Search

Effective cleaning of chromatography resin is required to prevent fouling and maximize the number of processing cycles which can be achieved. Optimization of resin cleaning procedures, however, can lead to prohibitive material, labor, and time requirements, even when using milliliter scale chromatography columns. In this work, high throughput (HT) techniques were used to evaluate cleaning agents for a monoclonal antibody (mAb) polishing step utilizing Fractogel(®) EMD TMAE HiCap (M) anion exchange (AEX) resin. For this particular mAb feed stream, the AEX resin could not be fully restored with traditional NaCl and NaOH cleaning solutions, resulting in a loss of impurity capacity with resin cycling. Miniaturized microliter scale chromatography columns and an automated liquid handling system (LHS) were employed to evaluate various experimental cleaning conditions. Cleaning agents were monitored for their ability to maintain resin impurity capacity over multiple processing cycles by analyzing the flowthrough material for turbidity and high molecular weight (HMW) content. HT experiments indicated that a 167 mM acetic acid strip solution followed by a 0.5 M NaOH, 2 M NaCl sanitization provided approximately 90% cleaning improvement over solutions containing solely NaCl and/or NaOH. Results from the microliter scale HT experiments were confirmed in subsequent evaluations at the milliliter scale. These results identify cleaning agents which may restore resin performance for applications involving fouling species in ion exchange systems. In addition, this work demonstrates the use of miniaturized columns operated with an automated LHS for HT evaluation of chromatographic cleaning procedures, effectively decreasing material requirements while simultaneously increasing throughput. Biotechnol. Bioeng. 2016;113: 1251-1259. © 2015 Wiley Periodicals, Inc.

A library of displacer analogues with varying degrees of electrostatic, hydrophobic and hydrogen bonding moieties was evaluated for their ability to enhance the selectivity of multimodal (MM) chromatography under high loading conditions. The library was screened for displacement of model proteins using a robotic liquid handling system and selective batch separations were achieved for proteins that were inseparable with linear gradient chromatography. Trends in protein displacement were identified and displacers with higher hydrophobicity and net charge exhibited improved protein displacements. Proteins that interacted with the resins primarily via electrostatic interactions were more readily displaced than those that possessed a significant hydrophobic contribution to their binding. In addition, multimodal displacers were found to be more selective than single mode electrostatic displacers. Column chromatography studies were also carried out and baseline separations were achieved for model protein pairs using selective displacement. Finally, operation of these columns in the desorption mode resulted in baseline separation of model proteins which were not separable by selective displacement chromatography. This study indicates that the inherent selectivity of MM resins can be augmented by the selectivity of the displacer under non-linear competitive binding conditions, creating new opportunities for protein separations not possible using traditional gradient operations.

A protein library exhibiting a range of properties was employed to study protein binding behavior in hydroxyapatite systems. Chromatographic retention on ceramic hydroxyapatite (CHT) chromatography was determined using a sodium chloride gradient in the presence of different phosphate concentrations. Results from the column experiments were then analyzed using various quantitative structure property relationship (QSPR) based modeling approaches. Using the experimental data set in concert with new molecular descriptors, QSPR classification models were generated to provide improved understanding of protein binding in CHT systems. In addition, nonlinear SVM QSPR prediction models were generated and employed as a predictive tool for protein affinity in CHT. Interestingly, a class of descriptors which describe synergistic binding with both metal chelation and cation exchange interactions on the angstrom length scale was found to play a vital role for protein binding in all of the models developed for CHT. The importance of this descriptor suggests the importance of synergistic binding in CHT, which has not been previously described in the literature. This study provides a deeper understanding into the mechanisms and selectivity of protein adsorption in CHT and will help to create predictive models which could be used for methods development in bioseparation processes.

