In the present study, both of these challenges were addressed through the development of an innovative mixed mode resin with 2-amino-4methylpentanoic acid ligands that combines weak cation exchange with moderate hydrophobic interactions.
Journal of Chromatography A 1687 (2023) 463696 Contents lists available at ScienceDirect Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma Purification of hydrophobic complex antibody formats using a moderately hydrophobic mixed mode cation exchange resin Wolfgang Koehnlein a , Annika Holzgreve b , Klaus Schwendner a , Romas Skudas b , Florian Schelter a,∗ a b Roche Diagnostics GmbH, Nonnenwald 2, 82377 Penzberg, Germany Merck KGaA, Frankfurter Str 250, 64293 Darmstadt, Germany a r t i c l e i n f o Article history: Received 19 September 2022 Revised 21 November 2022 Accepted 30 November 2022 Available online December 2022 Keywords: Complex antibody formats Downstream purification Hydrophobicity Mixed mode chromatography Product-related impurities a b s t r a c t Immunoglobulins of complex formats possess great potential for increased biopharmaceutical efficacy However, challenges arise during their purification as the removal of numerous product-related impurities typically requires several expensive chromatographic steps Additionally, many complex antibody formats have a high hydrophobicity which impairs the use of conventional mixed mode chromatography In the present study, both of these challenges were addressed through the development of an innovative mixed mode resin with 2-amino-4methylpentanoic acid ligands that combines weak cation exchange with moderate hydrophobic interactions Supported by high throughput partition coefficient screens for identification of preferable pH and salt concentration ranges in bind and elute mode, this mixed mode resin successfully demonstrated efficient impurity separation from an extremely hydrophobic bispecific antibody with a single unit operation High purity (>97%) was obtained as a result of significant reduction of product-related impurities as well as process-related host cell proteins (>3 log scale), while maintaining satisfactory recovery (70%) This also supports that highly hydrophobic antibody formats can be efficiently purified using a resin with moderate hydrophobic characteristics Studies involving additional antibodies possessing different formats and a wide range of hydrophobicity confirmed the broad applicability of the new resin In view of its high selectivity and robust operating ranges, as well as the elimination of the need for an additional column step, the novel resin enables simplified downstream processing and economic manufacturing of complex antibody formats © 2022 The Authors Published by Elsevier B.V This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Introduction In search of continual enhancement of recombinant monoclonal antibodies (mAbs), development focuses on increasingly complex molecular constructs and mechanisms of action In this regard, bispecific antibodies (bsAbs) or even multispecific formats have emerged as an important class of biopharmaceuticals [1] Their particular property of targeting more than one antigen raises the expectations of improved drug efficacy in comparison to conventional monospecific mAbs [2] A great potential also arises from the generation of a broad spectrum of distinct complex antibody formats that may facilitate the fit for specific applications Various engineering strategies using full-length antibodies, specific fragment or special modules create combinatorial diversity and offer ∗ Corresponding author E-mail address: Florian.Schelter@roche.com (F Schelter) more flexibility regarding valency and specificity or even enable novel functionalities [1,3,4] Among the numerous approaches described for the formation of bsAbs [5], the CrossMAb technology is a pioneer that performs exchange and crossover of antibody domains to enable correct light chain association with the respective heavy chain counterparts [6,7] This technology offers stable constructs suitable for drug development and production rendering versatile CrossMAb formats available today [8] The opportunities for enhanced drug efficacy offered by the various antibody formats are, however, associated with increased challenges for product manufacturing at high yield and purity [8,9] In addition to common process-related impurities, like host cell proteins (HCP) and DNA, complex antibody formats tend to form product-related impurities at higher number and level than standard mAbs (see Fig 1) The more different components are required for correct antibody formation, the more single byproducts, mis-paired forms, or aggregates may occur that appear as low molecular weight (LMW) and high molecular weight (HMW) https://doi.org/10.1016/j.chroma.2022.463696 0021-9673/© 2022 The Authors Published by Elsevier B.V This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) W Koehnlein, A Holzgreve, K Schwendner et al Journal of Chromatography A 1687 (2023) 463696 Fig Characterization of the antibodies used pI: Isoelectric point; HCP: host cell protein; HMW: high molecular weight; LMW: low molecular weight impurities in the product [9] Hence, type and amount of impurities are antibody-specific and usually depend on the format Although, respective platform technologies have been developed to facilitate the formation of the desired bsAb