Separation of full and empty adeno-associated virus capsids by anion-exchange chromatography using choline-type salts.

Development of clinically desirable adeno-associated virus (AAV) vectors with optimal genome design requires rapid and accurate analytical methods to assess AAV quality. Anion-exchange (AEX) chromatography provides a powerful analytical method for full/empty AAV capsid ratio determination. However, the current AEX methodology for separation of empty and full AAV capsids largely relies on the use of the highly toxic tetramethylammonium chloride (TMAC). Here, we describe a novel analytical AEX method for separation of empty and full AAV capsids that uses only non-toxic, choline-type compounds that contain structural similarity to the quaternary ammonium ligand present on the surface of AEX resin. Choline-Cl gradient, combined with sensitive fluorescence detection, allowed a safe and effective separation of empty and full AAV capsids with reproducible empty/full ratio determination. The choline-based assay was suitable for commonly used serotypes, AAV2, AAV5, AAV6, and AAV8. The limit of detection was ∼3.9 × 108 virus particles in the assay. A gradient-hold step-gradient elution with choline-Cl resulted in enhanced baseline separation of empty and full AAV8 capsids. In summary, the use of choline-Cl in the AEX assay is recommended for empty/full capsid ratio determination and other applications in AAV production, and it eliminates the necessity of using toxic TMAC.

Evaluation of restricted access media for the purification of cell culture-derived Orf viruses.

Recently, multimodal chromatography using restricted access media (RAM) for the purification of nanoparticles, such as viruses has regained increasing attention. These chromatography resins combine size exclusion on the particle shell and adsorptive interaction within the core. Accordingly, smaller process-related impurities, for example, DNA and proteins, can be retained, while larger product viruses can pass unhindered. We evaluated a range of currently available RAM, differing in the shells' pore cut-off and the core chemistry, for the purification of a cell culture-derived clarified model virus, namely the Orf virus (ORFV). We examined impurity depletion and product recovery as relevant criteria for the evaluation of column performance, as well as scale-up robustness and regeneration potential for evaluating a multiple use application. The results indicate that some columns, for example, the Capto Core, enable both a high DNA and protein removal, while others, for example, the Monomix Core 60 (MC60), are more suitable for DNA depletion. Furthermore, column regeneration is facilitated by using columns with larger shell pores (5000 vs. 700 kDa) and weaker binding interactions (anion exchange vs. multimodal). According to these findings, the choice of RAM resins should be selected according to the respective feed sample composition and the planned number of application cycles.

Salt-tolerant cation exchanger-containing sulfate groups as a viable alternative for mixed-mode type and heparin-based affinity resins.

Ion-exchange chromatography is still one of the most popular protein separation techniques. Before chromatographic separation, the high salt concentration in various samples necessitates additional steps. Therefore, low salt tolerance of ion-exchange resins is a drawback that needs to be addressed. Herein, the differences in salt tolerance and hydrophobicity of strong cation-exchange TOYOPEARL resins of sulfonium and sulfate-types were investigated. Despite only a minor structural difference, differences in selectivity and salt tolerance between the sulfate and sulfonic groups were detected. In silico calculations were also carried out for model substances representing the sulfonium and sulfate groups, wherein significant differences in hydrophobicity was observed. These experiments confirmed the hypothesis that the salt tolerance, higher affinity, and selectivity for certain vitamin K dependent clotting factors are interrelated and dependent on the presence of the sulfate group. Separation of clotting factor IX from the prothrombin complex concentrate further to confirmed the affinity for these proteins. The results show that the use of only a resin with the sulfate ligand and not with the sulfonic acid ligand allows for a facile and rapid separation of clotting factor IX and other vitamin K dependent clotting factors.

Evaluation of hydrophobic-interaction chromatography resins for purification of antibody-drug conjugates using a mimetic model with adjustable hydrophobicity.