Selective desorption on ceramic hydroxyapatite (CHT) was implemented for the purification of monomeric monoclonal antibody (mAb) from associated aggregates and other post-protein A step impurities. A robotic liquid handling system was employed to carry out a parallel batch screen of selective desorbents on a post-protein A step mAb mixture. The effect of different phosphate concentrations was also investigated. Selective batch separations were achieved between monomeric mAb and associated aggregates/impurities. The batch screen results also established optimal mobile phase conditions for each selective desorbent. These initial batch results were then used to guide column separations, and baseline separation of monomeric mAb from associated aggregates and impurities was achieved, validating the screening results. Selective desorption also resulted in improved separations on CHT, with 100% yield of pure monomeric mAb as compared to 61% and 79%, respectively, for conventional linear and step gradient operations. This proof of concept study demonstrates selective desorption on CHT as an effective separation technique for the purification of monomeric mAb from associated aggregates and other post-protein A step impurities in a single process step.

Robotic high-throughput batch screening was carried out to examine the efficacy of positively charged, alkyl based molecules as chemically selective displacers. Steric mass action (SMA) dynamic affinity plots were constructed, identifying some of the displacers as SMA selective. Saturation transfer difference NMR experiments were conducted, identifying other displacers as chemically selective. These analyses identified molecules which gave selective separations, but did not conform to either the SMA or chemically selective model. Finally, dynamic light scattering experiments were conducted to directly detect micelle formation of the displacers. The use of micelles potentially adds an orthogonal dimension of selectivity to the displacement process.

In this paper we examine whether adding a more retained protein to the feed will mitigate displacer-protein interactions in the column, thus affecting the displacement modality that occurs (chemically selective vs. traditional displacement chromatography). STD-NMR experiments were carried out to probe displacer-protein interactions for the chemically selective displacer chloroquine diphosphate and the results indicated that this displacer only had measurable interactions with the protein alpha-chymotrypsinogen A. For a two component feed mixture containing ribonuclease A and alpha-chymotrypsinogen A, the separation resulted in the displacement of ribonuclease A, with the more hydrophobic alpha-chymotrypsinogen A remaining on the column. On the other hand, when the experiment was repeated with cytochrome c added to the feed, all three feed proteins were displaced. Column simulations indicated that the combination of sample self-displacement occurring during the introduction of the feed, along with the dynamics of the initial displacement process at the column inlet was responsible for this behavior. These results indicate that for this class of hydrophobic-based selective displacers, in order for the protein to be selectively retained, the protein should be the most strongly retained feed component.

A parallel batch screening technique was employed to identify chemically selective displacers which exhibited exclusive separation behavior for the protein pair α-chymotrypsin/ribonuclease A on a strong cation exchange resin. Two selective displacers, 1-(4-chlorobenzyl)piperidin-3-aminesulfate and N′1′-(4-methyl-quinolin-2-yl)-ethane-1,2-diamine dinitrate, and one non-selective displacer, spermidine, were selected as model systems to investigate the mechanism of chemically selective displacement chromatography. Saturation transfer difference (STD) NMR was used to directly evaluate displacer–protein binding. The results indicated that while binding occurred between the two chemically selective displacers and the more hydrophobic protein, α-chymotrypsin, no binding was observed with ribonuclease A. Further, the non-selective displacer, spermidine, was not observed to bind to either protein. Importantly, the binding event was observed to occur primarily on the aromatic portion of the selective displacers. Extensive molecular dynamic simulations of protein–displacer–water solution were also carried out. The MD results corroborated the NMR findings demonstrating that the binding of selective displacers occurred primarily on hydrophobic surface patches of α-chymotrypsin, while no significant long term binding to ribonuclease A was observed. The non-selective displacer did not show significant binding to either of the proteins. MD simulations also indicated that the charged amine group of the selective displacers in the bound state was primarily oriented towards the solvent, potentially facilitating their interaction with a resin surface. These results directly confirm that selective binding between a protein and displacer is the mechanism by which chemically selective displacement occurs. This opens up many possibilities for future molecular design of selective displacers for a range of applications. Biotechnol. Bioeng. 2009;102: 1428–1437. © 2008 Wiley Periodicals, Inc.