product, such as the knob in hole (KIH) technology in case of CrossMAb [5,10,11], the processing of complex antibody formats is still often difficult, elaborate and limited regarding yield and purification success Generally, purification of complex formats may benefit from special adaptations of the chromatography-based downstream processing established for standard mAbs [12] The procedure typically comprises up to four sequential steps i.e., affinity chromatography (mostly Protein A-based), ion exchange chromatography (IEX) either with anions (AEX) or cations (CEX), hydrophobic interaction chromatography (HIC), and the so-called mixed mode or multimodal chromatography (MMC) primarily combining electrostatic and hydrophobic interactions [1,9] Respective modifications mainly consider improved separation of product-related impurities from the antibody monomer, as was also shown for the use of MMC to achieve efficient bsAb purification [13] This type of chromatography depends on the interplay of special resins with the fine-tuned conditions of binding and elution generated by ionic and hydrophobic interactions together with the effects of hydrogen bonding [14] Thereby, pH primarily controls ionic interactions and binding capacity in favor of high recovery, whereas conductivity (i.e salt concentration) rather determines hydrophobic interactions supporting separation capability rendering these two parameters essential for optimization [15,16] In addition, improvement in MMC separation performance and selectivity may emerge from the development of new functional groups/ligands and the use of additives [17] A further challenge for the purification of antibodies with complex formats may arise from particular physical properties, such as high hydrophobicity Several bsAb with complex formats investigated in our lab display high to extremely high hydrophobicity that counteracts successful application of available MMC The alternatively used two-step process of CEX and HIC, however, increases effort, time, and costs of the manufacturing process To establish an efficient MMC step for highly hydrophobic antibodies, a new resin with refined characteristics is desirable providing adequate moderate affinity for a bind and elute mode, while maintaining favorable separation properties for impurity elimination Moreover, a high-quality resin should offer a broad window of operation resulting from distinct graduation of binding characteristics in the range of standard pH and salt conditions to facilitate the development of robust processes Based on these demands, the present study investigated the usability of the newly developed mixed mode cation exchange chromatography (MMCEX) resin, Eshmuno®CMX, for the purification of antibodies with complex formats and high hydrophobicity In this study, partition coefficient (Kp) screens as a high throughput tool [18] has been used to analyze the binding behavior of antibodies as a function of pH and salt conditions Initial experiments indicated that functional groups of desired moderate hydrophobicity for the resin might be found among carboxylic acids Based on these experiments, 2-Amino-4-methylpentanoic acid was selected The most hydrophobic antibody representing a trivalent bispecific (2+1) format [7] was then used to evaluate the operational window of the novel resin for high product yield and efficient removal of product-related as well as process-related impurities A refined bind and elute mode was verified by column runs confirming efficient separation of LMW, HMW and HCP impurities The application space of Eshmuno® CMX was further examined with two other antibodies covering a wider hydrophobicity range and different impurity profiles In summary, this study shows that highly hydrophobic complex antibody formats can be efficiently purified using a MMC with moderate hydrophobic characteristics Materials and methods 2.1 Antibody material Humanized IgG1 antibodies with standard and complex formats were produced in Chinese hamster ovary (CHO) cells and characterized by isoelectric point and impurity profiles (methods W Koehnlein, A Holzgreve, K Schwendner et al Journal of Chromatography A 1687 (2023) 463696 Fig Determination of relative hydrophobicity of the antibodies using commercially available pharmaceutical antibodies for linear regression analysis described in the following sub-sections) as shown in Fig In addition, the relative hydrophobicity of the antibodies was determined by a standard HIC method [19] using a TSKgel® Ether-5PW HPLC Column, 10 μm (Tosoh Bioscience) After conditioning, mAb amounts of 20 μg (sample volume 20 μL) were applied to a gradient run (flow rate: 0.8 mL/min) consisting of eluent A (25 mM sodium phosphate buffer with 1.5 M ammonium sulfate, pH 7.0) and eluent B (25 mM sodium phosphate buffer, pH 7.0) starting after equilibration with 100% eluent A and reaching 100% eluent B after 60 The retention times of eluted mAbs were determined at peak maximum The values of commercially available pharmaceutical antibodies together with the molecule with lowest retention time (represented by 0) and the molecule with highest retention time (represented by 1) served for regression analysis to determine the relative hydrophobicity of the antibodies under study (Figs and 2) The eluate obtained from affinity chromatography columns (Protein A) was conditioned to pH and after depth and sterile