Antibody drug conjugates are cytotoxic pharmaceuticals, designed to destroy malignant cells. A cytotoxic molecule is attached to an antibody that binds specific to a cancer-cell surface. Given the high toxicity of the drugs, strict safety standards have to be kept. For this reason, an antibody drug conjugates model was developed with fluorescein 5-isothiocyanate as the nontoxic payload surrogate. Due to the similar hydrophobicity, this model is used to establish a suitable purification process and characterization method for antibody drug conjugates. Because of the pH dependent solubility of fluorescein, the hydrophobicity of conjugates can be modulated by the pH value. Based on the complex heterogeneity and hydrophobicity of the conjugates a chromatographic purification is challenging. Hydrophobic interaction chromatography is used for analytical as well as for preparative separation. Because of the increased hydrophobicity of the conjugates compared to native antibody, hydrophobic interaction chromatography often suffer from resolution and recovery problems. Conjugates were separated differing on the number of payloads attached to the antibody. For this matter, the drug-antibody ratio is determined and used as a quantitative term. The conjugates are purified at high recoveries and resolution by step gradients using suitable resins, allowing the separation of the target drug-antibody ratio.

Size distribution analysis of influenza virus particles using size exclusion chromatography.

Size exclusion chromatography is a standard method in quality control of biopharmaceutical proteins. In contrast, vaccine analysis is often based on activity assays. The hemagglutination assay is a widely accepted influenza quantification method, providing no insight in the size distribution of virus particles. Capabilities of size exclusion chromatography to complement the hemagglutination assay are investigated. The presented method is comparatively robust regarding different buffer systems, ionic strength and additive concentrations. Addition of 200mM arginine or sodium chloride is necessary to obtain complete virus particle recovery. 0.5 and 1.0M arginine increase the hydrodynamic radius of the whole virus particles by 5nm. Sodium citrate induces virus particle aggregation. Results are confirmed by dynamic light scattering. Retention of a H1N1v strain correlates with DNA contents between 5ng/mL and 670ng/mL. Quantitative elution of the virus preparations is verified on basis of hemagglutination activity. Elution of hemagglutination inducing compounds starts at a flow channel diameter of 7000nm. The universal applicability is demonstrated with three different influenza virus samples, including an industrially produced, pandemic vaccine strain. Size distribution of the pandemic H1N1v 5258, H1N1 PR/8/34, and H3N2 Aichi/2/68 preparations spreads across inter- and intra-particle volume and extends to the secondary interaction dominated range. Thus, virus particle debris seems to induce hemagglutination. Fragments generated by 0.5% Triton™ X-100 treatment increase overall hemagglutination activity.

Mono- and polyprotic buffer systems in anion exchange chromatography of influenza virus particles.

Different ions typically used in downstream processing of biologicals are evaluated for their potential in anion exchange chromatography of an industrially produced, pandemic influenza H1N1 virus. Capacity, selectivity and recovery are investigated based on single step elution parallel chromatography experiments. The inactivated H1N1 feedstream is produced in Madin-Darby Bovine Kidney cells. Interesting effects are found for sodium phosphate and sodium citrate. Both anions are triprotic kosmotropes. Anion exchange chromatography generally offers high scalability to satisfy sudden demands for vaccines, which may occur in case of an emerging influenza outbreak. Appropriate pH conditions for H1N1 adsorption are determined by Zeta potential measurements. The dynamic binding capacity of a salt tolerant polyamine-type resin is up to 6.4 times greater than the capacity of a grafted Q-type resin. Pseudo-affinity interactions of polyamines with the M2 protein of influenza may contribute to the obtained capacity increase. Both resins achieve greater capacity in sodium phosphate buffer compared to Tris/HCl. A recovery of 67% and DNA clearance close to 100% without DNAse treatment are achieved for the Q-type resin. Recovery of the virus from the salt tolerant resin requires the use of polyprotic acids in the elution buffer. 85% of the DNA and 60% of the proteins can be removed by the salt tolerant resin. The presence of sodium phosphate during anion exchange chromatography seems to support stability of the H1N1 particles in presence of hydrophobic cations.

Routes to improve binding capacities of affinity resins demonstrated for Protein A chromatography.

Protein A chromatography is a well-established platform in downstream purification of monoclonal antibodies. Dynamic binding capacities are continuously increasing with almost every newly launched Protein A resin. Nevertheless, binding capacities of affinity chromatography resins cannot compete with binding capacities obtained with modern ion exchange media. Capacities of affinity resins are roughly 50% lower. High binding capacities of ion exchange media are supported by spacer technologies. In this article, we review existing spacer technologies of affinity chromatography resins. A yet known effective approach to increase the dynamic binding capacity of Protein A resins is oligomerization of the particular Protein A motifs. This resembles the tentacle technology used in ion exchange chromatography. Dynamic binding capacities of a hexameric ligand are roughly twice as high compared to capacities obtained with a tetrameric ligand. Further capacity increases up to 130mg/ml can be realized with the hexamer ligand, if the sodium phosphate buffer concentration is increased from 20 to 100mM. Equilibrium isotherms revealed a BET shape for the hexamer ligand at monoclonal antibody liquid phase concentrations higher than 9mg/ml. The apparent multilayer formation may be due to hydrophobic forces. Other quality attributes such as recovery, aggregate content, and overall purity of the captured monoclonal antibody are not affected.