One of the major impediments to the implementation of displacement chromatography for the purification of biomolecules is the need to collect fractions from the column effluent for time-consuming offline analysis. The ability to employ direct online monitoring of displacement chromatography would have significant implications for applications ranging from analytical to preparative bioseparations. To this end, a set of novel fluorescent displacers were rationally designed using known chemically selective displacers as a template. Fluorescent cores were functionalized with different charge moieties, creating a homologous library of displacers. These compounds were then tested on two protein pairs, alpha-chymotrypsinogen A/ribonuclease A and cytochrome c/lysozyme, using batch and column displacement experiments. Of the synthesized displacers, two were found to be highly selective while one was determined to be a high-affinity displacer. Column displacements using one of the selective displacers yielded complete separation of both protein pairs while facilitating direct online detection using UV and fluorescence detection. Saturation transfer difference NMR was also carried out to investigate the binding of the fluorescent displacers to proteins. The results indicated a selective binding between the selective displacers and alpha-chymotrypsinogen A, while no binding was observed for ribonuclease A, confirming that protein-displacer binding is responsible for the selectivity in these systems. This work demonstrates the utility of fluorescent displacers to enable online monitoring of displacer breakthroughs while also acting as efficient displacers for protein purification.

We describe an affinity-based strategy for designing selective protein displacers for the chromatographic purification of proteins. To design a displacer that is selective for a target protein, we attached a component with affinity for the target protein to a resin-binding component; we then tested the ability of such displacers to selectively retain the target protein on a resin relative to another protein having a similar retention time. In particular, we synthesized displacers based on biotin, which selectively retained avidin as compared to aprotinin on SP Sepharose high performance resin. In addition, we have extended this approach to develop an affinity-peptide-based displacer that discriminates between lysozyme and cytochrome c. Here, a selective displacer was designed from a lysozyme-binding peptide that had been identified and optimized previously using phage-display technology. Our results suggest a general strategy for designing highly selective affinity-based displacers by identifying molecules (e.g., peptides) that bind to a protein of interest and using an appropriate linker to attach these molecules to a moiety that binds to the stationary phase.

Effective cleaning of chromatography resin is required to prevent fouling and maximize the number of processing cycles which can be achieved. Optimization of resin cleaning procedures, however, can lead to prohibitive material, labor, and time requirements, even when using milliliter scale chromatography columns. In this work, high throughput (HT) techniques were used to evaluate cleaning agents for a monoclonal antibody (mAb) polishing step utilizing Fractogel(®) EMD TMAE HiCap (M) anion exchange (AEX) resin. For this particular mAb feed stream, the AEX resin could not be fully restored with traditional NaCl and NaOH cleaning solutions, resulting in a loss of impurity capacity with resin cycling. Miniaturized microliter scale chromatography columns and an automated liquid handling system (LHS) were employed to evaluate various experimental cleaning conditions. Cleaning agents were monitored for their ability to maintain resin impurity capacity over multiple processing cycles by analyzing the flowthrough material for turbidity and high molecular weight (HMW) content. HT experiments indicated that a 167 mM acetic acid strip solution followed by a 0.5 M NaOH, 2 M NaCl sanitization provided approximately 90% cleaning improvement over solutions containing solely NaCl and/or NaOH. Results from the microliter scale HT experiments were confirmed in subsequent evaluations at the milliliter scale. These results identify cleaning agents which may restore resin performance for applications involving fouling species in ion exchange systems. In addition, this work demonstrates the use of miniaturized columns operated with an automated LHS for HT evaluation of chromatographic cleaning procedures, effectively decreasing material requirements while simultaneously increasing throughput. Biotechnol. Bioeng. 2016;113: 1251-1259. © 2015 Wiley Periodicals, Inc.

Affinity interaction based resins have long been the foundation for various purification platforms of biological molecules. For the purification of rAAV, the affinity resin AVB Sepharose HP (AVB) has been used to successfully purify a number of different serotypes. However, to date the use of AVB for the purification of the AAV8 serotype has been relatively unsuccessful due in part to modest binding capacity. In this work we investigate the use of a new, prototype affinity resin for the purification of AAV8 from crude feed streams. Performance of the AVB and prototype resins were directly compared for the purification of selected AAV8 feed streams. Differences in capacities, loading steps, wash step tolerances and elution steps will be discussed. In addition, the resulting performance of purifying AAV8 using the prototype resin is directly compared against the best case performance of the AVB resin purifying a different serotype more commonly used on AVB. Finally, the scale-up use of the prototype resin from bench to manufacturing will be discussed. These results demonstrate that the prototype affinity resin is a viable option for the scaleable, chromatographic purification of AAV8 from crude harvest streams.

