationally select parental monoclonal antibodies (mAbs) to make bispecific antibodies requiring correct heavy / light chain pairing. However, although CSA interrogates the HC / LC pairing phenomenon, it does not directly address the universal purification challenges presented by bispecific formats. Indeed, a bispecific candidate that cannot be separated from mispaired impurities will be unsuccessful as a therapeutic candidate. Because these impurities are derived from the same building blocks as the target molecule, their physicochemical properties are often too similar to be effectively separated chromatographically. Accordingly, there is a need in the art for improved methods for identifying the building blocks for bispecific antibodies that will best translate into therapeutic candidates. More generally, there is also a need for improved methods for identifying the building blocks for multicomponent polypeptides in general.SUMMARY

[0005] This application relates to processes for screening and selecting potential components for inclusion in multicomponent polypeptides, and methods of constructing multicomponent peptides that incorporate the selected components. In some embodiments, those processes and methods involve measurement of a physical property such as conductivity. For example, the process of measuring the conductivity of a panel of potential components can be used as part of a method of generating multicomponent polypeptides. The disclosed methods allow prediction of how well a panel of desired multicomponent species might be separated from undesired contaminating species, and prediction of which candidate components would best facilitate such separation. The processes described herein into screening and production processes facilitates more efficient screening, and results in identifying multicomponent peptides that make better therapeutic candidates because they are able to more easily be purified away from contaminants.

[0006] Some embodiments described herein relate to a method of screening potential components for multicomponent polypeptides, wherein the multicomponent polypeptides each comprise at least one building block A and at least one building block B. and wherein the screening comprises multiple steps to identify preferred candidate components. Those steps include identifying and individually purify ing a group of molecules (pool A) wherein each molecule comprises one or more copies of a candidate building block A, as well as identifying and individually purifying a group of molecules (pool B) wherein each molecule comprises a one or more copies of a candidate building block B. Those steps also include determining the conductivity of each pool A molecule and of each pool B molecule. After determining the conductivity of the pool A and pool B molecules, the difference in conductivity (ACond or AConductivity) between each molecule of pool A and each molecule of pool B is determined. The steps also include selecting a set of candidate multicomponent polypeptides that comprise a building block A-building block B pair, wherein the pool A and pool B molecules comprising the building blocks in the selected multicomponent peptides have a larger ACond than the ACond between the pool A and pool B molecules comprising the building block pairs in the non-selected multicomponent polypeptides.

[0007] In some embodiments, the conductivity of each pool A molecule and of each pool B molecule is determined using cation-exchange chromatography (CIEX), anion-exchangechromatography (AIEX), hydrophobic interaction chromatography, or reversed phase chromatography.

[0008] In some embodiments, the selected multicomponent peptides have a ACond between the pool A and pool B molecules greater than 5 mS / cm. In some embodiments, the selected candidate multicomponent polypeptides have a ACond between the pool A and pool B molecules greater than 7 mS / cm. In some embodiments, the selected building block A came from a pool A molecule measured to have a conductivity within the top 25% of tested pool A molecules, and the selected building block B came from a pool B molecule measured to have a conductivity within the bottom 25% of tested pool B molecules. In some embodiments, the selected building block A came from a pool A molecule measured to have a conductivity' within the top 10% of tested pool A molecules, and the selected building block B came from a pool B molecule measured to have a conductivity within the bottom 10% of tested pool B molecules.

[0009] Particular embodiments relate to methods of screening in which each pool A molecule, each pool B molecule, or each pool A molecule and each pool B molecule comprises an immunoglobulin. In some of these embodiments, each pool A molecule, each pool B molecule, or each pool A molecule and each pool B molecule comprises an antibody. In some embodiments, building block A comprises all or part of an antibody heavy chain. In some embodiments, building block A comprises an antibody arm. In some embodiments, building block B comprises all or part of an antibody heavy chain. In some embodiments, building block B is an antibody arm. In some embodiments, building block A and building block B comprise all or part of an antibody heavy chain. In some embodiments, building block A and building block B comprise an antibody arm.

[0010] The multicomponent polypeptide of the method of screening may comprise a variety of different classes of polypeptides. In some embodiments, the multicomponent polypeptide is a bispecific antibody. In some embodiments, building block A comprises an antibody arm and building block B comprises an Fc-linked scFv. In some embodiments, the multicomponent polypeptide is an antibody drug conjugate or antibody fusion. In some embodiments, the multicomponent polypeptide comprises an antibody fusion, and wherein the antibody is fused to a cytokine.

[0011] The methods disclosed herein may combine analysis of ACond with further steps as part of the screening or design process. In some embodiments, the method further comprises estimating the absolute conductivity of the selected multicomponent polypeptides by summingthe conductivity of the pool A molecule and pool B molecule that supply the building blocks for each multicomponent polypeptide, and keeping in the selected set only those multicomponent polypeptides within the upper 50% of the estimated summed conductivities. In some embodiments, the method comprises keeping only those multicomponent polypeptides with an absolute conductivity greater than 11 mS / cm. In some embodiments, the absolute conductivity is measured by summing the conductivity of each molecule as determined using cation-exchange chromatography (CIEX), anion-exchange chromatography (AIEX), hydrophobic interaction chromatography, or reversed phase chromatography.

[0012] In some embodiments, the method further comprises comparing the CIEX elution profile of each selected multicomponent polypeptide with the CIEX elution profiles of the pool A molecule and the pool B molecule from which its component parts were derived, and keeping only those multicomponent polypeptides whose principal elution peak have less than 25% overlap in peak area with the principal elution peak of either the pool A molecule or the pool B molecule. In some embodiments, there is less than 10% overlap in peak area between the multicomponent peptide's principal peak and the principal peak of either the pool A molecule or the pool B molecule.

[0013] Some embodiments described herein relate to a method of constructing a multicomponent immunoglobulin molecule comprising the steps of (1) measuring the surface charge of each member of a pool A of immunoglobulins and measuring the surface charge of each member of a pool B of immunoglobulins; (2) selecting an immunoglobulin from pool A (immunoglobulin A) and an immunoglobulin from pool B (immunoglobulin B) that have a difference in conductivity (ACond) greater than 5 mS / cm. as measured by CIEX; and (3) combining a building block A that comprises part or all of immunoglobulin A with a building block B that comprises part or all of immunoglobulin B in the multicomponent immunoglobulin molecule. In some embodiments, the selected immunoglobulin A and immunoglobulin B have a ACond greater than 7 mS / cm, as measured by CIEX.

[0014] In some embodiments of the method of constructing a multicomponent polypeptide, each pool A molecule and each pool B molecule comprises an immunoglobulin. In some of these embodiments, each pool A molecule, each pool B molecule, or each pool A molecule and each pool B molecule comprises an antibody. In some embodiments, building block A comprises all or part of an antibody heavy chain. In some embodiments, building block A comprises an antibody arm. In some embodiments, building block B comprises all or part of anantibody heavy chain. In some embodiments, building block B is an antibody arm. In some embodiments, building block A and building block B comprise all or part of an antibody heavy chain. In some embodiments, building block A and building block B comprise an antibody arm.

[0015] The multicomponent polypeptide of the method of constructing a multicomponent polypeptide may comprise a variety of different classes of polypeptides. In some embodiments, the multicomponent polypeptide is a bispecific antibody. In some embodiments, building block A comprises an antibody arm and building block B comprises an Fc-linked scFv. In some embodiments, the multicomponent polypeptide is an antibody drug conjugate or antibody fusion. In some embodiments, the multicomponent polypeptide comprises an antibody fusion, and wherein the antibody is fused to a cytokine.

[0016] The methods of constructing a multicomponent polypeptide disclosed herein may combine analysis of ACond with further steps as part of the screening or design process. In some embodiments, the method further comprises estimating the absolute conductivity of the selected multicomponent polypeptides by summing the conductivity of the pool A molecule and pool B molecule that supply the building blocks for each multicomponent polypeptide, and keeping in the selected set only those multicomponent polypeptides within the upper 50% of the estimated summed conductivities. In some embodiments, the method comprises keeping only those multicomponent polypeptides with an absolute conductivity greater than 11 mS / cm. In some embodiments, the absolute conductivity is measured by summing the conductivity of each molecule as determined using cation-exchange chromatography (CIEX), anion-exchange chromatography (AIEX), hydrophobic interaction chromatography, or reversed phase chromatography.BRIEF DESCRIPTION OF THE FIGURES

[0017] FIGS. 1 A- IF are diagrams that show different antibodies and antibody-based structures. FIG. 1 A is an example of a basic monospecific bivalent antibody, which contains two identical arms capable of binding antigen A (Building Block A) attached through their respective Fc regions. FIG. IB shows an exemplary' bispecific antibody in which an arm capable of binding antigen A is linked to an arm capable of binding to antigen B (Building Block B) through their respective Fc regions, where the interaction's specificity is facilitated by complementary chargepair mutations in those Fc regions. FIG. 1C shows an exemplary AmAb, which is comprised of a standard antibody arm capable of binding antigen A (Building Block A) and an antibody armcapable of binding antigen B that has an scFv in place of the Fab region (Building Block B'), where the scFv optionally contains a stabilizing disulfide bond. FIG. ID is a schematic breakdown of a bispecific antibody (as shown in FIG. IB) that shows the percent identity with each of the parent antibodies from which it was developed and the percentage of the bispecific antibody that is engineered to facilitate the bispecific antibody’s generation. FIG. IE is a schematic breakdown of an AmAb (as shown in FIG. 1C) that shows the percent identity' with each of the parent antibodies from which it w as developed and the percentage of the AmAb that is engineered to facilitate the bispecific antibody’s generation. FIG. IF is a diagram showing various species that may be formed during the creation of an AmAb from its constituent protein chains, including the desired AmAb (the “correct species”) and a range of different potential contaminants such as mispaired homodimers, single-arm species, and aggregates.

[0018] FIGS. 2A-2E show an overview of cation exchange chromatography (CIEX) purification profiles of different exemplary antibodies and antibody-based molecules. FIG. 2A is an overlay of the CIEX elution profiles for a parental antibody capable of binding antigen A (Building Block A homodimer), a parental antibody capable of binding antigen B (Building Block B homodimer), and a hetero-IgG bispecific antibody generated from those parental antibodies that is capable of binding both antigen A and antigen B. FIG. 2B is an overlay of the CIEX elution profiles for a parental antibody capable of binding antigen A (Building Block A homodimer), a parental antibody capable of binding antigen B (Building Block B homodimer), and an AmAb generated from those parental antibodies that is capable of binding both antigen A and antigen B. FIG. 2C is an overlay of the CIEX elution profiles for a parental antibody capable of binding antigen A (Building Block A homodimer), a parental antibody that is an scFv-based homodimer capable of binding antigen B (Building Block B' homodimer), an scFv-linked antibody with the ClmAb format, and an scFv-linked antibody with the C2mAb format. FIG. 2D is an overlay of the CIEX elution profiles for a parental antibody capable of binding antigen A (Building Block A homodimer), a parental antibody capable of binding antigen B (Building Block B homodimer), an scFv-linked antibody with the C lmAb format, an scFv-linked antibody with the BlmAb format, and an scFv-linked antibody with the NlmAb format. FIG. 2E is an overlay of the CIEX elution profiles for a parental antibody capable of binding antigen A (Building Block A homodimer), a cytokine, and an antibody-cytokine fusion comprising an antibody capable of binding antigen A and a cytokine receptor.

[0019] FIG. 3 is a schematic overview of purification-based Building Block Compatibility Screening (pBBCS)-aided CIEX separation. The separation between parental mAbs (AConductivity) indicates the extent to which common contaminants such as homodimeric and single-arm species will separate from the AmAb species during CIEX.

[0020] FIGS. 4A and 4B show an overlay of the CIEX profiles for 8 unique mAbs against Target A (FIG. 4A) and an overlay of the CIEX profiles for 17 unique mAbs against Target B (FIG 4B).

[0021] FIGS. 5A-5C show representative CIEX profiles for AmAb designs, overlaid with the CIEX elution profiles of their respective pairs of parental antibodies with various AConductivities. FIG. 5A shows the overlay of an AmAb’s elution profile along with its parent antibodies that have a AConductivity of 3.6 mS / cm. FIG. 5B shows the overlay of an AmAb’s elution profile along with its parent antibodies that have a AConductivity of 6.3 mS / cm. FIG. 5C shows the overlay of an AmAb’s elution profile along with its parent antibodies that have a AConductivity of 10.5 mS / cm.