Kinetic plots in aqueous size exclusion chromatography of monoclonal antibodies and virus particles.

The growing importance of monoclonal antibodies and virus particles has led to a pressure for faster size exclusion chromatography. In recent years, numerous small particle columns for size exclusion chromatography of biologicals have been introduced. Small particles are a strategy to reduce analysis time. In the following study, opportunities of small particles in size exclusion chromatography of large biomolecules are investigated. Poppe plots reveal that the lower particle size limit depends on the size of the sample molecule. Hydrodynamic radii of monoclonal antibody monomer, aggregates and H1N1 as well as the diffusion coefficients were determined. Considering this sample compound dependency, kinetic plots referring to the resolution of a distinct compound pair instead of the plate number of a single analyte are more meaningful. Plate times were found to be equivalent with 4 and 2µm particles for a monoclonal antibody aggregate separation at resolutions smaller than 1.8. Quantification of a H1N1 in clarified cell culture can be accomplished with 17µm and 13µm particles at equal plate times at resolutions smaller than 2.5. Virus polydispersity is likely to be affected by run times of several hours at room temperature and shear forces resulting from particles smaller than 10µm. Comparatively high flow rates should be applied in size exclusion chromatography of the 100nm H1N1 virions.

Purification of PEGylated proteins, with the example of PEGylated lysozyme and PEGylated scFv.

PEGylation is a common and highly accepted possibility for half-life prolongation of proteins by increasing the hydrodynamic size. The chromatographic purification of PEGylated protein, using PEG (poly-ethylene glycol) of different PEG chain lengths, with the example of lysozyme and a scFv, is described in detail here, and helpful suggestions for the purification of other PEGylated proteins are listed. The relevant characterization methods for PEGylated proteins, important for the successful purification, are also described. The purification starts with a CEX (cation exchange) chromatography leading to about 95 % purity for polishing HIC (hydrophobic interaction chromatography) is described.

Mixed electrolytes in hydrophobic interaction chromatography.

An essential part of the modulation of protein-binding capacity in hydrophobic interaction chromatography is the buffer-salt system. Besides using "single" electrolytes, multicomponent electrolyte mixtures may be used as an additional tool. Both the protein solubility and the binding capacity depend on the position of a salt in the so-called Hofmeister series. Specific interactions are observed for an individual protein-salt combination. For salt mixtures, selectivity, recovery, and binding capacity do not behave like for the single salts that are positioned in between the two mixed components in the Hofmeister series, as the continuous correlation would suggest. Thus, finding strategies for mixed salts could potentially lead to improved capacities in hydrophobic interaction chromatography. Mixtures of ammonium sulfate, sodium citrate, sodium sulfate, sodium chloride, sodium acetate, and glycine were used to investigate the binding capacities for lysozyme and a monoclonal antibody on various hydrophobic resins. Resin capacity for two investigated proteins increases when mixtures consisting of a chaotropic and a kosmotropic salt are applied. It seems to be related to the rather basic isoelectric points of the proteins.

Separation of proteins from human plasma by sample displacement chromatography in hydrophobic interaction mode.

Sample displacement chromatography (SDC) in reversed-phase and ion-exchange modes was introduced approximately 20 years ago. This method was first used for the preparative purification of peptides and proteins. Recently, SDC in ion-exchange mode was also successfully used for enrichment of low-abundance proteins from human plasma. In this paper, the use of SDC for the separation of plasma proteins in hydrophobic interaction mode is demonstrated. By use of two or more columns coupled in series during sample application, and subsequent elution of detached columns in parallel, additional separation of bound proteins was achieved. Further low-abundance, physiologically active proteins could be highly enriched and detected by ESI-MS/MS.

Aldehyde PEGylation kinetics: a standard protein versus a pharmaceutically relevant single chain variable fragment.