Design of peptide affinity ligands against biological targets is important for a broad range of applications. Here, we report on de novo and combinatorial strategies for the design of high-affinity and high-specificity peptides against S-protein as a target. The peptide libraries employed in this study contain (1) consensus motif (CM) sequences identified from high-throughput phage combinatorial screening, (2) point mutations of CM sequences, and (3) de novo sequences rationally designed based on stereo-chemical information of the complex between S-protein and its natural ligand, S-peptide. In general, point mutations to CM allowed for modulating peptide affinity and specificity over a broad range. This is particularly useful in designing peptides with varying affinities and specificities for the target. De novo sequences, especially those based on the S-protein binding pocket, on average bound with higher affinities within a narrow range (10–100 nM) as compared to point mutations to CM (1 nM–2 $$\mu $$ M). As such, the approaches described here serve as a general guide for optimizing the design of peptide affinity ligands for a wide range of target proteins or applications.

The development and commercialization of a biological pharmaceutical, such as recombinant adeno associated virus (rAAV), requires rigorous quantitative and qualitative testing. An accurate particle titer measurement is important for the development of purification processes. Current methods for titering rAAV include qPCR, for measuring genome copy titer, and ELISA, for measuring particle titer. Additionally, methods such as dynamic light scattering (DLS) have been considered for assessment of particle size and aggregation but with limited success for rAAV. Here, a novel method is described called Nanoparticle Tracking Analysis (NTA), to provide rAAV titer (concentration) within minutes, as opposed to hours or days required by traditional techniques. NTA characterization was enabled by a gold labeling technique to increase the light scattered from rAAVs. With a purified gold labelled rAAV sample, NTA results indicated a monodisperse size distribution and a measured rAAV particle titer equivalent to the concentration based on genome copies by qPCR. Particle titer determination can also be relevant for assessment of product stability. NTA was used to assess the stability of rAAV over a 3 month thermal stress test. The NTA method enabled the detection of a reduction of the primary rAAV particle peak within 1 month, as well as further reductions over time. These initial results demonstrate the potential uses of NTA for product characterization.

Continued interest in using rAAV vectors as therapies for a range of indications with varying prevalence often translates to increased manufacturing scales. A significant challenge that is encountered is the use of different manufacturing materials during the transition between early research and development and the clinical manufacturing scale. For example, manufacturing components, ranging from bioprocessing bags and hard plastics to bioprocess tubing and filters can be constructed of different materials and are not often engineered for use with rAAV. In this work we investigate the loss of rAAV to non-specific surface adsorption on various materials of construction. In addition, the effect of different background matrices was also investigated. These data demonstrate that surface adsorptive loss of rAAV can be significant, and that losses can vary significantly even for materials of similar construction. To mitigate adsorptive loss, several surfactants and other components were examined for their ability to reduce surface adsorption of rAAV. A number of these agents were identified as viable options for cGMP manufacturing. These results provide insight for the development of scaleable, more robust and higher yielding manufacturing processes to meet the continued demand for rAAV therapeutics.