[0022] FIG. 6 is a two-dimensional visualization of 136 unique combinations of Building Blocks A and B (as shown in FIGS. 4A and 4B), plotted by mean protein A yield (average of Target A and Target B mAb yields) and AConductivity.

[0023] FIG. 7 is a correlational analysis of predicted AmAb elution position (calculated as the mean of the two parental mAb elution positions) and observed AmAb elution position.

[0024] FIGS. 8A and 8B show correlational analyses of AConductivity7against the separation of AmAb from either superimposed Building Block A mAb (FIG. 8A) or superimposed Building Block B mAb (FIG. 8B) in CIEX. Data points are shaded to reflect qualitative separation behavior as either poor (grey squares) or good (black circles). The vertical dotted line represents a threshold AConductivity7of 7 mS / cm, above which good separation was consistently observed.

[0025] FIG. 8C is a breakdown of panel size after each step of pBBCS using the panels shown in FIGS 4A and 4B as potential combinations. FIG 8C applies a AConductivity threshold of 7 mS / cm or greater, which was identified in FIGS. 8A and 8B as a threshold above-which good separation w ould be expected for resultant bispecific antibodies.

[0026] FIGS. 9A-9G show the characterization of a self-associating AmAb subspecies observed in CIEX. FIG. 9A shows the CIEX elution profile of an AmAb antibody, with the two principal noted peaks highlighted. FIGS. 9B and 9C are the SEC elution profiles of collectedprotein from Peak 1 and Peak 2 from FIG. 9A, respectively. As shown in FIGS. 9B and 9C, Peak 2 exhibited an additional and substantially earlier SEC elution position than Peak 1. FIGS. 9D and 9E show intact mass spectrometry of collected protein from Peak 1 and Peak 2 from FIG.9 A, respectively. That intact mass spectrometry revealed that Peak 2 was composed of AmAb. with a mass identical to that of Peak 1, suggesting that Peak 2 was a self-associating multimer of AmAbs. FIGS. 9F and 9G are CIEX re-runs of collected protein from Peak 1 and Peak 2, respectively. The appearance of a peak identical in size to Peak 1 in Peak 2’s re-run provides further indication that Peak 2 was a self-associating multimer of AmAbs.

[0027] FIGS 10A-10D show the characterization of a self-associating AmAb subspecies observed in CIEX. FIG. 10A shows the SEC elution profile of the protein A eluate obtained when purifying the AmAb. FIG. 10B shows the elution times of Peak 1 and Peak 2 plotted with the molecular weights of monomers, dimers, or trimers of the AmAb. overlaid on top of an SEC standard curve. FIGS. 10C and 10D show intact mass spectrometry of collected protein from Peak 2 of AmAbs that differ only in the absence or presence of a disulfide in the scFv region. That intact mass spectrometry revealed that the disulfide bond facilitated maintenance of a dimer molecule in Peak 2 during mass spectrometry.

[0028] FIGS. 11 A-l 1C are an analysis of the affinity of the material in Peak 1 and Peak 2 from FIG. 10A for Target A and Target B. FIGS. 11 A and 1 IB show in vitro analysis of the affinity of the material in Peak 1 and Peak 2. FIG. 11C shows a cell-based functional assay to measure on and off-target activity.

[0029] FIGS. 12A and 12B show an analysis of the separation of monomer and multimer antibodies in relation to the conductivity of the monomer antibody. FIG. 12A shows an exemplary CIEX elution profile for an AmAb, which notes the increased contact points with CIEX resin, potential intermol ecular basic patches, and increased elution conductivity of a multimeric species. FIG12B shows a plot of the CIEX conductivity of a various AmAb monomers against the separation of Peak 1 (monomer) and Peak 2 (multimer) in CIEX elution profiles, with qualitatively poor separation noted when the conductivity of the monomer AmAb was below 11 mS / cm. FIG12C shows a plot of the CIEX conductivity for Target B mAbs against separation of Peak 1 (monomer) and Peak 2 (multimer) in CIEX elution profiles, possibly due to a dominant role played by scFvs in defining how multimeric AmAbs interact with the CIEX column.

[0030] FIGS. 13A and 13B show analytical SEC of the Protein A eluates for two AmAbs, containing identical Fab arm sequences but different scFv arm sequences. Pairing the Target Al building block with the Target Bl building block produced the AlxBl AmAb, while pairing the Target Al building block with the Target B2 building block produced the AlxB2 AmAb. Both the AlxBl AmAb and the A1 / B2 AmAb show the presence of aggregation in their respective SEC elution profdes in FIG. 13A and FIG. 13B.

[0031] FIGS. 14A and 14B show an example in which two AmAbs, containing identical Fab arm sequences but different scFv arm sequences, generated aggregates with differing degrees of separation after CIEX. Pairing the Target Al building block with the Target Bl building block (AlxBl) produced an AmAb with an elution conductivity of 5.6 mS / cm (FIG. 14A), whereas pairing the same Target Al building block with the Target B2 building block (AlxB2) generated an AmAb with an elution conductivity of 11.8 mS / cm (FIG. 14B). Elution conductivities for Bl and B2 building blocks (expressed as mAbs) were 7.3 and 16.9 mS / cm, respectively. The aggregate from the AlxBl AmAb fails to separate from the monomer form during SEC, as evidenced by the HMW species detected in the same SEC fractions (FIG. 14A). In contrast, the aggregate from the AlxB2 AmAb separated well from the monomer form during SEC. as evidenced by the HMW species being detected in separate SEC fractions (FIG. 14B)

[0032] FIGS. 15A-15F show an analysis of aggregation in a panel of AmAbs. FIGS 15A- 15D show the SEC and CIEX elution profiles of four AmAbs, in which the proportion of aggregate peak in SEC correlated with the abundance of aggregate peak in CIEX. FIG. 15E shows a plot of the average aggregate abundance with ten different Target B building blocks, with some scFv sequences having a higher propensity toward aggregation. FIG. 15F shows a plot of the size of Peak 2 observed with CIEX against the percent aggregation measured in SEC, with a trendline for relation included.

[0033] FIG. 16 shows an overview' of pBBCS-based strategies for improving AmAb purification. By increasing AConductivity, pBBCS improves the resolution between AmAbs and contaminants such as homodimers and single-arm species. Furthermore, pBBCS aims to increase the overall conductivity of the AmAb molecule (which can be accurately predicted from the parental mAb elution profiles) to enable effective separation of aggregates due to the more pronounced avidity effects of high-conductivity AmAb molecules.DETAILED DESCRIPTION

[0034] Bispecific antibodies, engineered to recognize two targets simultaneously, demonstrate exceptional clinical potential for the therapeutic intervention of complex diseases. These molecules are often composed of multiple polypeptide chains of differing sequences. To meet industrial scale productivity, enforcing the correct quaternary7assembly of these chains is critical. More broadly, the same demands on assembly are present in production of multicomponent polypeptides in general. The purification of multicomponent polypeptides resulting from a single cell expression is widely attractive from a manufacturing standpoint enabling lower costs and improve timelines. However, such approach may render cell extracts with undesirable species that must be separated from the molecule of interest. The methods disclosed herein provide rational workflows for predicting success, based on investigating and identifying the native properties of component building blocks (e.g., antibodies) that can influence the purification of multicomponent polypeptides or their component building blocks.

[0035] This application describes Purification-Based Building Block Compatibility Screening (pBBCS), a method to predict the assembly of multicomponent polypeptides (e.g., bispecific antibodies) with the appropriate building blocks to enable superior purification properties. pBBCS utilizes analysis of physical properties such as the surface charge of those components, and facilitates quickly focusing optimization efforts on only promising candidate components. Therefore, pBBCS can play a critical role in the next-generation workflows for multicomponent polypeptides to enable lower costs, higher speed, and an increase in overall success rates.

[0036] One property7that maybe applied in the context of pBBCS is elution conductivity of proteins. Each multicomponent protein exhibits a unique surface charge that which impacts and defines the behavior of the that protein, just as the surface charge of contaminating unpaired or mispaired species impacts and defines their behavior. These surface charge properties are important during certain types of chromatography7(e.g., cation exchange chromatography (CEX)), a critical step in multicomponent protein purification. In some embodiments, the methods described herein use in silico prediction of bispecific CEX profiles to define compatibility requirements for component proteins. As an exemplary application of these methods, the Examples below demonstrate the predictive power of pBBCS using a panel of Fab- scFv-Fc molecules by showing that distinct surface charge properties (as evidenced by differences in elution conductivities) of the two Fvs can improve separation of mispaired species.The Examples also show that aggregates, a major contaminant in scFv-containing bispecifics, can be effectively separated through selection of Fv pairs predicted to generate a sufficiently high surface charge (i.e., a high CEX elution conductivity).Definitions of general terms and expressions

[0037] In order that the present disclosure can be more readily understood, certain terms are first defined. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure is related. As used in this application, except as otherwise expressly provided herein, each of the following terms shall have the meaning set forth below. Additional definitions are set forth throughout the application.

[0038] Units, prefixes, and symbols are denoted in their Systeme International de Unites (SI) accepted form.

[0039] As used in the present disclosure and claims, the singular forms “a,” "an.'’ and "the ” include plural forms unless the context clearly dictates otherwise. Unless specifically stated or obvious from context, as used herein, the term “or” is understood to be inclusive. The term “and / or” as used in a phrase such as “A and / or B” herein is intended to include both “A and B,” “A or B,” “A,” and “B ” Likewise, the term “and / or” as used in a phrase such as “A, B, and / or C” is intended to encompass each of the following embodiments: A, B, and C; A, B. or C; A or C; A or B; B or C; A and C; A and B; B and C; A (alone); B (alone); and C (alone).

[0040] It is understood that wherever embodiments are described herein with the language “comprising,” otherwise analogous embodiments described in terms of “consisting of” and / or “consisting essentially of’ are also provided. In this disclosure, “comprises,” “comprising.” “containing” and “having” and the like can have the meaning ascribed to them in U.S. Patent law and can mean “includes,” “including,” and the like; “consisting essentially of’ or “consists essentially” likewise have the meaning ascribed in U.S. Patent law and the terms are open-ended, allowing for the presence of more than that which is recited so long as basic or novel characteristics of that which is recited is not changed by the presence of more than that which is recited, but excludes prior art embodiments.

[0041] The terms “about” or “comprising essentially of’ refer to a value or composition that is within an acceptable error range for the value or composition as determined by one of ordinary skill in the art, which will depend in part on how the value or composition is measured or determined, i.e., the limitations of the measurement system. For example, “about” or“comprising essentially of’ can mean within 1 or more than 1 standard deviation per the practice in the art. Alternatively, “about” or “comprising essentially of’ can mean a range of up to 20%. Furthermore, particularly with respect to biological systems or processes, the terms can mean up to an order of magnitude or up to 5-fold of a value. When values or compositions are provided in the application and claims, unless otherwise stated, the meaning of “about” or “comprising essentially of’ should be assumed to be within an acceptable error range for that value or composition.

[0042] The term "antibody" means an immunoglobulin molecule that recognizes and specifically binds to a target, such as a protein, polypeptide, peptide, carbohydrate. polynucleotide, lipid, or combinations of the foregoing. As used herein, the term "antibody" encompasses polyclonal antibodies, monoclonal antibodies, chimeric antibodies, humanized antibodies, fully human antibodies, recombinant antibodies, multispecific antibodies, and bispecific antibodies. An antibody can be of any the five major classes of immunoglobulins: IgA, IgD, IgE, IgG, and IgM, or subclasses (isotypes) thereof (e.g., IgGl, IgG2, IgG3, IgG4, IgAl and IgA2), based on the identity of their heavy-chain constant domains referred to as alpha, delta, epsilon, gamma, and mu, respectively. The different classes of immunoglobulins have different and well-known subunit structures and three-dimensional configurations. For example, a common configuration for an antibody has two full length antibody heavy chains and two full length antibody light chains.