filtration (to remove precipitates) was used as starting material for resin investigation rial was combined with the mixture (Eshmuno® particles) and polymerization is started by adding Cerium (IV) nitrate After the grafting reaction, non-reacted components and starter are removed by extensive washing using acidic, basic and solvent mixtures The following chromatography resins were used for comparative studies: CaptoTM MMC ImpRes, (Cytiva); NuviaTM cPrimeTM (Bio-Rad Laboratories), Eshmuno® HCX (Merck KGaA, Darmstadt, Germany) 2.3 Execution of partition coefficient (Kp) screens Kp screens served as a high-throughput screening tool to evaluate the binding behavior of antibodies and impurities as a function of pH and salt concentration [18] The filter plates (96 multiwell plates, AcroprepAdvTM , 0.45μm, Polypropylen, Pall Corporation) containing 50 μL of the respective resin (50 % slurry in water (v/v)) per well and 300 μL water were prepared by a Microlab STARlet robot (Hamilton) equipped with a shaker A total of 48 defined buffer conditions (six pH values x eight salt concentrations) were investigated The Tecan Freedom EVO® 200 liquid handling station (software Evoware®, Tecan Deutschland GmbH) created a stepwise increase in pH and sodium sulfate molarity in the wells by combining buffer stock solutions of each pH with increasing volumes of the respective buffer stock solution containing 1.4 M sodium sulfate In addition, an equilibration buffer plate and a load plate containing the protein sample were produced The system was equipped with one liquid handling arm (processing volumes of 10 μL to 10 0 μL), one excentric gripper, a Te-ShakeTM , a Te-StackTM for storage of microplates, a Te-SlideTM for plate transport, an Infinite® M200 plate reader for measuring protein concentrations (software Magellan), and an integrated centrifuge (Rotanta 46RSC, Hettich) 2.2 Chromatography resins Eshmuno® CMX (Merck KGaA, Darmstadt, Germany prepared by graft polymerization of 2-Amino-4-methylpentanoic acid was used for the purification experiments with hydrophobic antibodies Additionally, two prototypes have been prepared by graft polymerization of a carboxyl group or 2-aminopropanoic acid The synthesis of the prototypes was performed by dissolving the respective amino acid in deionized water Then an acrylic compound like acrylic acid chloride or acrylic acid was added to form a monomer which is necessary for the grafting process Afterwards the functional group -OH containing base mate3 W Koehnlein, A Holzgreve, K Schwendner et al Journal of Chromatography A 1687 (2023) 463696 Initial Kp screens evaluating the MMC resins CaptoTM MMC ImpRes, NuviaTM cPrimeTM , and Eshmuno® HCX as well as the prototypes of Eshmuno® CMX with different functional groups/ligands were performed in a buffer of acetate, MES, HEPES, or TAPS 20 mM Each buffer was used in its optimum buffering range and titrated with NaOH to adjust the following pH values: 5.0, 5.7, 6.4, 7.1, 7.8, and 8.5 Additionally, buffers were combined with the following sodium sulfate concentrations: 20, 160, 300, 440, 580, 720, 860, and 10 0 mM The second series of Kp screens evaluating Eshmuno® CMX with the three hydrophobic antibodies was executed with 25 mM Tris/acetate buffer adjusted to the pH values: 4.0, 5.0, 6.0, 7.0, 8.0, and 9.0 combined with the following sodium sulfate concentrations: 25, 75, 150, 250, 350, 500, 650, and 800 mM Antibody solutions used to prepare the load plate were additionally subjected to concentration and diafiltration using 30 kDa TFF membrane Biomax® 30kDa, Pellicon® (Merck KGaA, Darmstadt, Germany) in 25 mM Tris/acetate, pH 6.5 to enable subsequent adjustment of load density and pH conditions in the plates In the first series of Kp screens the load density accounted for g/L resin, in the second series a load density of 30 g/L resin was applied Execution of Kp screens on the filter plates consisted of the following steps (each with 300 μL pipetting volume): two times of equilibration, one loading step followed by stripping and regeneration After each step, centrifugation collected the flow-through in a plate and measurement of protein concentration and further analytics were performed The load (300 μL) derived from the load plate was incubated for 30 on a shaker to reach an equilibrium between bound and unbound state, before the flow-through was collected in a flow-through plate The protein concentrations of the load plate and flow-through plate were determined by an UV absorbance based method using the Infinite® M200 plate reader The binding capacity of the resin for the protein was visualized by contour plots displaying the portion of protein flow-through Additional analysis was performed by SE-HPLC (Section 2.5) to determine LMW, main peak monomer, and HMW levels displayed as flow-through contour plots (Tosoh Bioscience) using an UV based detector (at 280 nm) on the UltiMateTM 30 0 RSLC instrument (Thermo Fisher Scientific) Analysis was performed with a mobile phase of 0.20 M potassium phosphate and 0.25 M potassium chloride, pH 6.2 and a flow rate of 0.3 mL/min at room temperature (19-26°C) The injected sample volume was 10 μL (5 g/L) resulting in 50 μg per load Samples of lower concentration were used undiluted 2.6 Analytical capillary electrophoresis (CE-SDS) To examine obtained samples under denatured conditions, the capillary electrophoresis system