The mPEG-aldehyde PEGylation with two different PEG sizes and two proteins was experimentally determined with respect to yield, conversion, and selectivity. The kinetic behavior of these PEGylation reactions was simulated using a numerically solved set of differential equations. We show that the assumption of an inactivation of mPEG-aldehyde is crucial for the simulation of the overall PEGylation and that the inactivation is pH-dependent. We further demonstrate that ideal PEGylation parameters such as pH, temperature, reaction time, and protein concentration need to be chosen carefully depending on the protein and PEG size. In terms of selectivity and yield, we show that the reaction should be stopped before the highest mono-PEG concentration is reached. Moreover, room temperature and a slightly acidic pH of approximately 6 are good starting points. In conclusion, selectivity can be optimized choosing a shorter reaction time and a reduced reaction temperature.

Selectivity of monolithic supports under overloading conditions and their use for separation of human plasma and isolation of low abundance proteins.

Human serum albumin (HSA) and immunoglobulin G (IgG) represent over 75% of all proteins present in human plasma. These two proteins frequently interfere with detection, determination and purification of low abundance proteins that can be potential biomarkers and biomarker candidates for various diseases. Some low abundance plasma proteins such as clotting factors and inhibitors are also important therapeutic agents. In this paper, the characterization of ion-exchange monolithic supports under overloading conditions was performed by use of sample displacement chromatography (SDC). If these supports were used for separation of human plasma, the composition of bound and eluted proteins in both anion- and cation-exchange mode is dependent on column loading. Under overloading conditions, the weakly bound proteins such as HSA in anion-exchange and IgG in cation-exchange mode are displaced by stronger binding proteins, and this phenomenon was not dependent on column size. Consequently, small monolithic columns with a column volume of 100 and 200 µL are ideal supports for high-throughput screening in order to develop new methods for separation of complex mixtures, and for sample preparation in proteomic technology.

Solubility and binding properties of PEGylated lysozyme derivatives with increasing molecular weight on hydrophobic-interaction chromatographic resins.

Dynamic binding capacities and resolution of PEGylated lysozyme derivatives with varying molecular weights of poly (ethylene) glycol (PEG) with 5 kDa, 10 kDa and 30 kDa for HIC resins and columns are presented. To find the optimal range for the operating conditions, solubility studies were performed by high-throughput analyses in a 96-well plate format, and optimal salt concentrations and pH values were determined. The solubility of PEG-proteins was strongly influenced by the length of the PEG moiety. Large differences in the solubilities of PEGylated lysozymes in two different salts, ammonium sulfate and sodium chloride were found. Solubility of PEGylated lysozyme derivatives in ammonium sulfate decreases with increased length of attached PEG chains. In sodium chloride all PEGylated lysozyme derivatives are fully soluble in a concentration range between 0.1 mg protein/ml and 10 mg protein/ml. The binding capacities for PEGylated lysozyme to HIC resins are dependent on the salt type and molecular weight of the PEG polymer. In both salt solutions, ammonium sulfate and sodium chloride, the highest binding capacity of the resin was found for 5 kDa PEGylated lysozyme. For both native lysozyme and 30 kDa mono-PEGylated lysozyme the binding capacities were lower. In separation experiments on a TSKgel Butyl-NPR hydrophobic-interaction column with ammonium sulfate as mobile phase, the elution order was: native lysozyme, 5 kDa mono-PEGylated lysozyme and oligo-PEGylated lysozyme. This elution order was found to be reversed when sodium chloride was used. Furthermore, the resolution of the three mono-PEGylated forms was not possible with this column and ammonium sulfate as mobile phase. In 4 M sodium chloride a resolution of all PEGylated lysozyme forms was achieved. A tentative explanation for these phenomena can be the increased solvation of the PEG polymers in sodium chloride which changes the usual attractive hydrophobic forces in ammonium sulfate to more repulsive hydration forces in this hydrotrophic salt.

Investigation of salt properties with electro-acoustic measurements and their effect on dynamic binding capacity in hydrophobic interaction chromatography.