[0043] As used herein, the term “antibody heavy chain” refers to an antibody heavy chain, consisting of a variable region and a constant region as defined for a full-length antibody. A full-length antibody heavy chain is a polypeptide consisting in N-terminal to C- terminal direction of an antibody heavy chain variable domain (VH), an antibody constant heavy chain domain 1 (CHI), an antibody hinge region (HR), an antibody heavy chain constant domain 2 (CH2), and an antibody heavy chain constant domain 3 (CH3), abbreviated as VH-CH1-HR- CH2-CH3. The term heavy chain when used in reference to an antibody can refer to any distinct type, e.g., alpha (a), delta (5), epsilon (s), gamma (y), and mu (p), based on the amino acid sequence of the constant domain, which give rise to IgA, IgD, IgE, IgG, and IgM classes of antibodies, respectively, including subclasses of IgG (e.g., IgGl, IgG2, IgG3, and IgG4) and subclasses of IgA (e.g., IgAl and IgA2). Heavy chain amino acid sequences are known in the art.

[0044] As used herein, the term “antibody light chain” refers to an antibody light chain, consisting of a variable region and a constant region as defined for a full-length antibody. A full-length antibody light chain is a polypeptide consisting in N-terminal to C- terminal direction of an antibody light chain variable domain (VL), and an antibody light chain constant domain (CL), abbreviated as VL-CL. The term light chain when used in reference to an antibody can refer to any distinct type, e.g., kappa (K) or lambda (X) based on the amino acid sequence of the constant domains. Light chain amino acid sequences are known in the art.

[0045] The term "antibody fragment" refers to a portion of an intact antibody. An "antigen-binding fragment," "antigen-binding domain," or "antigen-binding region," refers to a portion of an intact antibody that binds to an antigen. An antigen-binding fragment can contain the antigenic determining regions of an intact antibody (e.g., the complementarity determining regions (CDR)). Examples of antigen-binding fragments of antibodies include, but are not limited to Fab, Fab', F(ab')2, and Fv fragments, linear antibodies, and single chain antibodies. An antigen-binding fragment of an antibody can be derived from any animal species, such as rodents (e.g., mouse, rat. or hamster) or humans, or can be artificially produced.

[0046] The term “multispecific antibody” means that an antigen binding protein is capable of specifically binding to two or more different antigens. A subcategory of multispecific antibodies is "bispecific antibodies," which are capable of specifically binding to two different antigens. As used herein, an antibody "‘specifically binds” to a target antigen when it has a significantly higher binding affinity for, and consequently is capable of distinguishing, that antigen, compared to its affinity for other unrelated proteins, under similar binding assay conditions. Antigen binding proteins that specifically bind an antigen may have an equilibrium dissociation constant (KD) < 1 X 1 O'6M. The antigen binding protein specifically binds antigen with “high affinity” when the KD is < 1 x 10‘8M.

[0047] An "isolated antibody" refers to an antibody population that comprises a single species of antibody. For example, a particular isolated antibody consists of an antibody population having a single heavy chain amino acid sequence and a single light chain amino acid sequence, which binds to a single epitope. An isolated antibody can, however, have crossreactivity to other antigens, such as related molecules from different species. Also, a population of antibodies may still be an "isolated antibody" when contaminated by small amounts of other antibody species. In particular, an isolated antibody may contain less than 5%, less than 4%, less than 3%, less than 2%, less than 1%, or no other antibody species.

[0048] A "monoclonal antibody" refers to a homogeneous antibody population involved in the highly specific recognition and binding of a single antigenic determinant, or epitope. Thisis in contrast to polyclonal antibodies that typically include different antibodies directed against different antigenic determinants. Furthermore, "monoclonal" antibody refers to such antibodies made in any number of manners including but not limited to by hybridoma, phage selection, recombinant expression, and transgenic animals.

[0049] The terms "variable region" or "variable domain" are used interchangeably and are common in the art. The variable region typically refers to a portion of an antibody, generally, a portion of a light or heavy chain, typically about the amino-terminal 110 to 120 amino acids or 110 to 125 amino acids in the mature heavy chain and about 90 to 115 amino acids in the mature light chain, which differ extensively in sequence among antibodies and are used in the binding and specificity of a particular antibody for its particular antigen. The variability' in sequence is concentrated in those regions called complementarity determining regions (CDRs) while the more highly conserved regions in the variable domain are called framework regions (FR). From N-terminus to C-terminus, naturally occurring light and heavy chain variable regions both typically conform with the following order of these elements: FR1, CDR1, FR2, CDR2, FR3, CDR3 and FR4. Without wishing to be bound by any particular mechanism or theory, it is believed that the CDRs of the light and heavy chains are primarily responsible for the interaction and specificity of the antibody with antigen.

[0050] The terms "VL" and "VL domain" and "VH region" are used interchangeably to refer to the light chain variable region of an antibody.

[0051] The terms "VH" and "VH domain" and "VH region" are used interchangeably to refer to the heavy chain variable region of an antibody.

[0052] The terms "constant region" and "constant domain" are interchangeable and have their common meaning in the art. The constant region is an antibody portion, e g., a carboxyl terminal portion of a light and / or heavy chain which is not directly involved in binding of an antibody to antigen, but which can exhibit various effector functions, such as interaction with the Fc receptor. The constant region of an immunoglobulin molecule generally has a more conserved amino acid sequence relative to an immunoglobulin variable domain.

[0053] As used herein, the terms "Fc region” and “Fc domain" refer to a C-terminal region of an IgG heavy chain; in case of an IgGl antibody, the C-terminal region comprises - CH2-CH3 (see above).

[0054] As used herein, the term "antibody arm" refers to the combination of an antibody heavy chain and an antibody light chain. In some embodiments, an antibody comprises twoantibody arms joined together by their respective heavy chain Fc regions. In some embodiments relating to bispecific antibodies, the bispecific antibody comprises two distinct antibody arms joined together by their respective heavy chain Fc regions.

[0055] The term "interface" refers to the association surface that results from interaction one or more amino acids in a first antibody domain with one or more amino acids of a second antibody domain. Exemplary interfaces include the CHI / CL, VH / VL, CH2-CH2, and CH3 / CH3 interfaces. In some embodiments, the interface includes, for example, hydrogen bonds, electrostatic interactions, or salt bridges between the amino acids forming an interface.

[0056] The term "chimeric antibody" refers to an antibody wherein the amino acid sequence is derived from two or more species. Typically, the variable region of both light and heavy chains corresponds to the variable region of antibodies derived from one species of mammals (e.g., mouse, rat, rabbit, etc.) with the desired specificity, affinity, and capability while the constant regions are homologous to the sequences in derived from another (usually human) to avoid eliciting an immune response in that species.

[0057] A "humanized antibody" refers to a chimeric antibody comprising amino acid residues from non-human CDRs and amino acid residues from human framework regions and constant regions. A humanized antibody may comprise substantially all of at least one. and typically two, variable domains, in which all or substantially all of the CDRs correspond to those of a non-human antibody, and all or substantially all of the FRs correspond to those of a human antibody. A humanized antibody optionally may comprise at least a portion of an antibody constant region derived from a human antibody. A "humanized form" of an antibody, e.g., a non- human antibody, refers to an antibody that has undergone humanization. Typically, humanized antibodies are human immunoglobulins in which residues from the CDRs are replaced by residues from the CDRs of a non-human species (e.g., mouse, rat, rabbit, hamster) that have the desired specificity, affinity, and capability. Accordingly, humanized antibodies are also referred to as "CDR grafted" antibodies. Early examples of methods used to generate humanized antibodies are described in U.S. Pat. 5,225,539; Roguska et al.. Proc. Natl. Acad. Sci., USA, 91(3):969-973 (1994), and Roguska et al., Protein Eng. 9(10):895-904 (1996). Many additional examples and methods relating to humanization of antibodies have subsequently been published.

[0058] A "human antibody" refers to an antibody having variable regions in which both the FRs and CDRs are derived from human germline immunoglobulin sequences. Furthermore, if the antibody contains a constant region, the constant region also is derived from human germlineimmunoglobulin sequences. The human antibodies of the disclosure can include amino acid residues not encoded by human germline immunoglobulin sequences (e.g., mutations introduced by random or site-specific mutagenesis in vitro or by somatic mutation in vivo). However, the term "human antibody," as used herein, is not intended to include antibodies in which CDR sequences derived from the germline of another mammalian species, such as a mouse, have been grafted onto human framework sequences. The terms "human antibodies" and "fully human antibodies" and are used synonymously.

[0059] The terms "polypeptide," "peptide," and "protein" are used interchangeably herein to refer to polymers of amino acids of any length. The polymer can be linear or branched, it can comprise modified amino acids, and it can be interrupted by non-amino acids. The terms also encompass an amino acid polymer that has been modified naturally or by intervention; for example, disulfide bond formation, glycosylation, lipidation, acetylation, phosphorylation, or any other manipulation or modification, such as conjugation with a labelling component. Also included within the definition are, for example, polypeptides containing one or more analogs of an amino acid (including, for example, unnatural amino acids, etc.), as well as other modifications know n in the art. It is understood that, because the polypeptides of this invention are based upon antibodies, in certain embodiments, the polypeptides can occur as single chains or associated chains.

[0060] The term antibody drug conjugate ("ADC") is used herein to refer to an antibody or antibody fragment that has been conjugated, or linked, to another small molecule, protein, or compound (the ADC's "cargo"). Such linkage is achieved via covalent linkage between the antibody or antibody fragment and its cargo, and such linkage is optionally facilitated by a linker. In some embodiments, the linked cargo is a drug or agent with an intended therapeutic effect.

[0061] The term "antibody fusion" refers to the combination of an antibody or antibody fragment with a non-antibody polypeptide into a single polypeptide. For example, an antibody fusion may comprise an antibody in which one or more heavy or light chains have been combined with a cytokine. An antibody fusion can be generated by joining two or more genes that originally coded for the separate components of the fusion into a single new gene that, when translated, results in the combined polypeptide.

[0062] As used herein, the term "host cell" can be any type of cell, e.g.. a primary’ cell, a cell in culture, or a cell from a cell line. In specific embodiments, the term "host cell" refers to a cell transfected with a nucleic acid molecule and the progeny or potential progeny of such a cell.Progeny of such a cell may not be identical to the parent cell transfected with the nucleic acid molecule, e.g., due to mutations or environmental influences that may occur in succeeding generations or integration of the nucleic acid molecule into the host cell genome.Screening Components for and Construction of Multicomponent Polypeptides

[0063] When producing multicomponent polypeptides, a challenge of production is to ensure that the various components are effectively paired, in the proper orientation, and in the correct stoichiometry'. Deficiencies in this regard results in accumulation of related, but undesired, polypeptides. The methods provided below aid in avoiding such deficiencies.

[0064] For example, recognizing the challenges that related antibody-based molecules and other multicomponent polypeptides present to purification of the desired species, a strategy has been developed to use surface charge as the basis for maximizing purification success rates. This strategy' focuses on the rational selection of building blocks that could be used as part of antibody -based molecules or other multicomponent polypeptides. This strategy increases the chance of generating desired polypeptides that will be well-resolved from contaminants during chromatographic purification or analysis. Therefore, the strategy facilitates identification of multicomponent polypeptides that can be more efficiently and effectively purified. Below, exemplary embodiments applying this strategy' are discussed in terms of preparing antibodybased molecules; these exemplary embodiments are intended to be applicable more broadly to multicomponent polypeptides in general.

[0065] In some embodiments, the developed strategy for building block pair selection is called purification-based Building Block Compatibility Screening (pBBCS). This is a 2-step assay that first requires empirically characterizing the physicochemical properties of building blocks (e.g., CIEX elution profiles) by expressing and purifying parental proteins from which the building blocks are derived. For example, in the context of a designing a bispecific antibody that could entail characterizing the properties of parental antibodies from which each arm of the bispecific antibody was derived. This first step may be adapted to a high-throughput process to enable the quick expression and purification of all candidate building blocks. Secondly, analysis of the physiochemical properties in question is performed to determine if two building blocks can be expected to combine into to generate a well-behaved multicomponent polypeptide. For example, in the context of a bispecific antibody that might be created from parental antibodies, analyses of the CIEX profiles of candidate parental antibodies (z.e., the candidate building blocks) may be used to determine if antibody arms can be combined to generate a well-behavedbispecific. In this exemplary setup, the second step would entail calculating the difference in elution conductivities of two building blocks (z.e., AConductivity or ACond) of all possible building block combinations for generating a bispecific. FIG. 3 is a schematic overview of pBBCS-aided CIEX separation that can be achieved by using AConductivity as an indicator of the extent to which common contaminant species will separate from the desired species during CIEX. Applying this in silico selection of compatible building blocks will reduce screening panel size and thereby save time and resources in the development of bispecifics.