Caliper Labchip GXII (PerkinElmer) was used to execute conventional SDS-PAGE in a chip format The included single components were separated according to molecular weight by disconnecting all non-covalently bound molecules from each other Analysis was performed using the HT Protein Express Reagent Kit (PerkinElmer) with a sample volume of μL (1 g/L) under non-reducing conditions (pH 8.7) Samples of lower concentration were used undiluted Samples were labeled by a fluorescent dye to enable laser detection 2.7 HCP electrochemiluminescent immunoassay (ECLIA) The residual Chinese Hamster Ovarian HCP (Host Cell Protein) content in the samples was determined by the ECLIA-HCP assay on a cobas ®e 801 immunoassay analyzer (Roche Diagnostics) The assay was based on a sandwich principle by using biotinylated polyclonal CHO HCP-specific antibodies to bind the target This complex was fixed to the solid phase by streptavidin-coated microparticles Thereafter, a ternary sandwich complex was formed by addition of a second polyclonal CHO HCP-specific antibody labeled with a ruthenium complex that allowed chemiluminescence detection The assay displayed a detection limit for CHO HCP of ng/mL, a quantification limit of 7.5 ng/mL and a linear measuring range up to 10 0 ng/mL 2.8 DNA measurement DNA originating from host cells was measured for the column runs using an automated quantitative PCR method performed in the 96-well format by the FLOW Flex system (Roche Diagnostics) consisting of three modules: The FLOW PCR SETUP Instrument used as FLOW primary sample handling system, the MagNa Pure 96 system for automated DNA extraction using the MagNA Pure LC Total Nucleic Acid Isolation Kit – High Performance, and the LightCycler® 480 system for DNA amplification and quantification using the Residual DNA CHO Kit specific for highly conserved CHO DNA regions The method had a detection limit of 0.4 pg/mL and a quantification limit of pg/mL with a linear measuring range up to 40 0 pg/mL 2.4 Eshmuno® CMX column runs Column runs for MAB A using a bind and elute mode were performed on columns with a bed height of 20 cm and a diameter of cm The residence time was corresponding to a linear flowrate of 300 cm/hour The dynamic binding capacity at pH 5.0 and 10% breakthrough resulted in 64 g/L Equilibration was done with column volumes (CV) of equilibration buffer (50 mM sodium acetate, 25 mM sodium sulphate, pH 5.0) followed by load application (density 10 g/L) The following washing step (wash 1: 100 mM MES, pH 5.5, 3.5 CV) served to override the buffer capacity of the resin With the next washing step, the pH was shifted to basic conditions (wash 2: 15 mM Tris/acetate, pH 9.0, CV) and the last washing step served to prepare elution conditions (wash 3: 50 mM sodium citrate, 100 mM sodium sulfate, pH 6.1, CV) The product was eluted with elution buffer (100 mM sodium succinate, 300 mM sodium sulfate, 10 CV) Two runs with different elution pH levels were performed (run at pH 6.0, run at pH 6.2) Finally, the column was cleaned in place (CIP) (100 mM arginine, pH 10.9, CV followed by M sodium hydroxide, 16 CV) Results and discussion 3.1 Evaluation of standard MMC for the purification of a highly hydrophobic molecule Initial experiments evaluated the impact of antibody hydrophobicity on the usability of available MMC resins for the purification process Kp screens performed with the eluate of Protein A chromatography over the range of standard pH (5-8.5) and salt concentrations (0-10 0 mM) were used to investigate the binding properties of CaptoTM MMC ImpRes, a common available MMC resin, for three antibodies with increasing relative hydrophobicity determined by standard HIC (Fig 2) Obtained contour plots showed a satisfactory window for binding and flow-through reflecting suitable elution conditions for tocilizumab, the antibody with low hydrophobicity (0.06) (Fig 3A) MAB C, the antibody with moderate 2.5 Size exclusion high performance liquid chromatography (SE-HPLC) The separation of HMW and LMW from the antibody monomers in the samples was analyzed by SE-HPLC under native conditions using a TSKgel® UP-SW30 0 column, μm, 4.6 mm x 300 mm W Koehnlein, A Holzgreve, K Schwendner et al Journal of Chromatography A 1687 (2023) 463696 Fig The impact of antibody hydrophobicity on the usability of different standard mixed mode resins A) Kp screen contour plots of three antibodies with increasing relative hydrophobicity (0.06, 0.33, 0.74) using Capto MMC Impress B) Kp screen contour plots for the highly hydrophobic MAB A performed on three different mixed mode resins hydrophobicity (0.33) displayed a smaller window with a narrowed transition zone in the low salt range, but still appeared acceptable In contrast, the highly hydrophobic MAB A (0.74) showed efficient binding, however, completely lacked an elution window under the pH and salt conditions used Two further tested MMC resins, NuviaTM cPrimeTM and Eshmuno® HCX, provided a similar picture regarding a strong binding of the highly hydrophobic antibody without offering suitable conditions for elution (Fig 3B) Only for Eshmuno® HCX, a small elution window at high pH and moderate salt concentration was indicated, however, at unsatisfactory recovery level Overall, the tested MMC resins were not applicable for the processing of the hydrophobic bsAb under study, as they impaired efficient elution and recovery This outcome may