A detailed study was carried out to combine the unique selectivity of ceramic hydroxyapatite (CHA) with the separation power of selective displacement chromatography. A robotic liquid handling system was employed to carry out a parallel batch screen on a displacer library made up of analogous compounds. By incorporating positively charged, metal chelating and/or hydrogen bonding groups into the design of the displacer, specific interaction sites on CHA were targeted, thus augmenting the selectivity of the separation. The effect of different mobile phase modifiers, such as phosphate, sulfate, lactate and borate, were also investigated. Important functional group moieties and trends for the design of CHA displacers were established. Selective batch separations were achieved between multiple protein pairs which were unable to be resolved using linear gradient techniques, demonstrating the applicability of this technique to multiple protein systems. The specific interaction moieties used on the selective displacer were found to dictate which protein was selectively displaced in the separation, a degree of control not possible using a mono-interaction type resin in displacement chromatography. Mobile phase modifiers were also shown to play a crucial role, augmenting the selectivity of a displacer in a synergistic fashion. Column separations were carried out using selective displacers and mobile phase modifiers identified in the batch experiments, and baseline separation of the previously unresolved protein pairs was achieved. Further, the elution order in these systems was able to be reversed while still maintaining baseline separations. This work establishes a new class of separations which combine the selectivities of multi-modal resins, displacers/eluents, and mobile phase modifiers to create unique selectivity windows unattainable using traditional modes of operation.

A library of molecular analogues to the selective displacer, N'1'-(4-methylquinolin-2-yl)ethane-1,2-diamine dinitrate, was employed to study the effects of changes in displacer chemistry on their efficacy for selective separations. High throughput screens were carried out using a robotic liquid handling system to examine the ability of these compounds to selectively displace proteins in batch adsorption systems. Experiments were conducted using the model protein pairs ribonuclease A/alpha-chymotrypsinogen A and cytochrome C/lysozyme on a strong cation exchanger. Selectivity pathway and DC-50 plots were constructed from the analogue screen data, and results indicated that minor changes in the molecular design of the displacer can have a significant impact on the separation behavior. Specifically, charge density and spacing of resin and protein interaction moieties were found to be important. The screen also identified a new displacer, 4-methyl-2-piperazin-1-yl-quinoline, which produced a more selective displacement than previously reported with the original compound. A steric mass action dynamic affinity plot was constructed to validate that this new displacer was acting as a chemically selective, rather than a steric mass action selective displacer. Finally, saturation transfer difference NMR experiments were conducted to examine protein-displacer interactions with these displacers and protein pairs. These results demonstrate how subtle changes in displacer design can be employed to fine-tune the separation performance of chemically selective displacers.

A robotic high-throughput displacer screen was developed and employed to identify chemically selective displacers for several protein pairs in cation exchange chromatography. This automated screen enabled the evaluation of a wide range of experimental conditions in a relatively short period of time. Displacers were evaluated at multiple concentrations for these protein pairs, and DC-50 plots were constructed. Selectivity pathway plots were also constructed and different regimes were established for selective and exclusive separations. Importantly, selective displacement was found to be conserved for multiple protein pairs, demonstrating the technique to be applicable for a range of protein systems. Although chemically selective displacers were able to separate protein pairs that had similar retention in ion exchange but different surface hydrophobicities, they were not able to distinguish protein pairs with similar surface hydrophobicities. This corroborates that displacer-protein hydrophobic interactions play an important role for this class of selective displacers. Important functional group moieties were established and efficient displacers were identified. These results demonstrate that the design of chemically selective displacers requires a delicate balance between the abilities to displace proteins from the resin and to bind to a selected protein. The use of robotic screening of displacers will enable the extension of chemically selective displacement chromatography beyond hydrophobic displacer-protein interactions to other secondary interactions and more selective displacement systems.

The recent resurgence of using rAAV vectors as commercially viable therapeutic agents has highlighted the need to identify stable capsid/genomic construct combinations which can both endure a typical cGMP manufacturing process and possess a long product shelf life. In this work we investigate the inherent stability of various capsids and genomic constructs using differential scanning calorimetry (DSC). DSC measures the amount of energy required to increase the temperature of a sample and reference and can be used to determine the Tm of the molecule in question. Numerous commercially relevant rAAV serotypes containing several genomic constructs (self-complimentary and oversized genomes) were subjected to melting curve analysis using DSC to compare stability. Additionally, various mobile phase modifiers were added to solution in order to probe the stability of rAAV vector, using organics and kosmotropic/chaotropic salts. These results demonstrate DSC as an effective tool to examine inherent rAAV vector capsid stability and support a role for the serotype and vector genome structure to affect the observed stability.