[0066] In some embodiments, the physiochemical property to be determined as part of pBBCS relates to the charge of proteins. Properties that may be used include those that allow prediction of how the surface charge building blocks will contribute to the surface charge of a multicomponent polypeptide, and thus contribute to an understanding of how well undesired species can be separated from a desired multicomponent polypeptide. Relevant charge-related properties that may be determined include the pH. pl, pKa, elution conductivity, surface residue charge, and number of charged residues of the proteins to be analyzed. The value or relative magnitude of charge-related properties can be determined by any range of methods. Methods that may be used in the determination of the above charge-related properties of proteins include electrophoretic mobility assays (e.g., capillary electrophoresis, electrophoretic light scattering, or membrane confined electrophoresis), steady state electrophoresis, ion exchange chromatography (e.g., anion exchange chromatography (AIEX) or cation-exchange chromatography (CIEX)), hydrophobic interaction chromatography, reversed phase chromatography, and mass spectrometry. In exemplary embodiments, the physiochemical property to be determined is the elution conductivity of the proteins in question, and the elution conductivity may be determined by chromatographic methods such as AIEX or CIEX.

[0067] In some embodiments, one or more building blocks can be selected from a pool of candidates molecules (which may be defined as. e.g., "pool A") that comprise candidate building blocks. In some embodiments, two or more building blocks can be selected from two pools of candidate molecules (which may be defined as, e.g., "pool A" and "pool B") that comprise candidate building blocks A and B, respectively. When analysing the one or more pools of candidates for potential inclusion in a multicomponent polypeptide, the physical property in question for each member of those pools may be analyzed. The determined values of that physical property for a given pool of candidates may then be compared to the values of that physical property for another component, or to the values of candidates from another pool. Insome embodiments, a physical property is measured for all members of a first pool (pool A) and of a second pool (pool B), and the those values are used to select two or more building blocks for a multicomponent polypeptide. In exemplary' embodiments, the physiochemical property to be determined in a pool A and a pool B is the elution conductivity of the proteins in question, and the elution conductivity may be determined by chromatographic methods such as AIEX or CIEX.

[0068] In some embodiments, the values of the physical property' to be determined for candidate building blocks are used to predict the value of the physical property of potential multicomponent polypeptides. As an approximation, the value of a physical property of a multicomponent polypeptide can be considered to be the average of the value of that physical property for each of its component parts. Therefore, when a priority is to be able to differentiate the multicomponent polypeptide from its component parts, building block selection may comprise comparing the property' values for candidate components and selecting those components with the greatest difference in magnitude. A large difference in component property magnitude results in the greatest chance that the multicomponent polypeptide's physical property' value will be distinguishable from the physical property 's values for the component parts.

[0069] In some embodiments, the phy sical property to be determined for the candidate building blocks is conductivity. Therefore, in some embodiments, the difference in elution conductivities (z.e., AConductivity or ACond) of all possible building block combinations is determined. The conductivity of candidate building blocks may' be determined by any method available in the art. In some embodiments, the conductivity' of each candidate building block is determined using cation-exchange chromatography (CIEX), anion-exchange chromatography (AIEX), hydrophobic interaction chromatography, or reversed phase chromatography.

[0070] In some embodiments, the method of screening or method of constructing multicomponent polypeptides comprises selecting component building blocks (e.g., a building block A and a building block B) for a multicomponent polypeptide based on those building blocks have a larger AConductivity than other candidate pairs of building blocks. For example, in some embodiments the selected building block A came from a pool A molecule measured to have a conductivity within the top 25% of tested pool A molecules, and the selected building block B came from a pool B molecule measured to have a conductivity' within the bottom 25% of tested pool B molecules. In some embodiments the selected building block A came from a pool A molecule measured to have a conductivity’ within the top 20% of tested pool A molecules, and the selected building block B came from a pool B molecule measured to have a conductivitywithin the bottom 20% of tested pool B molecules. In some embodiments the selected building block A came from a pool A molecule measured to have a conductivity within the top 15% of tested pool A molecules, and the selected building block B came from a pool B molecule measured to have a conductivity within the bottom 15% of tested pool B molecules. In some embodiments the selected building block A came from a pool A molecule measured to have a conductivity within the top 10% of tested pool A molecules, and the selected building block B came from a pool B molecule measured to have a conductivity7w ithin the bottom 10% of tested pool B molecules. In some embodiments the selected building block A came from a pool A molecule measured to have a conductivity within the top 5% of tested pool A molecules, and the selected building block B came from a pool B molecule measured to have a conductivity within the bottom 5% of tested pool B molecules. In some embodiments the selected building block A came from a pool A molecule measured to have a conductivity within the top 30% of tested pool A molecules, and the selected building block B came from a pool B molecule measured to have a conductivity within the bottom 30% of tested pool B molecules. In some embodiments the selected building block A came from a pool A molecule measured to have a conductivity' within the top 35% of tested pool A molecules, and the selected building block B came from a pool B molecule measured to have a conductivity within the bottom 35% of tested pool B molecules. In some embodiments the selected building block A came from a pool A molecule measured to have a conductivity within the top 40% of tested pool A molecules, and the selected building block B came from a pool B molecule measured to have a conductivity within the bottom 40% of tested pool B molecules.

[0071] In some embodiments, the method of screening or method of constructing multicomponent polypeptides comprises selecting component building blocks (e.g., a building block A and a building block B) for a multicomponent polypeptide based on those building blocks have at least a particular AConductivity. In some embodiments, the selected building blocks A and B, respectively, come from pool A and B molecules that have a AConductivity greater than 4 mS / cm. In some embodiments, the selected building blocks A and B, respectively, come from pool A and B molecules that have a AConductivity greater than 5 mS / cm. In some embodiments, the selected building blocks A and B, respectively, come from pool A and B molecules that have a AConductivity' greater than 6 mS / cm. In some embodiments, the selected building blocks A and B, respectively, come from pool A and B molecules that have a AConductivity greater than 7 mS / cm. In some embodiments, the selected building blocks A andB, respectively, come from pool A and B molecules that have a AConductivity greater than 8 mS / cm. In some embodiments, the selected building blocks A and B, respectively, come from pool A and B molecules that have a AConductivity' greater than 9 mS / cm.

[0072] In some embodiments, the pool of molecules being examined for candidate building blocks can be used in their entirety as part of a multicomponent polypeptide. For example, a cytokine or an scFv can be analyzed independently before being incorporated into a multicomponent polypeptide. In other embodiments, the pool of molecules being examined for candidate building blocks are only partially incorporated into a multicomponent polypeptide. For example, a pool of antibodies might be examined before using one or more of the domains or component antibody chains as components in the multicomponent polypeptide. In such cases of partial incorporation, measurement of a physical property of the pool antibody or other pool molecule can be used as a proxy for the physical property of the incorporated component.

[0073] In some embodiments, the pool molecules are immunoglobulins. In some embodiments, the pool molecules are antibodies or antibody fragments. When the pool molecules are antibodies or antibody fragments, the building blocks to be incorporated as components may be part or all of the antibody or antibody fragment. For example, the building block may be part or all of an antibody heavy chain, an antibody light chain, an antibody arm (i.e., a combination of an antibody heavy chain and an antibody light chain), an antibody fragment (e.g., a Fab fragment, a VH domain, a VL domain, an Fc region), or a polypeptide comprising domains from the antibody (e.g., an scFv).

[0074] Any multicomponent polypeptide may be constructed, and any components for such multicomponent polypeptides may be screened, using the methods described herein. In some embodiments, bispecific antibodies may be constructed using pools of antibodies that specifically bind to each of their two target antigens. In some embodiments, the bispecific antibodies to be constructed comprise two heavy chains and two light chains. In some embodiments, the bispecific antibodies to be constructed comprise two antibody arms, wherein each arm was taken from a separate pool of candidate antibodies. In some embodiments, the multicomponent polypeptide is comprises an antibody arm from a first pool and an Fc-linked scFv from a second pool. In some embodiments, the multicomponent polypeptide is a monovalent antibody comprising a first heavy chain from a first pool and a second heavy chain from a second pool.

[0075] In some embodiments, the multicomponent polypeptide to be constructed or screened comprises a combination of an antibody-derived component and a non-antibody-derived component. In some embodiments, the multicomponent polypeptide comprises an antibody fusion with anon-antibody protein. In some embodiments, the multicomponent polypeptide comprises an antibody -cytokine fusion. In some embodiments, the multicomponent polypeptide comprises an antibody-drug conjugate.

[0076] In some embodiments, the method of construction or screening comprises measuring the absolute conductivity of the components of the multicomponent polypeptide. In some embodiments, the method comprises estimating the absolute conductivity of the selected multicomponent polypeptides by summing the conductivity of the pool A molecule and pool B molecule that supply the building blocks for each multicomponent polypeptide, and keeping in the selected set only those multicomponent polypeptides within an upper range of the estimated summed conductivities. In some embodiments, the method comprises keeping in the selected set only those multicomponent polypeptides within the upper 60% of the estimated summed conductivities. In some embodiments, the method comprises keeping in the selected set only those multicomponent polypeptides within the upper 55% of the estimated summed conductivities. In some embodiments, the method comprises keeping in the selected set only those multicomponent polypeptides within the upper 50% of the estimated summed conductivities. In some embodiments, the method comprises keeping in the selected set only those multicomponent polypeptides within the upper 45% of the estimated summed conductivities. In some embodiments, the method comprises keeping in the selected set only those multicomponent polypeptides within the upper 40% of the estimated summed conductivities. In some embodiments, the method comprises keeping in the selected set only those multicomponent polypeptides within the upper 35% of the estimated summed conductivities. In some embodiments, the method comprises keeping in the selected set only those multicomponent polypeptides within the upper 30% of the estimated summed conductivities. In some embodiments, the method comprises keeping in the selected set only those multicomponent polypeptides within the upper 25% of the estimated summed conductivities. In some embodiments, the method comprises keeping in the selected set only those multicomponent polypeptides within the upper 20% of the estimated summed conductivities. In some embodiments, the method comprises keeping in the selected set only those multicomponent polypeptides within the upper 15% of the estimated summedconductivities. In some embodiments, the method comprises keeping in the selected set only those multicomponent polypeptides within the upper 10% of the estimated summed conductivities. In some embodiments, the method comprises keeping in the selected set only those multicomponent polypeptides within the upper 5% of the estimated summed conductivities.

[0077] In some embodiments, the method of construction or screening comprises selecting component building blocks (e.g., a building block A and a building block B) for a multicomponent polypeptides based on those building blocks have at least a particular absolute conductivity. In some embodiments, the absolute conductivity is determined by summing the conductivity of the component molecules, or summing the conductivity of the pool molecules from which the component molecules are derived. In some embodiments, the selected building blocks have an absolute conductivity greater than 4 mS / cm. In some embodiments, the selected building blocks have an absolute conductivity greater than 5 mS / cm. In some embodiments, the selected building blocks have an absolute conductivity greater than 6 mS / cm. In some embodiments, the selected building blocks have an absolute conductivity greater than 7 mS / cm. In some embodiments, the selected building blocks have an absolute conductivity greater than 8 mS / cm. In some embodiments, the selected building blocks have an absolute conductivity greater than 9 mS / cm. In some embodiments, the selected building blocks have an absolute conductivity greater than 10 mS / cm. In some embodiments, the selected building blocks have an absolute conductivity greater than 11 mS / cm. In some embodiments, the selected building blocks have an absolute conductivity greater than 12 mS / cm. In some embodiments, the selected building blocks have an absolute conductivity greater than 13 mS / cm. In some embodiments, the selected building blocks have an absolute conductivity greater than 14 mS / cm. In some embodiments, the selected building blocks have an absolute conductivity greater than 15 mS / cm.