be explained by the fact that all three standard MMC resins use phenyl groups as hydrophobic functional groups/ligands that mediate strong hydrophobic interaction acids containing a short hydrophobic linear or branched hydrocarbon part were found suitable as possible ligands for Eshmuno® CMX This new MMC resin material is based on weak cation exchange groups and displays a pronounced three-dimensional structure in the form of tentacles that offers extended steric access for binding and is, therefore, supposed to favor the separation of target molecules with moderate or minor differences (Fig 4A) The goal was to complement these binding properties with an appropriate component providing the required level of hydrophobic interaction To this end, three prototypes of the new resin containing functional groups/ligands with different hydrophobicity were tested in Kp screens with the highly hydrophobic MAB A A carboxyl group, 2-aminopropanoic acid, and 2-Amino-4methylpentanoic acid were selected as functional groups/ligands based on increasing hydrocarbon parts to build resins with increasing hydrophobicity (Fig 4B) The contour plots obtained for the carboxy group and 2-aminopropanoic acid were quite similar and showed wide areas with low or no binding capacity, indicating a low modulatory impact especially of pH (Fig 3B) These binding characteristics were not suitable to generate adequate operational windows In contrast, 2-Amino-4-methylpentanoic acid, the most hydrophobic functional group/ligand examined displaying a branched hydrocarbon part, showed a pronounced effect of pH resulting in a more graduated contour plot with separated areas of efficient binding and elution over a wider transition zone A high binding rate appeared at low salt content over the range > pH 4.5, while increasing salt concentration decreased binding affinity and yielded efficient elution in the range of moderate salt and neutral to weak acidic pH conditions These results indicated a suitable operational window for a bind and elute mode for MAB A along with favorable preconditions for the elimination of the included impurities Thus, 2-Amino-4-methylpentanoic acid could be identified as an appropriate functional group/ligand to provide moderate hy- 3.2 Development of a novel MMC resin with moderate hydrophobic interaction The impurity profile of MAB A, containing in addition to high amounts of HMW and HCP also a significant LMW fraction (Fig 1), required processing in a bind and elute mode Therefore, the development of a new resin suitable for highly hydrophobic molecules aimed to provide a broad operating window that clearly identifies proper conditions for binding and elution This is facilitated by a wide transition zone with gradual differences from binding to nonbinding conditions for the antibody monomer across standard pH and salt concentrations In this regard, the findings with available MMC resins indicated that the hydrophobic interaction component had to be weakened towards moderate binding affinity In search of structures providing desired moderate hydrophobicity, carboxylic W Koehnlein, A Holzgreve, K Schwendner et al Journal of Chromatography A 1687 (2023) 463696 Fig Development of a novel mixed mode resin for the purification of hydrophobic molecules A) Structure of the three prototypes with tentacle chemistry B) Kp screen contour plots for three prototypes of Eshmuno® CMX with different functional groups/ligands of increasing hydrophobicity using the highly hydrophobic MAB A drophobic interactions In combination with weak cation exchange groups, the novel Eshmuno® CMX resin showed promising binding characteristics for the design of robust purification processes for highly hydrophobic antibodies with complex formats pH range with slightly raised ionic strength was found suitable to prepare elution while simultaneously removing further LMW An appropriate elution window could then be reached with increasing salt concentrations exceeding 250 mM At a closer view, the data indicated a balance between purity and recovery in this zone that was more dependent on pH than on salt conditions A more acidic pH appeared to support higher purity due to improved HCP reduction at the cost of lower product recovery, while a shift towards neutral pH pointed to increased recovery associated with slightly increased HCP content Thus, the contour plots served as an overall survey to assess how individual performance parameters can be modulated to obtain desired product yield and quality Although Kp screens are operated as a static process, they provide useful information for the design of a dynamic column process 3.3 Evaluation of Eshmuno® CMX for the purification of a highly hydrophobic bsAb with a complex format In the next step, the space for suitable modes of purification for the MAB A should be further specified, also in regard to the separation capacity for the major impurities HMW, LMW, and HCP that were present at relatively high levels (Fig 1) To this end, a Kp screen was repeated focusing on salt concentrations of 0800 mM over the range of pH 4-9 (Fig 5) that largely reproduced the outcome for protein binding regarding the initial experiments with 2-Amino-4-methylpentanoic acid as functional group/ligand In addition, the screen enclosed contour plots reflecting the binding behavior of the main peak and of the individual impurity types Obtained results