The pH dependence in hydrophobic interaction chromatography (HIC) is usually discussed exclusively in terms of protein dependence and there are no clear defined trends. Many of the deviations from an ideal solution are caused solely by the high salt concentration, as protein concentration is usually negligible. So pH dependency in hydrophobic interaction chromatography could also be the result of pH dependent changes of ion properties from the salt solution. The possibility that pH dependent ion hydration or ion association in highly concentrated salt solutions may influence the dynamic protein binding capacity onto HIC resins was investigated. In buffer solutions commonly used in HIC e.g. sodium chloride, ammonium sulphate and sodium citrate pH dependent maxima in the electro-acoustic signals were found. These maxima are related to an increase of the ion sizes by hydration or ion association. At low ionic strength the maxima are in the range between 4.5 and 6 and they increased in concentrated electrolyte solutions to values between 6 and 8. The range of these maxima is in the same region as dynamic protein binding capacity maxima often observed in HIC. For a qualitative interpretation of this phenomenon of increased protein stabilization by volume exclusion effect extended scaling theory can be used. This theory predicts a maximum of protein stabilization if the ratio of salt ion diameter to water is 1.8. According to the hypothesis raised here, if the pH dependent ratio of salt ion diameter to water approaches this value the transport of the protein in the pore system is less restricted and an increase in binding capacity can be produced.

Fast determination of conditions for maximum dynamic capacity in cation-exchange chromatography of human monoclonal antibodies.

Dynamic binding capacity (DBC) measurements of cation-exchange resins were performed with two human monoclonal antibodies. DBC showed a pH dependent maximum, which was shifted to lower pH values with increasing buffer concentrations and increasing salting-out effect of the buffer anion according to the Hofmeister series. As this downshift correlates well with zeta potential values, a measurement of the latter allows the determination of the pH value for maximum DBC under a given set of conditions. Thus, the use of zeta potential values can accelerate the purification process development and helps to understand the protein adsorption mechanism.

Resin characterization by electro-acoustic measurements.

The electro-acoustic effects, namely the ion vibration potential (IVP) and the colloidal vibration current (CVI), colloidal vibration potential (CVP) first described by P. Debye [P. Debye, J. Chem. Phys. 1 (1933) 13], are a result of charge separation of bound or free ions at different degrees by ultrasonic waves. Today commercial instruments are available to investigate liquid homogeneous and heterogeneous systems. In the present paper the application of this technique for the characterization of salts, protein solutions and resins for biochromatography is shown and valuable information about resins can be derived in a short time. Various resins were investigated with the following results: (1) the CVI magnitude is dependent of several parameters (such as particle size distribution, volume fraction, density difference); (2) the CVI is influenced by the surface modification of the resins. Polymeric modifications decrease the value of CVI. The CVI is generally lower for high capacity resins; (3) the measurement of the electro-acoustic effects can be used to detect small changes in resins. The CVI is dependent of the amount of adsorbed protein in "native" and denatured state.

Comparison between mass transfer properties of weak-anion-exchange resins with graft-functionalized polymer layers and traditional ungrafted resins.

Glycidylmethacrylate was grafted to Toyopearl HW-65M and subsequently modified with diethylamine to obtain a weak anion exchanger. The degree of grafting was varied from 11 to 50%. The binding capacity for bovine serum albumin was 11 mg/ml for the lowest degree of grafting and 97 mg/ml for the highest degree of grafting. The maximum binding capacity was observed at 27% degree of grafting. The mass transfer properties of the grafted resins and an ungrafted resin(Toyopearl DEAE 650M) were investigated assuming rectangular isotherms. Simple models for reaction kinetics, pore- and surface diffusion and film diffusion were used to describe the concentration-time data in batch mode. The data were best fitted by a pore diffusion model. The estimated pore diffusion coefficients (D(P)) for bovine serum albumin were fitted by a polynome to the degree of grafting with an maximum value at 27% of D(P) = 1.95-10(-11) m2/s. Compared to published data of other ungrafted resins and to the molecular diffusion coefficient of bovine serum albumin in free solution of D(P) = 5.6 10(-11) m2/s, the diffusion in grafted layers seems to be accelerated. The breakthrough curves for columns packed with various resins showed a decrease in sharpness with increasing degree of grafting which could not be described by a simple pore diffusion model using the calculated transport coefficients from batch mode. The shape of the breakthrough curves could be well described by a combined film and pore diffusion model. For the ungrafted Toyopearl DEAE 650M resin the breakthrough curve is more favorable and the influence of film diffusion to the mass transfer is reduced. It can be concluded that grafting will increase the capacity and the pore diffusion in batch mode but in column operation the grafting layer has a film resistance which plays an important role in the overall mass transfer.