[0078] In some embodiments, the method of construction or screening comprises comparing the elution profile of each selected multicomponent polypeptide with the elution profiles of the pool A molecule and the pool B molecule from which its component parts were derived. In some embodiments, the elution profile is determined by CIEX chromatography. In some embodiments, the method comprises keeping, or proceeding with, only those multicomponent polypeptides whose principal elution peak has a small amount of overlap in peak area with the principal elution peak of either the pool A molecule or the pool B molecule. In some embodiments, the multicomponent polypeptide's elution peak has less than 30% overlap in peak area with the principal elution peak of either the pool A molecule or the pool B molecule. Insome embodiments, the multicomponent polypeptide's elution peak has less than 25% overlap in peak area with the principal elution peak of either the pool A molecule or the pool B molecule. In some embodiments, the multicomponent polypeptide's elution peak has less than 20% overlap in peak area with the principal elution peak of either the pool A molecule or the pool B molecule. In some embodiments, the multicomponent polypeptide's elution peak has less than 15% overlap in peak area with the principal elution peak of either the pool A molecule or the pool B molecule. In some embodiments, the multicomponent polypeptide's elution peak has less than 10% overlap in peak area with the principal elution peak of either the pool A molecule or the pool B molecule. In some embodiments, the multicomponent polypeptide's elution peak has less than 5% overlap in peak area with the principal elution peak of either the pool A molecule or the pool B molecule. In some embodiments, the multicomponent polypeptide's elution peak has no overlap in peak area with the principal elution peak of either the pool A molecule or the pool B molecule.Methods of Producing Proteins

[0079] A host cell, when cultured under appropriate conditions, synthesizes heterodimeric antibody that can subsequently be collected from the culture medium (if the host cell secretes it into the medium) or directly from the host cell producing it (if it is not secreted). The selection of an appropriate host cell will depend upon various factors, such as desired expression levels, polypeptide modifications that are desirable or necessary for activity (such as glycosylation or phosphorylation) and ease of folding into a biologically active molecule. A host cell may be eukaryotic or prokaryotic.

[0080] Mammalian cell lines available as hosts for expression are known in the art and include, but are not limited to, immortalized cell lines available from the American Type Culture Collection (ATCC) and any cell lines used in an expression system known in the art can be used to make the recombinant polypeptides of the invention. In general, host cells are transformed with a recombinant expression vector that comprises DNA encoding a desired heterodimeric antibody. Among the host cells that may be employed are prokaryotes, yeast, or higher eukaryotic cells. Prokaryotes include gram negative or gram positive organisms, for example E. coli or bacilli. Higher eukary otic cells include insect cells and established cell lines of mammalian origin. Examples of suitable mammalian host cell lines include the COS-7 line of monkey kidney cells (ATCC CRL 1651) (Gluzman et al., 1981. Cell 23: 175), L cells, 293 cells, C127 cells. 3T3 cells (ATCC CCL 163), Chinese hamster ovary (CHO) cells, or their derivatives such as Veggie CHO and related cell lines which grow in serum-free media (Rasmussen et al..1998, Cytotechnology 28: 31), HeLa cells, BHK (ATCC CRL 10) cell lines, and the CV1 / EBNA cell line derived from the African green monkey kidney cell line CV1 (ATCC CCL 70) as described by McMahan et al., 1991, EMBO J. 10: 2821, human embryonic kidney cells such as 293, 293 EBNA or MSR 293, human epidermal A431 cells, human Colo205 cells, other transformed primate cell lines, normal diploid cells, cell strains derived from in vitro culture of primary tissue, primary explants, HL-60, U937, HaK or Jurkat cells. Optionally, mammalian cell lines such as HepG2 / 3B, KB, NIH 3T3 or S49, for example, can be used for expression of the polypeptide when it is desirable to use the polypeptide in various signal transduction or reporter assays. Alternatively, it is possible to produce the polypeptide in lower eukaryotes such as yeast or in prokaryotes such as bacteria. Suitable yeasts include Saccharomyces cerevisiae, Schizosaccharomyces pombe, Kluyveromyces strains, Candida, or any yeast strain capable of expressing heterologous polypeptides. Suitable bacterial strains include Escherichia coli, Bacillus subtilis, Salmonella typhimurium, or any bacterial strain capable of expressing heterologous polypeptides.

[0081] If the antibody or antibody fragment is made in yeast or bacteria, it may be desirable to modify the product produced therein, for example by phosphorylation or glycosylation of the appropriate sites, in order to obtain a functional product. Such covalent attachments can be accomplished using known chemical or enzymatic methods. A polypeptide can also be produced by operably linking the isolated nucleic acid of the invention to suitable control sequences in one or more insect expression vectors, and employing an insect expression system. Materials and methods for baculovirus / insect cell expression systems are commercially available in kit form from, e.g., Invitrogen, San Diego, Calif.. U.S.A, (the MaxBac® kit), and such methods are known in the art, as described in Summers and Smith, Texas Agricultural Experiment Station Bulletin No. 1555 (1987), and Luckow and Summers, Bio / Technology 6:47 (1988). Cell-free translation systems could also be employed to produce polypeptides, such as antibodies or antibody fragments, using RNAs derived from nucleic acid constructs disclosed herein. Appropriate cloning and expression vectors for use with bacterial, fungal, yeast, and mammalian cellular hosts are described by Pouwels et al. (Cloning Vectors: A Laboratory Manual, Elsevier, New York, 1985). A host cell that comprises an isolated nucleic acid of the invention, preferably operably linked to at least one expression control sequence, is a “recombinant host celk’.

[0082] Cell lines may be selected through determining which cell lines have high expression levels and constitutively produce antigen binding proteins with the desired binding properties. A cell line from the B cell lineage that does not make its own antibody but has a capacity to make and secrete a heterologous antibody can be selected.PARTICULAR EMBODIMENTS

[0083] Particular embodiments of the invention include the following.1. A method of screening potential components for multicomponent polypeptides, wherein the multicomponent polypeptides each comprise at least one building block A and at least one building block B, and wherein the screening comprises: a. identifying and individually purify ing a group of molecules (pool A) wherein each molecule comprises one or more copies of a candidate building block A; b. identifying and individually purifying a group of molecules (pool B) wherein each molecule comprises a one or more copies of a candidate building block B; c. determining the conductivity of each pool A molecule and of each pool B molecule; d. determining the difference in conductivity (ACond) between each molecule of pool A and each molecule of pool B; and e. selecting a set of candidate multicomponent polypeptides that comprise a building block A-building block B pair, wherein the pool A and pool B molecules comprising the building blocks in the selected multicomponent peptides have a larger ACond than the ACond between the pool A and pool B molecules comprising the building block pairs in the non-selected multicomponent polypeptides.2. The method of embodiment 1, wherein the conductivity of each pool A molecule and of each pool B molecule is determined using cation-exchange chromatography (CIEX), anion-exchange chromatography (AIEX), hydrophobic interaction chromatography, or reversed phase chromatography.The method of embodiment 2, wherein the conductivity of each pool A molecule and of each pool B molecule is determined using CIEX. The method of embodiment 3, wherein the selected multicomponent peptides have a ACond between the pool A and pool B molecules greater than 5 mS / cm. The method of embodiment 4, wherein the selected multicomponent peptides have a ACond between the pool A and pool B molecules greater than 7 mS / cm. The method of embodiment 1, wherein the selected building block A came from a pool A molecule measured to have a conductivity within the top 25% of tested pool A molecules, and wherein the selected building block B came from a pool B molecule measured to have a conductivity within the bottom 25% of tested pool B molecules. The method of embodiment 6, wherein the selected building block A came from a pool A molecule measured to have a conductivity within the top 10% of tested pool A molecules, and wherein the selected building block B came from a pool B molecule measured to have a conducti vity within the bottom 10% of tested pool B molecules. The method of any of embodiments 1-7, wherein each pool A molecule and each pool B molecule comprises an immunoglobulin. The method of embodiment 8, wherein each pool A molecule and each pool B molecule comprises an antibody. The method of any of embodiments 1-9, wherein building block A comprises all or part of an antibody heavy chain. The method of embodiment 10, wherein building block A comprises an antibody arm. The method of any of embodiments 1-11, wherein building block B comprises all or part of an antibody heavy chain. The method of embodiment 12, wherein building block B is an antibody arm. The method of any of embodiments 9-13, wherein the multicomponent polypeptide is a bispecific antibody.The method of any of embodiments 9-14, wherein building block A comprises an antibody arm, and wherein building block B comprises an Fc-linked scFv. The method of any of embodiments 8-15, wherein the multicomponent polypeptide is an antibody drug conjugate or antibody fusion. The method of embodiment 16, wherein the multicomponent polypeptide comprises an antibody fusion, and wherein the antibody is fused to a cytokine. The method of any of embodiments 1-17, wherein the method further comprises estimating the absolute conductivity of the selected multicomponent polypeptides by summing the conductivity of the pool A molecule and pool B molecule that supply the building blocks for each multicomponent polypeptide, and keeping in the selected set only those multicomponent polypeptides within the upper 50% of the estimated summed conductivities. The method of embodiment 18, wherein the method comprises keeping only those multicomponent polypeptides with an absolute conductivity greater than 11 mS / cm. The method of any of embodiments 1-19, wherein the method further comprises comparing the CIEX elution profile of each selected multicomponent polypeptide with the CIEX elution profiles of the pool A molecule and the pool B molecule from which its component parts were derived, and keeping only those multicomponent polypeptides whose principal elution peak have less than 25% overlap in peak area with the principal elution peak of either the pool A molecule or the pool B molecule. The method of embodiment 20, where there is less than 10% overlap in peak area between the multicomponent peptide's principal peak and the principal peak of either the pool A molecule or the pool B molecule. A method of constructing a multicomponent immunoglobulin molecule comprising the steps of: a. measuring the surface charge of each member of a pool A of immunoglobulins and measuring the surface charge of each member of a pool B of immunoglobulins;b. selecting an immunoglobulin from pool A (immunoglobulin A) and an immunoglobulin from pool B (immunoglobulin B) that have a difference in conductivity' (ACond) greater than 5 mS / cm, as measured by CIEX; and c. combining a building block A that comprises part or all of immunoglobulin A with a building block B that comprises part or all of immunoglobulin B in the multicomponent immunoglobulin molecule. The method of embodiment 22, wherein the selected immunoglobulin A and immunoglobulin B have a ACond greater than 7 mS / cm, as measured by CIEX. The method of embodiment 22 or embodiment 23, wherein each pool A molecule and each pool B molecule comprises an immunoglobulin. The method of embodiment 24, wherein each pool A molecule and each pool B molecule comprises an antibody. The method of any of embodiments 22-25, wherein building block A comprises all or part of an antibody heavy chain. The method of embodiment 26, wherein building block A comprises an antibody arm. The method of any of embodiments 22-27, wherein building block B comprises all or part of an antibody heavy chain. The method of embodiment 28, wherein building block B is an antibody arm. The method of any of embodiments 25-29, wherein the multicomponent polypeptide is a bispecific antibody. The method of any of embodiments 25-30, wherein building block A comprises an antibody arm, and wherein building block B comprises an Fc-linked scFv. The method of any of embodiments 24-31, wherein the multicomponent polypeptide is an antibody drug conjugate or antibody fusion. The method of embodiment 32, wherein the multicomponent polypeptide comprises an antibody fusion, and wherein the antibody is fused to a cytokine.34. The method of any of embodiments 22-33, wherein the method further comprises, prior to combining the building blocks, estimating the absolute conductivity of multicomponent polypeptides comprising the selected components by summing the conductivity of the pool A molecule and pool B molecule that supply the building blocks for each multicomponent polypeptide, and constructing only those multicomponent polypeptides within the upper 50% of the estimated summed conductivities.35. The method of embodiment 34, wherein the method comprises constructing only those multicomponent polypeptides with an absolute conductivity greater than 11 mS / cm.EXAMPLESExample 1: Plasmid Construction, Cell Culture, and Protein Expression

[0084] Unless otherwise indicated, the plasmid construction, cell culture, plasmid transfection, protein expression, and antibody purification referenced in subsequent examples were performed as described in this Example.