revealed a suitable window for HMW reduction around pH and low salt concentration, for LMW reduction at low salt and basic conditions around pH 8, and for HCP reduction at basic pH values largely independent from the salt concentration Importantly, these patterns did not show significant impurity flowthrough in the window identified for efficient main peak elution, thus indicating a high selectivity of the new resin The combination of the contour plots provided the basis to identify the ranges of adequate conditions for each step of the purification process Thus, optimal conditions for antibody binding were found at acidic conditions around pH and low ionic strength, at which part of HMW could be simultaneously removed As a next step, a shift to basic pH 8-9 at the same conductivity appeared appropriate to remove LMW and HCP fractions, while the antibody was still strongly bound The weak acid to neutral 3.4 Bind and elute column purification process for a highly hydrophobic bsAb using Eshmuno® CMX Based on the Kp screen results, a procedure for an Eshmuno® CMX column run was tested selecting the following bind and elute mode The resin was equilibrated at pH 5.0 and low ionic strength Loading under these conditions should result in efficient capturing of the antibody Thereafter, incubation at pH 5.5 intended to overcome the buffer capacity of the resin Following a shift to pH 9.0, a second washing step aimed to eliminate LMW and HCP from the still tightly bound antibody A third wash step at pH 6.1 and slightly increased conductivity (100 mM sodium sulfate) served to prepare elution of the product that was done with 300 mM salt Due to the assumed impact of pH, runs at two different elution pH values, pH 6.0 (run 1) and pH 6.2 (run 2), were performed The protocol resulted in the elution profiles shown in Fig 6A The respective purification process is documented by analytical SE-HPLC W Koehnlein, A Holzgreve, K Schwendner et al Journal of Chromatography A 1687 (2023) 463696 Fig Kp screen contour plots of the highly hydrophobic MAB A on Eshmuno® CMX displaying the flow-through of total protein, host cell protein (HCP), high molecular weight (HMW) impurities, main peak, and low molecular weight (LMW) impurities Fig Bind and elute purification process of the highly hydrophobic MAB A on Eshmuno® CMX A and C) Elution profiles of run (elution pH 6.0) and run (elution pH 6.2) B and D) overlays of the analytical size exclusion chromatography performed for run and run Colour coding: black = load, blue = wash 1, pink = wash 2, green = eluate, brown = post eluate W Koehnlein, A Holzgreve, K Schwendner et al Journal of Chromatography A 1687 (2023) 463696 Fig CE-SDS data obtained for the bind and elute purification process of the highly hydrophobic MAB A in two runs using non-reduced conditions Colour coding: black = load, blue = wash 1, pink = wash 2, brown = eluate, green = post-eluate Table Removal of product-related and process-related impurities by Eshmuno® CMX column runs Step definition IgG (g/L) Native purity by SE-HPLC (Sum in %) HMW Main peak LMW 15.8 5.9 6.4 0.6 10.5 1.2 2.3 14.2 83.6 98.2 97.2 84.8 5.9 0.6 0.5 1.1 Load Run 1: eluate elution pH 6.0 Run 2: eluate elution pH 6.2 Run 2: fraction post end pooling criteria HCP (ng/mg) DNA (pg/mg) Step yield (%) Purity by CE-SDS (Sum in %) 12353 2.3 3.3 31 5759 1367 1375 4168 64.9 71.2 7.5 HMW Main peak LMW 0.07 pH 5) This shortcoming may be counterbalanced by using a successive standard flow-through AEX polishing at high capacity and product recovery (around 95%) [20], that may combine DNA removal with efficient virus reduction in one step The third molecule investigated was MAB C, a glycoengineered mAb with a standard format and moderate relative hydrophobicity of 0.33 displaying a more usual impurity profile at a lower level (Fig 1) In a similar manner as found for the two bsAbs, efficient protein binding was detected in the low salt range and elution seemed feasible at conditions of higher salt and pH or above (Fig 9) HMW and LMW reduction occurred under nearly the same conditions, mainly at acidic pH and at low salt up to pH 8, where the antibody was tightly bound Therefore, loading at pH and subsequent washing at a pH seemed suitable to remove significant fractions of HMW and LMW together with HCP Favorable elution conditions appeared around 400 mM salt and pH Again, a lower HCP content could be expected at lower pH values The results showed that the binding properties of this mAb on Eshmuno® CMX had changed in comparison to CaptoTM MMC ImpRes (Fig 3) The reduced hydrophobic interaction of Eshmuno® CMX resulted in a consistent and wide transition zone with increasing salt concentrations and, thus, may constitute a practical alternative to resins with a stronger hydrophobic component as well as to the application of pure cation exchange resins In view of the presented examples, the application of Eshmuno® CMX appeared feasible in a similar bind and elute mode offering high selectivity for efficient purification of the two complex bispecific antibody formats as well as of the standard mAb In all three cases, the binding window was found in the range of pH without salt addition, where simultaneously HMW fractions could be washed out A shift to highly basic conditions served for impurity removal The optimal pH level for this step depended on the pI of the antibody, which determined the range in which the antibody was still strongly bound due to cationic interaction Thus, for MAB B having a pI of 8.0 the wash step was suitable at pH 7.5, whereas for MAB A with a pI of 9.4 the wash step could be ex- 3.5 General applicability of Eshmuno® CMX for the purification of hydrophobic molecules The applicability of Eshmuno® CMX for the purification of hydrophobic antibodies was further evaluated by performing Kp screens with other molecules, such as MAB B, a bsAb of a symmetric 1+1 format with a high relative hydrophobicity of 0.58 Among the LMW components, the impurity profile included a ¾ antibody fragment (Fig 1) Again, obtained contour plots revealed good loading conditions for protein binding in the range of pH and low salt, while favorable elution conditions could be deduced in the range of pH 5-6 and elevated salt concentrations higher than 350 mM (Fig 8) HMW behavior was similar as observed for MAB A, where flow-through appeared under acidic conditions The major difference appeared for LMW flow-through, which was only low at basic conditions but increased around pH and 150 mM salt Therefore, loading at pH and subsequent washing appeared suitable to partially remove HMW and LMW fractions As strong antibody binding occurred up to pH 7.5, a washing step at this pH level was useful to reduce further LMW in addition to a part of the HCP fraction In terms of the elution window in the range of pH to pH 6, more acidic conditions seem to provide the possibility for higher product purity due to lower HCP content, however, most probably at the cost of lower recovery W Koehnlein, A Holzgreve, K Schwendner et al Journal of Chromatography A 1687 (2023) 463696 Fig Kp screen contour plots of MAB C on Eshmuno® CMX displaying the flow-through of total protein, host cell protein (HCP), high molecular weight (HMW) impurities, main peak, and low molecular weight (LMW) impurities ecuted at pH 9.0 Especially, HCP could be removed in part with this washing step at elevated pH and low salt concentration because all three antibodies showed sufficient binding under these conditions In contrast, strong HCP binding occurred in the acidic pH range largely independent from the salt concentration, while the final HCP content in the elution pool was mainly influenced by the elution pH Type, amount, and binding behavior of LMW were more specific for the individual antibody, so that their removal occurred under different conditions Kp screens identified best conditions for elution in a window of pH 5-6 with salt concentrations around 300 mM and above This window fell in a wide transition zone of gradual binding differences indicating that elution conditions regarding optimal selectivity and product recovery can be carefully adapted and balanced between pH and salt to establish reproducible and robust column processes Overall, the results obtained for Eshmuno® CMX indicate a generic approach for a bind and elute mode with particular benefit for efficient purification of highly hydrophobic molecules and especially bsAbs with complex formats The single unit operation step, moreover, offers the option of replacing a two-column procedure using CEX and HIC As the mAbs under study range from moderate up to extremely high hydrophobicity (relative hydrophobicity of 0.33- 0.74) the novel MMC resin appears suitable to cover a wide range of antibodies of different hydrophobicity grade This may include standard antibodies with low hydrophobicity lacking strong binding, so that the new MMC resin may be also applied in a flow-through mode moderate hydrophobic properties like the novel MMC Eshmuno® CMX The combination of weak CEX and moderate hydrophobic interactions allows for robust operation under standard pH and salt conditions during binding, washing, and final antibody elution, resulting in high product purity and yield The successful purification of three antibodies with widely different formats, impurity profiles, and hydrophobicities showed the high versatility of Eshmuno® CMX In view of its favorable separation properties and high selectivity, the resin can also provide an alternative approach for further purification challenges, such as the separation of glycosylation variants and antibody drug conjugates with different drug to antibody ratios Moreover, by combining characteristics of CEX and HIC in a single unit operation, implementation of the novel resin provides the opportunity for simplification and cost-saving during downstream processing Integration into a platform purification approach for mAb manufacturing is therefore favored Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper CRediT authorship contribution statement Wolfgang Koehnlein: Conceptualization, Methodology, Formal analysis, Writing – original draft Annika Holzgreve: Investigation, Writing – original draft Klaus Schwendner: Investigation Romas Skudas: Supervision, Funding acquisition, Writing – review & editing Florian Schelter: Conceptualization, Supervision, Funding acquisition, Writing – review & editing Conclusions The present study demonstrated the effective purification of highly hydrophobic complex antibody formats with a resin with 10 W Koehnlein, A Holzgreve, K Schwendner et al Journal of Chromatography A 1687 (2023) 463696 Data availability [8] M Surowka, W Schaefer, C Klein, Ten years in the making: application of CrossMAb technology for the development of