[0085] Plasmid Construction: genes to be expressed as polypeptides were synthesized by Twist Bioscience. Genes were cloned individually into vectors for mammalian transient expression using the Golden Gate assembly method. See Engler C et al.. PLoS One 3:e3647 (2008). which is incorporated herein by reference in its entirety) All chains containing Fc regions were constructed using a human IgGl scaffold (IgGl-SEFL2.2) carrying an aglycosylation mutation (N297G), as well as an engineered disulfide bond in the CH2 domain. {See Estes et al., iScience 24: 103447 (2021); and Jacobsen FW et al., J. Biol. Chem. 292: 1865-75 (2017); which are incorporated herein by reference in their entirety’) All molecules requiring hetero-Fc pairing included charge pair mutations (“CPM”) vl 03 in the CH3 regions to promote heterodimerization. {See Estes et al., iScience 24: 103447 (2021)) Optionally, scFv sequences contained an engineered disulfide to increase stability of the Fv. After sequencing confirmation by Sanger, transfection-grade DNA was prepared using Maxi plasmid purification kits (Qiagen).

[0086] Cell Culture, Plasmid Transfection, and Protein Expression: proteins were transiently expressed in suspension human embryonic kidney 293-6E cells (NRC-BRI) using the PEImax transfection reagent. Prior to transfection, the cells were maintained in FreeStyle F-17 medium (Gibco, catalog # A1383502) with 0. 1% Kolliphor P188 (Sigma, catalog # K4894), 25 pg / ml G418 (Gibco, catalog # 10131027) and 6 mM L-glutamine (Gibco, catalog # 25030149).To achieve an optimal density of 2x106viable cells per mL for transfection, cells were passaged 26 hours prior to transfection. The purified plasmids were mixed at a ratio of 1 : 1 (HC:LC) for mAbs, 1 : 1 : 1 : 1 for Hetero-IgGs, 1 : 1 for IgG-scFvs, 1 : 1 : 1 for AmAbs and other 3-chain formats containing scFvs. For each mL of cell media, 0.5 pig DNA was incubated with 1.5 pL PEImax reagent (Polysciences, catalog # 24765-2) in 100 pL FreeStyle F-17 medium for 10 minutes, then added to the cell culture. 24 hours following transfection, cell cultures were fed with Tryptone N1 solution (Organotechnie, catalog # 19553) and glucose (Thermo Fisher, catalog # A2494001) to a final concentration of 2.5 g / L and 4.5 g / L, respectively. Seven days following transfection, the conditioned medium was harvested for purification.

[0087] Two-step antibody purification with Protein A and cation exchange chromatography: Following harvest of 40 mL cultures, the conditioned medium was purified by Protein A affinity capture (1 ml HiTrap MabSelect SuRe, GE Life Sciences, catalog # GE11- 0034-93) equilibrated with 25 mM Tris, 100 mM NaCl. pH 7.4. Protein were eluted with 100 mM sodium acetate, pH 3.6 followed immediately by buffer exchange into 10 mM sodium acetate, 150 mM NaCl, pH 5.2 using a 5 ml HiTrap Desalting column (GE Life Sciences, catalog # GE17-1408-01). Final eluate volumes were approximately 1.8 mL.

[0088] For cation-exchange chromatography (CIEX), Protein A eluates were diluted with 20 ml of 20 mM MES, pH 6.2 and loaded onto a 1 ml cation ion-exchange column (HiTrap SP- HP, GE Life Sciences, catalog # GE29-0513-24) at 1 ml / min. After washing with 8 column volumes of dilution buffer at 1 ml / min, the samples were eluted with a linear 0-400 mM NaCl gradient over 40 column volumes at 0.4 ml / min. Purity of purified samples were analyzed with non-reducing micro capillary electrophoresis (MCE) and analytical size exclusion chromatography (SEC). For non-reducing MCE, 6 pL of proteins were mixed with 21 pL of sample buffer (8.4 mM Tns-HCl pH 7.0, 7.98% Glycerol, 2.38 mM EDTA, 2.8% SDS and 2.4 mM lodoacetamide), heated to 85°C for 10 min, and analyzed using a Caliper LabChip GXII Touch instrument (PerkinElmer). For analytical SEC, protein samples were analyzed on an Acquity HPLC instrument (Waters) using a Zenix-C column (300 A, 3 micron, 4.6x300 mm) (Sepax Technologies) with 100 mM sodium phosphate pH 6.8, 250 mM NaCl as the running buffer at 0.45 ml / min flow rate. Fraction pools of 90% or higher purity7(as determined by SEC and MS) were collected, and their concentrations were determined using Nanodrop 8000 (Thermo Fisher). The final pooled samples were independently analyzed by SEC to confirm purity.Example 2: Correlation of Antibody Structure with Chromatogra hic Separation

[0089] To gain a deeper insight into the molecular basis for chromatographic separation of bispecific antibodies, a series of experiments were performed to correlate the CIEX profdes of exemplary bispecific antibodies to the characteristics of their constituent protein modules.

[0090] The exemplary bispecific antibodies examined are show in in FIGS. 1A-1C, which are diagrams of a series of different antibody structures showing the Fc, Fv, Fab, CDR, and constant (VH, VL, CL, and CHI) regions identified on one of the arms of each antibody. FIG. 1 A shows the monospecific and bivalent antibody structure used in most natural antibodies. Such an antibody contains two arms that can serve as a building block in constructing bispecific antibodies (labelled as “Building Block A”). Two bispecific antibodies constructed by combining Building Block A with other building blocks are shown in FIGS. IB and 1C. FIG. IB shows an exemplary bispecific antibody in which an arm capable of binding antigen A is linked to an arm capable of binding to antigen B (Building Block B) through their respective Fc regions, where the interaction's specificity is facilitated by complementary charge-pair mutations in those Fc regions. FIG. 1C shows an exemplary AmAb. which is comprised of a standard antibody arm capable of binding antigen A (Building Block A) and an antibody arm capable of binding antigen B that has an scFv in place of the Fab region (Building Block B'). While the various building blocks can be combined to form desired AmAbs or other bispecific antibodies, those same building blocks may be present as. or form into, impurities that are difficult to purify away from the desired species. For example, FIG. IF shows exemplary impurities that may be present alongside the desired AmAb species when expressing the components of that AmAb.

[0091] Even with considerable alterations of the quaternary7structures, most of the final protein sequence in bispecific antibodies is derived from the initial building blocks (FIGS. ID and IE). For example, when charge pairing mutations (CPMs) are deployed to drive chain pairing in some bispecific antibodies (FIG. ID), those changes lead to only minimal differences (< 1%) from the sequence of parental antibodies, and even these changes are purposely buried in the interface. (See Estes el al. , iScience 24: 103447 (2021), which is incorporated herein by reference in its entirety) Similarly, only approximately 2% of an AmAb’s sequence is not derived from the parental antibodies (FIG. IE).

[0092] In view of the high percentage of bispecific antibodies’ sequence that is derived from parental antibodies, experiments were performed to test whether the surface charge of a given bispecific antibody is a composite of the two individual parent antibodies from which thebispecific antibody’s components were derived. The plasmids necessary for these experiments were constructed and used to express protein as described in Example 1 above. Briefly, sequences encoding each heavy and light chain of a Hetero-IgG with two distinct Fvs capable of binding antigen A and antigen B were cloned into the pTT25GS vector using the Golden Gate assembly method, with one plasmid encoding each polypeptide chain. Plasmids encoding the needed heavy and light chains of the Hetero-IgG’ s parental antibodies (i.e., a Building Block A Homodimer and a Building Block B Homodimer, capable of binding antigen A and antigen B, respectively)) were constructed using the same method and the same plasmid vector. Subsequently, each set of plasmids necessary for expressing a given bispecific antibody was transiently expressed in a population of suspension human embryonic kidney 293-6E cells (NRC- BRI) using the PEImax transfection reagent. Seven days following transfection, the conditioned medium from each population of transfected cells was harvested for antibody purification.

[0093] The Hetero-IgG and each parental antibody was separately purified from the conditioned medium using Protein A affinity capture (1 ml HiTrap MabSelect SuRe, GE Life Sciences, catalog # GE11-0034-93), as described in Example 1.

[0094] After elution, each antibody solution was analyzed using CIEX chromatography to obtain a CIEX elution profile for those antibody solutions. Raw chromatogram traces were exported (UNICORN 7, Cytiva) and CIEX profiles were generated with conductivity plotted against A280nm (Microsoft Excel). Elution positions, defined as the conductivity at which maximum A280nm signal was detected, were manually identified. Peak Separation was calculated as the difference in conductivities between the peaks indicated. Negative values were assigned as appropriate for Peak Separation when the AmAb elution position was earlier than both parental mAbs. For cases in which Peak 2 was directly overlapped with Peak 1, the peak separation was scored as zero. Separation was also qualitatively scored as either “good” or “bad” (as shown in FIGS. 8A-8B) based on the overlap observed between the AmAb main peak and the corresponding mAbs.

[0095] The CIEX elution profile for the Hetero-IgG bispecific, Building Block A Homodimer, and Building Block B Homodimer w ere overlaid in a single chart, showm in FIG. 2A. In FIG. 2A, each antibody’s elution peak is labelled, illustrating that the Hetero-IgG eluted at a conductivity between those of the building block homodimers. Thus, because the CIEX elution profile can be used as a surrogate for surface charge, the Hetero-IgG has a surface charge between those of the building block homodimers. Similarly, the CIEX elution profile for anAmAb, Building Block A Homodimer, and Building Block B Homodimer were overlaid in a single chart, shown in FIG. 2B. As with the Hetero-IgG, the AmAb had a surface charge between those of the building block homodimers.

[0096] Next, CIEX analysis was performed with other bispecific antibody formats containing other architectures and valencies. The plasmids necessary for these experiments were constructed and used as described in Example 1 above, using the same purification methods as described in Example 1. For each of these bispecific antibodies, a CIEX elution profile w as obtained, and subsequently these elution profiles were overlaid to generate FIGS. 2C and 2D. As seen in FIGS. 2C and 2D, adding one or two scFvs resulted in a bispecific molecule that elutes at a conductivity near the midpoint of the two parental mAbs. This was true regardless of the configuration used, as antibodies with each of the tested ClmAb, C2mAb, Bl mAh, and N1 mAh formats eluted at a conductivity betw een the conductivities of their parental antibodies (FIG. 2D). The fusion position of the scFv module appears to induce shifts in the elution positions. For example, fusion of the scFv to the N-terminus of an antibody (NlmAb), imparted a greater apparent positive charge and delayed CIEX migration compared to other scFv fusion positions (ClmAb and BlmAb) (FIG. 2D). This conductivity shift towards the later eluting antibody is likely due to the scFv module in NlmAb remaining fully exposed to the solvent, while one of the negatively charged Fab fragments from Target A becomes more concealed by the body of that same scFv. Consequently, ClmAb and BlmAb, with the scFv module occluded at the N- terminus by either the Fc domain or by a Fab, display very7similar elution profdes (FIG. 2D). The primary7contaminant in the panel of antibodies depicted in FIG. 2D was a noncovalent homodimer that maintains a mAb-like quaternary structure due to HC / HC mispairing via the CH3 dimer, mimicking the CIEX profile of the corresponding parental mAh. This contaminant may be the result of an imbalance in chain expression, yvith the scFv-containing polypeptide chain expressing at a low er level and the excess scFv-free chain driving homodimer formation. The data in FIG. 2C also demonstrates that the stoichiometry of components (z.e., the number of scFvs) also impacts a bispecific antibody’s CIEX elution profile. A bispecific antibody with two scFvs (C2mAb) displayed a greater apparent positive charge than a bispecific antibody yvith a single scFv (ClmAb) (FIG. 2C).

[0097] The effects of non-antibody building blocks on the CIEX elution profiles of bispecific antibodies were also analyzed. The deployment of cytokines and other native protein ligands for the assembly of bispecific formats (specifically, antibody fusions) are attractive. Forexample, such antibody fusions can be used to generate a molecule capable of accurately mimicking the cytokine binding to the corresponding receptor. CIEX analysis was performed with a fusion of an antibody targeting antigen A and a cytokine, and that fusion’s CIEX elution profile was compared to that of the parental antibody and the unbound cytokine. The plasmids necessary for these experiments were constructed and used as described in Example 1 above, using the same purification methods as described in Example 1. For each of these bispecific antibodies, a CIEX elution profile was obtained, and subsequently these elution profiles were overlaid to generate FIG. 2E. These data showed that the antibody-cytokine fusion eluted at a conductivity between that of the parental antibody (i.e.. Building Block A homodimer) and the cytokine building block.