therapeutic bispecific antibodies and antibody fusion proteins, MAbs 13 (2021) 1967714, doi:10.1080/19420862 2021.1967714 [9] S.W Chen, W Zhang, Current trends and challenges in the downstream purification of bispecific antibodies, Antib Ther (2021) 73–88, doi:10.1093/abt/ tbab007 [10] A.M Merchant, Z Zhu, J.Q Yuan, A Goddard, C.W Adams, L.G Presta, P Carter, An efficient route to human bispecific IgG, Nat Biotechnol 16 (1998) 677–681, doi:10.1038/nbt0798-677 [11] J.B Ridgway, L.G Presta, P Carter, Knobs-into-holes’ engineering of antibody CH3 domains for heavy chain heterodimerization, Protein Eng (1996) 617– 621, doi:10.1093/protein/9.7.617 [12] P Gagnon, Technology trends in antibody purification, J Chromatogr A 20 (2012) 1221 57-70, doi:10.1016/j.chroma.2011.10.034 [13] J Tang, X Zhang, T Chen, Y Wang, Y Li, Removal of half antibody, hole-hole homodimer and aggregates during bispecific antibody purification using MMC ImpRes mixed-mode chromatography, Protein Expr Purif 167 (2020) 105529, doi:10.1016/j.pep.2019.105529 [14] V Halan, S Maity, R Bhambure, A.S Rathore, Multimodal chromatography for purification of biotherapeutics - a review, Curr Protein Pept Sci 20 (2019) 4– 13, doi:10.2174/1389203718666171020103559 [15] R.A Chmielowski, S Meissner, D Roush, T.O Linden, E Glowacki, J Konietzko, J Nti-Gyabaah, Resolution of heterogeneous charged antibody aggregates via multimodal chromatography: a comparison to conventional approaches, Biotechnol Prog 30 (2014) 636–645, doi:10.1002/btpr.1908 [16] Y.F Lee, S Kluters, M Hillmann, T von Hirschheydt, C Frech, Modeling of bispecific antibody elution in mixed-mode cation-exchange chromatography, J Sep Sci 40 (2017) 3632–3645, doi:10.10 02/jssc.20170 0313 [17] X Santarelli, C Cabanne, Mixed mode chromatography: a novel way toward new selectivity, Curr Protein Pept Sci 20 (2019) 14–21, doi:10.2174/ 1389203718666171024121137 [18] P McDonald, B Tran, C.R Williams, M Wong, T Zhao, B.D Kelley, P Lester, The rapid identification of elution conditions for therapeutic antibodies from cation-exchange chromatography resins using high-throughput screening, J Chromatogr A 1433 (2016) 66–74, doi:10.1016/j.chroma.2015.12.071 [19] M Haverick, S Mengisen, M Shameem, A Ambrogelly, Separation of mAbs molecular variants by analytical hydrophobic interaction chromatography HPLC: overview and applications, MAbs (2014) 852–858, doi:10.4161/mabs 28693 [20] J Weaver, S.M Husson, L Murphy, SR Wickramasinghe, Anion exchange membrane adsorbers for flow-through polishing steps: Part II Virus, host cell protein, DNA clearance, and antibody recovery, Biotechnol Bioeng 110 (2013) 500–510, doi:10.1002/bit.24724 The authors not have permission to share data Acknowledgments The authors thank Alisha Koch, Roxana Disela, Svetlana Melnik, Hakyoung Kim, Katja Schubert, Karin Reindl and Alexander Kurtenbach for expert technical support We also want to thank Raena Morley for proof reading of the manuscript and Marlena Surowka for providing the molecule pictograms In addition, we thank Heidemarie Peuker (BISs - Biomedical Investigation Services) for diligent preparation of the manuscript This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors References [1] U Brinkmann, R.E Kontermann, The making of bispecific antibodies, MAbs (2017) 182–212, doi:10.1080/19420862.2016.1268307 [2] R.E Kontermann, Dual targeting strategies with bispecific antibodies, MAbs (2012) 182–197, doi:10.4161/mabs.4.2.190 0 [3] S Dengl, K Mayer, F Bormann, H Duerr, E Hoffmann, B Nussbaum, M Tischler, M Wagner, A Kuglstatter, L Leibrock, C Buldun, G Georges, U Brinkmann, Format chain exchange (FORCE) for high-throughput generation of bispecific antibodies in combinatorial binder-format matrices, Nat Commun 11 (2020) 4974, doi:10.1038/s41467- 020- 18477- [4] A.F Labrijn, M.L Janmaat, J.M Reichert, P Parren, Bispecific antibodies: a mechanistic review of the pipeline, Nat Rev Drug Discov 18 (2019) 585–608, doi:10.1038/s41573- 019- 0028- [5] C Klein, C Sustmann, M Thomas, K Stubenrauch, R Croasdale, J Schanzer, U Brinkmann, H Kettenberger, J.T Regula, W Schaefer, Progress in overcoming the chain association issue in bispecific heterodimeric IgG antibodies, MAbs (2012) 653–663, doi:10.4161/mabs.21379 [6] W Schaefer, J.T Regula, M Bähner, J Schanzer, R Croasdale, H Dürr, C Gassner, G Georges, H Kettenberger, S Imhof-Jung, M Schwaiger, K.G Stubenrauch, C Sustmann, M Thomas, W Scheuer, C Klein, Immunoglobulin domain crossover as a generic approach for the production of bispecific IgG antibodies, Proc Natl Acad Sci U S A 108 (2011) 11187–11192, doi:10.1073/pnas.1019002108 [7] C Klein, W Schaefer, J.T Regula, The use of CrossMAb technology for the generation of bi- and multispecific antibodies, MAbs (2016) 1010–1020, doi:10.1080/19420862.2016.1197457 11 ... difficult, elaborate and limited regarding yield and puri? ?cation success Generally, puri? ?cation of complex formats may benefit from special adaptations of the chromatography-based downstream processing... that highly hydrophobic complex antibody formats can be efficiently purified using a MMC with moderate hydrophobic characteristics Materials and methods 2.1 Antibody material Humanized IgG1 antibodies... Journal of Chromatography A 1687 (2023) 463696 Fig Determination of relative hydrophobicity of the antibodies using commercially available pharmaceutical antibodies for linear regression analysis