[0098] Overall, the data in FIGS. 2A-2E demonstrate that each bispecific antibody’s conductivity was a blend of the conductivity of its parent antibodies and components. These observations suggest that the physicochemical properties of a bispecific are derived from the building blocks of which they are composed. Thus, these findings highlight that the surface charge of individual Fvs can be used to predict the elution profiles of bispecific molecules generated from those Fvs.Example 3: Applying Purification-Based Building Block Compatibility Screening

[0099] Given the challenges that related antibody-based molecules present to bispecific purification and the data from Example 2, it was recognized that surface charge could be used as the basis for a strategy to maximize purification success rates through the rational selection of building blocks. Such a strategy would present an increased chance of generating bispecific antibodies that are well-resolved from contaminants during chromatographic purification or analysis.

[0100] To that end, strategy for building block pair selection was developed, which is called Purification-Based Building Block Compatibility Screening (pBBCS). This 2-step assay first requires empirically characterizing the physicochemical properties of building blocks (e.g., CIEX elution profiles) by expressing and purifying parental mAbs from which the building blocks are derived. The first step may be adapted to a high-throughput process to enable the quick expression and purification of all selected building blocks. Secondly, analyses of the CIEX profiles are used to determine if two building blocks can be combined to generate a well-behaved bispecific. The second step calculates the difference in elution conductivities of two building blocks i.e., AConductivity or ACond) of all possible building block combinations for generatinga bispecific. FIG. 3 is a schematic overview of pBBCS-aided CIEX separation that can be achieved by using AConducti vity as an indicator of the extent to which common contaminants such as homodimeric and single-arm species will separate from the AmAb species during CIEX. Applying this in silico selection of compatible building blocks will reduce screening panel size and thereby save time and resources in the development of bispecifics.

[0101] An exemplary series of experiments were performed to demonstrate the predictive power of pBBCS using the AmAb format (see FIG. 1C). Like other bispecifics, the three-chain AmAb often presents production challenges due to the appearance of unwanted quaternary structure. Thus, pBBCS was applied to identify and pair two building blocks displaying a large AConductivity that enabled peak resolution in CIEX relative to contaminating structures. First, to build a library of building blocks for pBBCS, 8 unique antibodies against Target A (Pool A) and 17 unique antibodies against Target B (Pool B) were expressed and purified. Pairwise combination of the HC / LC building block from an antibody in Pool A and an Fc-linked scFv generated from an antibody from Pool B allows for 136 different potential AmAbs.

[0102] The CIEX elution profile w as determined for each of the antibodies in Pool A and Pool B using the protocol outlined in Example 1. These elution profiles w ere then overlaid to generate the charts in FIG. 4A (Pool A) and FIG 4B (Pool B). The antibodies in each pool had a wide range of different pi’s, as indicated by the wide range of conductivities observed in the CIEX profiles. The main peaks for the antibodies capable of binding Target A were centered at conductivities ranging from approximately 6 to 13 mS / cm, while the antibodies capable of binding Target B w ere centered at conductivities ranging from approximately 6 to 17 mS / cm. Pairwise comparison of each antibody in Pool A with each antibody in Pool B (z.e., review of all 136 combinations) revealed that AConductivity between their respective main peaks ranged from 0 to approximately 11 mS / cm. The AConductivity for each potential AmAb is provided in TAB E 1 below-.TABLE 1

[0103] As an initial threshold, all combinations with a AConductivity greater than 5 mS / cm were used to generate AmAbs with a HC / LC building block from Pool A and an relinked scFv derived from a Pool B antibody. Forty-five of the combinations met this threshold (leaving 91 that did not). Each of these 45 AmAbs was expressed in 293 -6E cells and purified with Protein A chromatography as described in Example 1. To ensure sufficient antibody for subsequent experiments, a secondary threshold of having purified antibody solutions of 100 mg / L or greater was set. This led to the deselection of a further 11 AmAbs, leaving 34 samples for further testing. A plot of the AConductivities and mean protein A yield for each AmAb is shown in FIG. 6. with the upper-right quadrant showing the qualifying 34 antibodies. Eight AmAbs predicted to have poor separation from contaminants (AConductivities below 5 mS / cm; from the lower-right quadrant in FIG. 6) were also purified using the same method, for use as negative controls.

[0104] The 34 remaining candidate AmAb samples and eight control AmAb samples were analyzed using CIEX, and the conductivity of the main peak present for each AmAb was determined. FIG. 7 shows a plot of the measured conductivity and the predicted conductivity (z.e., the midpoint of the two parent mAbs). The resultant AmAb elution conductivities strongly correlated with the elution conductivity predicted by pBBCS (z.e., the midpoint of the two mAbs) (R = 0.978) (FIG. 7).

[0105] Superimposing AmAbs with their corresponding parental mAbs led to overlays that showed that a high AConductivity leads to greater separation during CIEX. FIGS. 5A-5C show exemplary overlays of AmAbs with their corresponding parental mAbs. FIG 5 A shows an overlay in which a relatively low AConductivity between parental mAbs (A cond. = 3.6 mS / cm) resulted in an AmAb with a peak that largely overlapped with the parental mAb peaks (i.e., resulted in poor separation). FIG. 5B shows an overlay with a AConductivity between parental mAbs of 6.3 mS / cm, which resulted in an AmAb with a peak that only slightly overlapped with the parental mAb peaks. FIG. 5C shows an overlay in which a relatively high AConductivity between parental mAbs (A cond. = 10.5 mS / cm) resulted in an AmAb with a peak that did not overlap with the parental mAb peaks. Analyzing these overlays for all AmAbs in the panel indicated that complete separation of the AmAb from their corresponding parental mAbs was achieved when AConductivity was greater than approximately 7 mS / cm (FIGS. 8 A and 8B). For bispecifics with AConductivity > 7 mS / cm, 100% (17 out of 17) of AmAbs showed completeseparation from parental antibody peaks; in contrast, only 8% (2 out of 25) of AmAbs with AConductivity < 7 mS / cm showed complete separation. Interestingly, certain molecules with very low AConductivities (< 3 mS / cm) eluted at positions earlier than both parental mAbs, indicating that predictions about separation may not be accurate below a certain AConductivity threshold. When the original 136 combinations were filtered using a AConductivity cut-off of > 7 mS / cm, 87.5% of AmAb combinations (119 out of 136) could be deselected based on elution conductivity and yield data from the parental mAbs (FIG. 8C), enabling screening efforts to focus on molecules with higher likelihoods of success. By applying a AConductivity cutoff of 7 mS / cm, it can be determined whether that bispecific offers the necessary C1EX peak resolution to be easily separated from its impurities. Generally, such impurities are subspecies with deviated quaternary assemblies of the selected building blocks.

[0106] In addition, comparing isoelectric point (pl) predictions to elution conductivities of parental mAbs showed that pl predictions (calculated using the Sillero method) are modestly predictive of CIEX behavior (FIG. 8D). Thus, while the predicted pl of parental antibodies can be used as a proxy to predict bispecific antibody conductivity, empirical CIEX data remains the superior method to implement the pBBCS workflow.Example 4: Identification and Removal of a High-Molecular-Weight Contaminant Generated Through scFv Self-Association

[0107] In the data gathered from the panel of antibodies examined Example 3, several AmAbs contained a second, smaller peak eluting after the main peak in CIEX (FIG. 9A). To gain a better understanding of the protein present in each peak, eluate from peak 1 and peak 2 were each subjected to size exclusion chromatography (SEC), liquid chromatography-mass spectrometry (LC-MS), and a CIEX re-run (FIGS 9B-9G). In addition, protein A eluate of the same AmAb (i.e., without initial CIEX separation) was subjected to SEC.

[0108] For analytical SEC. protein samples were analyzed on an Acquity HPLC instrument (Waters) using a Zenix-C column (300 A, 3 micron. 4.6x300 mm)(Sepax Technologies) with 100 mM sodium phosphate pH 6.8, 250 mM NaCI as the running buffer at 0.45 ml / min flow rate.

[0109] Intact mass measurements were collected on CIEX fractions (approximately 5-15 pg) using non-reduced LC-MS. The LC-MS system consisted of an Agilent 1290 Infinity II UPLC connected to an Agilent 6230 TOF mass spectrometer, and Zorbax RRHD 300SP-C8 2.1x50mM (1.8 pm) UPLC column at a flow rate of 0.5 mL / min and heated to 70°C. Thechromatographic solvents consisted of 0. 1 % trifluoroacetic acid in water (Buffer A) and 0. 1 % trifluoroacetic acid in 90% n-propanol (Buffer B), which were run on an isocratic gradient of 20% B for 4 minutes, 72% B for 2 minutes, 90% B for 0.5 minutes, and 95% B for 0.5 minutes. Scans for m / z were acquired at a rate of 0.7 spectra per second. Using Mass Hunter BioConfirm Software v 10.0 (Agilent), the summed spectra were deconvoluted and reports were generated comparing deconvoluted mass vs. counts.

[0110] SEC of material from peaks 1 and 2 showed that both solutions exhibited a SEC peak eluting at approximately 6 minutes corresponding to the desired AmAb. while only peak 2 exhibited a second SEC peak with a substantially earlier SEC elution position (FIGS. 9B and 9C). Consistent with this, SEC of protein A eluate contained the same two peaks (FIG. 10A). The differing elution times, and the earlier elution time of the contaminant, indicate this second species was too large to be a product of Fc mispairing. Intact mass spectrometry revealed that Peak 2 was composed of AmAb, with a mass identical to that of Peak 1 (FIGS. 9D and 9E). The identical mass of the two samples suggests that Peak 2 was a self-associating multimer of AmAbs.[OHl] As scFv sequences have been previously reported to be associated with aggregation, an experiment was performed to determine whether the AmAb self-association was scFv-mediated. Support for scFv-mediated AmAb aggregation was revealed from analysis of a subset of AmAb designs, identical to the original panel except for the addition of a stabilizing disulfide engineered into the scFv. The disulfide-containing AmAbs yielded comparable CIEX profiles, but LC-MS analysis showed that Peak 2 contained a mixture of AmAb and covalent dimeric AmAb (FIGS. 10C and 10D). This suggests that covalent dimerization is mediated by an intermol ecul ar disulfide formed via the engineered cysteines, serving as supporting evidence of an intermolecular scFv-mediated interaction. Plotting SEC elution times onto a standard curve (FIG. 10B) supported the hypothesis that the high molecular weight species may be a dimer or possibly a higher-order assembly of AmAbs.

[0112] The soluble antigen binding affinity of the purified proteins from each peak were measured. To measure the binding affinities (KD equilibrium dissociation constant) of the purified antibodies (bispecific or monoclonal) to soluble antigens, the antibodies were first captured onto streptavidin SAX biosensor tips with a biotinylated (1 biotin / molecule), polyclonal capture antibody (Jackson ImmunoResearch, catalog# 109-005-098) and then incubated with a dilution series of each soluble (monovalent) antigen. This assay format was chosen so thebivalent antibodies would be immobilized and on the biosensor tips and tested versus the same serial dilution of each soluble antigen. The measured quantitative KD affinities thus represent monovalent 1: 1 binding interaction and can be directly compared. Experiments were run in a ForteBio Octet HTX instrument using the 96-tip mode with standard 5 Hz data acquisition rate at 27 °C and 1000 RPM. Using Genedata Screener V17 software, raw Octet binding data was processed with installed SPR kinetic curve fitting package and globally fit to a 1:1 binding model to determine the association rate constant (£a) and the dissociation rate constant (kd). The equilibrium dissociation constant (A'D) was then calculated the as a ratio of kd'K. The KD values are average ± standard error from three independent experiments. As shown in FIGS. 11 A-l IB and Tables 2 and 3 below, the material in Peak 1 and Peak 2 show comparable binding affinities in vitro for both Target A and Target B.T BLE 2T BLE 3

[0113] Cell based functional assays were also performed with the purified proteins from each peak. Chemokine release by Target A-expressing cells co-cultured with Target B stableexpressing cells. On Day 1, Target A-expressing cells were harvested, counted and seeded to assay plates. On Day 2, tested molecules were titrated in assay medium from >50nM 1:3 down for 12 points. Tested molecules were transferred to assay plates pre-seeded with Target A-expressing cells and incubated at room temperature for 30 minutes. Cells stably expressing Target B (Target B binding indicates the crosslinking of the assayed molecules) were harvested, counted and resuspended to the desired concentration. Cells stably expressing Target B were then transferred to assay plates. The ratio of Target A- to Target B-expressing cells was set to 1 : 1 based on assay development data. Assay plates were incubated at 37°C with 5% CO2. On Day 3, chemokine levels in culture supernatant were measured. The same experiment was performed using Target A-expressing cells co-cultured with parental cells which do not express Target B, to investigate the crosslinking-independent activities of the tested molecules. Chemokine levels were calculated and plotted against the antibody concentration as log| Agonist | vs. response - Variable slope (four parameters) (GraphPad Prism, Version 7.04). The EC50 value, bottom and top of the curve for each assayed molecule was reported by Prism (GraphPad). The data from these assays is presented in Table 4 below.TABLE 4

[0114] These cell-based functional assays distinguished the species present in each peak. As shown in Table 4, Peak 2 demonstrated non-targeted activity, which was expected only when Target A receptors are crosslinked. This suggests that at least two arms recognizing this receptor are present (i.e., as a dimer of AmAbs). These findings indicate that scFv-containing formats present the additional obstacle of removing high molecular weight multimers. Removal of these aggregates is beneficial, as they pose liabilities toward the therapeutic mechanism-of-action. In addition, purifying therapeutic molecules to high purity is important for accurate assessment of functional activity. Importantly, Peak 1 and Peak 2 maintained their CIEX elution positions after approximately 4 weeks at 4° C, indicating that the AmAb and AmAb-multimer subspecies show a stable conformation and can therefore likely be separated.

[0115] Some AmAbs showed separation of monomeric and multimeric species in CIEX, while others showed severe overlap. Building on the promising results of pBBCS in the removal of homodimers and half-Abs, the establishment of a method capable of predicting separation of this newly identified impurity in silico would increase the applicability of pBBCS. To this end, experiments were performed to gain a deeper understanding of the molecular basis for aggregate behavior during CIEX.

[0116] By selecting for high AConductivity, it is possible to address the impurities resulting from the imbalanced chain ratios (i.e., homodimeric or single-arm AmAb). However, the AmAb multimer described in the previous section maintains the 1: 1 chain ratio of the AmAb monomer, and it was therefore reasoned that a high AConductivity alone may not be sufficient to predict the separation of aggregates. Instead, it was hypothesized that an aggregate would display increased electrostatic surface patches per particle compared to the AmAb monomer during CIEX, and possibly the presence of larger intermolecular electrostatic patches formed at the interface of two or more AmAbs (FIG. 12A). Thus, an AmAb with a high affinity for CIEX resin would generate multimers with a proportionally higher affinity for the column matrix, thereby increasing the species separation. Indeed, comparing main peak conductivities with the degree of separation between the AmAb and the aggregate peak (FIG. 12B) showed that greater AmAb conductivity was correlated with greater separation from the aggregate. Because AmAb conductivity can be accurately predicted from the building blocks conductivities (FIG. 7), pBBCS can also inform the design of bispecifics with predictable separation from aggregate species. Analysis of the data showed that an AmAb conductivity of ~11 mS / cm or higher enabled effective separation of aggregates (FIG. 12B). In addition, it was observed that improved separation of aggregates correlated with elution conductivity of the Target B mAh (i.e., the scFv arm of the AmAb), suggesting that the scFv may play a dominant role in driving separation of AmAb monomers and multimers (FIG. 12C). FIGS. 14A and 14B show examples in which two AmAbs, containing identical Fab arm sequences but different scFv arm sequences, generated aggregates with differing degrees of separation after CIEX. Pairing Target Al building block with Target Bl building block (AlxBl) produced an AmAb with an elution conductivity of 5.6 mS / cm, whereas pairing the same Target Al with Target B2 (AlxB2) generated an AmAb with an elution conductivity of 11.8 mS / cm. Analytical SEC of the corresponding Protein A eluates shows the presence of aggregation in both molecules (FIGS. 13 A and 13B). However, the aggregate from the AlxBl AmAb fails to separate during SEC, as evidenced by the HMWspecies detected in each fraction by SEC (FIG. 14A, plotted as grey dots). Conversely, the aggregate from the AlxB2 AmAb is effectively separated, possibly due to increased avidity for the CIEX resin (FIG. 14B, plotted as grey dots). Thus, a building block can be “rescued’' through pairing with an appropriate partner if the resultant molecule’s conductivity is sufficiently high. This result highlights the importance of selecting compatible building blocks, as the performance of a bispecific is not solely driven by intrinsic properties of each individual module, but also by the unique properties resulting from their pairing.

[0117] Given the observations on scFv-mediated aggregation discussed above, it was plausible that some scFvs may be more prone to aggregation than others. Thus, screening scFvs for aggregation tendency would provide a valuable metric for deselecting building blocks. Given the combinatorial nature of the tested AmAb panel, this was an ideal opportunity to characterize unique scFv sequences in bispecifics with identical backgrounds. For example, it was observed that drastically different proportions of aggregate when a single Fab sequence (Building Block Al) is paired with four unique scFv sequences (Building Block B3, B4, B5, or B6) (see FIGS. 15A-15D). Therefore, although some aggregation may be an inevitable by-product for scFv- containing bispecifics, it can be minimized through rational selection of building blocks. Indeed, quantifying aggregate abundance according to Building Block B supports the hypothesis that some scFv sequences have a higher propensity toward aggregation (FIG. 15E). For the AmAbs with observable separation in CIEX, it was observed that the proportion of aggregate peak in SEC correlated with the abundance of aggregate peak in CIEX (FIGS. 15A-15D and 15F). This is likely because the correlation was diminished by the partial overlap in CIEX, and the heterogeneous SEC profiles of the aggregate peak.

Claims

CLAIMS1. A method of screening potential components for multicomponent polypeptides, wherein the multicomponent polypeptides each comprise at least one building block A and at least one building block B, and wherein the screening comprises: a. identifying and individually purifying a group of molecules (pool A) wherein each molecule comprises one or more copies of a candidate building block A; b. identifying and individually purifying a group of molecules (pool B) wherein each molecule comprises a one or more copies of a candidate building block B; c. determining the conductivity of each pool A molecule and of each pool B molecule; d. determining the difference in conductivity (ACond) between each molecule of pool A and each molecule of pool B; and e. selecting a set of candidate multicomponent polypeptides that comprise a building block A-building block B pair, wherein the pool A and pool B molecules comprising the building blocks in the selected multicomponent peptides have a larger ACond than the ACond between the pool A and pool B molecules comprising the building block pairs in the non-selected multicomponent polypeptides.

2. The method of claim 1, wherein the conduct! vify of each pool A molecule and of each pool B molecule is determined using cation-exchange chromatography (CIEX), anion- exchange chromatography (AIEX), hydrophobic interaction chromatography, or reversed phase chromatography.

3. The method of claim 2, wherein the conductivity of each pool A molecule and of each pool B molecule is determined using CIEX.

4. The method of claim 3, wherein the selected multicomponent peptides have a ACond between the pool A and pool B molecules greater than 5 mS / cm.

5. The method of claim 4, wherein the selected multicomponent peptides have a ACond between the pool A and pool B molecules greater than 7 mS / cm.

6. The method of claim 1 , wherein the selected building block A came from a pool A molecule measured to have a conductivity within the top 25% of tested pool A molecules, and wherein the selected building block B came from a pool B molecule measured to have a conductivity within the bottom 25% of tested pool B molecules.

7. The method of claim 6, wherein the selected building block A came from a pool A molecule measured to have a conductivity within the top 10% of tested pool A molecules, and wherein the selected building block B came from a pool B molecule measured to have a conductivity- within the bottom 10% of tested pool B molecules.

8. The method of any of claims 1-7, wherein each pool A molecule and each pool B molecule comprises an immunoglobulin.

9. The method of claim 8, wherein each pool A molecule and each pool B molecule comprises an antibody.

10. The method of any of claims 1-9, wherein building block A comprises all or part of an antibody heavy chain.

11. The method of claim 10, wherein building block A comprises an antibody arm.

12. The method of any of claims 1-11, wherein building block B comprises all or part of an antibody heavy chain.

13. The method of claim 12, wherein building block B is an antibody arm.

14. The method of any of claims 9-13, wherein the multicomponent polypeptide is a bispecific antibody.

15. The method of any of claims 9-14, wherein building block A comprises an antibody arm, and wherein building block B comprises an Fc-linked scFv.

16. The method of any of claims 8-15, wherein the multicomponent polypeptide is an antibody drug conjugate or antibody fusion.

17. The method of claim 16, wherein the multicomponent polypeptide comprises an antibody fusion, and wherein the antibody is fused to a cytokine.

18. The method of any of claims 1 -17, wherein the method further comprises estimating the absolute conductivity of the selected multicomponent polypeptides by summing the conductivity' of the pool A molecule and pool B molecule that supply the building blocks for each multicomponent polypeptide, and keeping in the selected set only those multicomponent polypeptides within the upper 50% of the estimated summed conductivities.

19. The method of claim 18, wherein the method comprises keeping only those multicomponent polypeptides with an absolute conductivity greater than 11 mS / cm.

20. The method of any of claims 1-19, wherein the method further comprises comparing the CIEX elution profile of each selected multicomponent polypeptide with the CIEX elution profiles of the pool A molecule and the pool B molecule from which its component parts were derived, and keeping only those multicomponent polypeptides whose principal elution peak has less than 25% overlap in peak area with the principal elution peak of either the pool A molecule or the pool B molecule.

21. The method of claim 20, where there is less than 10% overlap in peak area between the multicomponent peptide's principal peak and the principal peak of either the pool A molecule or the pool B molecule.

22. A method of constructing a multicomponent immunoglobulin molecule comprising the steps of: a. measuring the surface charge of each member of a pool A of immunoglobulins and measuring the surface charge of each member of a pool B of immunoglobulins; b. selecting an immunoglobulin from pool A (immunoglobulin A) and an immunoglobulin from pool B (immunoglobulin B) that have a difference in conductivity (ACond) greater than 5 mS / cm, as measured by CIEX; and c. combining a building block A that comprises part or all of immunoglobulin A with a building block B that comprises part or all of immunoglobulin B in the multicomponent immunoglobulin molecule.

23. The method of claim 22. wherein the selected immunoglobulin A and immunoglobulin B have a ACond greater than 7 mS / cm, as measured by CIEX.

24. The method of claim 22 or claim 23, wherein each pool A molecule and each pool B molecule comprises an immunoglobulin.

25. The method of claim 24, wherein each pool A molecule and each pool B molecule comprises an antibody.

26. The method of any of claims 22-25, wherein building block A comprises all or part of an antibody heavy chain.

27. The method of claim 26, wherein building block A comprises an antibody arm.

28. The method of any of claims 22-27, wherein building block B comprises all or part of an antibody heavy chain.

29. The method of claim 28, wherein building block B is an antibody arm.

30. The method of any of claims 25-29, wherein the multicomponent polypeptide is a bispecific antibody.

31. The method of any of claims 25-30, wherein building block A comprises an antibody arm, and wherein building block B comprises an Fc-linked scFv.

32. The method of any of claims 24-31, wherein the multicomponent polypeptide is an antibody drug conjugate or antibody fusion.

33. The method of claim 32, wherein the multicomponent polypeptide comprises an antibody fusion, and wherein the antibody is fused to a cytokine.

34. The method of any of claims 22-33, wherein the method further comprises, prior to combining the building blocks, estimating the absolute conductivity of multicomponent polypeptides comprising the selected components by summing the conductivity of the pool A molecule and pool B molecule that supply the building blocks for each multicomponent polypeptide, and constructing only those multicomponent polypeptides within the upper 50% of the estimated summed conductivities.

35. The method of claim 34, wherein the method comprises constructing only those multicomponent polypeptides with an absolute conductivity greater than 11 mS / cm.

Images