Research Article
Pharmaceutical Biotechnology
Novel Displacement Agents for Aqueous 2-Phase Extraction Can Be Estimated Based on Hybrid Shortcut Calculations

https://doi.org/10.1016/j.xphs.2016.06.006Get rights and content

Abstract

The purification of therapeutic proteins is a challenging task with immediate need for optimization. Besides other techniques, aqueous 2-phase extraction (ATPE) of proteins has been shown to be a promising alternative to cost-intensive state-of-the-art chromatographic protein purification. Most likely, to enable a selective extraction, protein partitioning has to be influenced using a displacement agent to isolate the target protein from the impurities. In this work, a new displacement agent (lithium bromide [LiBr]) allowing for the selective separation of the target protein IgG from human serum albumin (represents the impurity) within a citrate–polyethylene glycol (PEG) ATPS is presented. In order to characterize the displacement suitability of LiBr on IgG, the mutual influence of LiBr and the phase formers on the aqueous 2-phase system (ATPS) and partitioning is investigated. Using osmotic virial coefficients (B22 and B23) accessible by composition gradient multiangle light-scattering measurements, the precipitating effect of LiBr on both proteins and an estimation of both protein partition coefficients is estimated. The stabilizing effect of LiBr on both proteins was estimated based on B22 and experimentally validated within the citrate–PEG ATPS. Our approach contributes to an efficient implementation of ATPE within the downstream processing development of therapeutic proteins.

Introduction

Since the first biopharmaceutical drug (Humulin) was approved in 1982, the biopharmaceutical market has gained a significant growth with global revenues of 140 billion USD in 2013 and estimated sales of approximately 166 billion USD up to the year 2017.1, 2 Within the biopharmaceutical market, the production of mAbs contributes with global revenues of nearly 75 billion USD and thus accounts to almost 50% of all biopharmaceuticals produced.3 With a current approval rate of 4 new mAbs a year, global sales on mAbs are estimated to reach nearly $125 billion by the year 2020.3 The increased production of mAbs is caused by their enhanced demand as treatment for cancer, for immunologic disorders (e.g., multiple sclerosis, rheumatoid disorders, etc.) or against Alzheimer disease.4, 5

Process innovations in the production of mAbs focused primary on the upstream processing, often disregarding the downstream processing. As a result, costs were shifted from upstream to downstream processing creating a bottleneck.6 State-of-the-art downstream processing of mAbs bases almost exclusively on cost-intensive chromatographic steps and can account to 50%-80% of the total production cost.7 Even though chromatography cannot be replaced as a key technology of the downstream processing to achieve purity >99.9%, disadvantages like enormous feedstock pretreatment and limitations in capacity require other purification steps for initial product capture.

A promising technology allowing for an efficient capture of the target protein in an early stage of the downstream processing is the continuous extraction using an aqueous 2-phase system (ATPS).8, 9, 10 An ATPS is formed by either mixing 2 hydrophilic polymers (e.g., polyethylene glycol [PEG] and dextran) or a hydrophilic polymer and a kosmotropic salt (e.g., phosphate, sulfate, or citrate salt) exceeding a critical concentration.11, 12 APTSs in general exhibit a high biocompatibility, can easily be scaled up, and show high selectivity and recovery of the target protein.8, 11 The purification potential of ATPS for the extraction of biological compounds was first shown by Albertsson.13 Several other research groups later showed that the protein-partitioning behavior within salt–PEG ATPS could be selectively influenced using the displacement agent (DA) NaCl, allowing for a purification of IgG from the contaminants.9, 14, 15, 16 Further advances, substituting phosphate salts with the biodegradable citrate salt as a phase former enabled a purification of IgG from a hybridoma cell culture supernatant with a recovery yield of 99%.15 Nevertheless, the selection of DAs enabling a selective and efficient IgG purification based on aqueous 2-phase extraction (ATPE) remained almost exclusively empirical and is still limited to chlorides, especially to NaCl.

The use of a DA has a direct impact on the phase composition of the ATPS. Unfortunately, the mutual influence between NaCl and the phase formers of the salt–PEG ATPS is often unattended with only few publications dealing with this topic.17, 18, 19, 20, 21 These publications focus on the effect of NaCl on the ATPS binodal curves. It was shown that NaCl has a significant influence on the ATPS phase composition by steadily expanding the ATPS region with increasing NaCl concentration.17, 18, 19, 20, 21

To facilitate the selection of DAs and thus enable the application of the ATPE on an industrial level, suitable design tools have to be available allowing for the selection of suitable ATPS and DAs other than NaCl based on a minimal experimental screening effort.

Regarding ATPS (polymer–polymer, polymer–salt), including a DA in the absence of proteins, thermodynamic modeling is already feasible using the Electrolyte Perturbed-Chain Statistical Associating Fluid Theory (ePC-SAFT) equation of state.22, 23, 24 Moreover, the influence of different process parameters like temperature, type of phase former, and phase former concentration on the ATPS phase composition is accessible.

However, protein partitioning (e.g., IgG) within the ATPS cannot be calculated using ePC-SAFT due to their nonspherical shape, surface characteristics, size, and solvent content. As molecules are regarded as chains of spherical segments within ePC-SAFT, the complex physicochemical properties of the proteins including their interaction with the surrounding molecules cannot be accessed sufficiently. In our approach, these protein–protein and protein–solute interactions are captured by the second osmotic virial coefficients B22 and the cross virial coefficient B23 accessible via composition gradient multiangle light-scattering (CG-MALS) measurements.

In our previous publication,25 a hybrid shortcut calculation based on the cross virial coefficient B23 allowing for the estimation of the protein partition coefficients within salt (citrate, phosphate)–PEG2000 ATPS as a function of the NaCl and phase former concentration was successfully presented. In this work, this hybrid shortcut estimation is applied to citrate–PEG ATPS including the new DA lithium bromide (LiBr) to access the partition coefficients of IgG and human serum albumin (HSA).

LiBr allowed for a selective purification of the target protein IgG from the contaminant HSA (already identified as a representative impurity by the studies6, 15, 26) within a 14.5% (w/w) citrate and 8% (w/w) PEG2000 ATPS. The displacement suitability of LiBr with regard to IgG was investigated with respect to the mutual influence of LiBr and the phase formers (citrate, PEG) in combination with the cross virial coefficient B23. The influence of different amounts of LiBr on the tie-line slope (TLS) and on the 2-phase regime of the ATPS was determined and illustrated in phase diagrams.

Furthermore, as the precipitation of the target protein has to be avoided to circumvent a loss of product,27 the ability to estimate the precipitating effect of a DA on the target proteins within the ATPS is crucial. Protein precipitation and crystallization in solution is related to the protein–protein interactions which can be expressed by the second osmotic virial coefficient B22.28, 29 Protein precipitation and crystallization is generally favored with increasing attractive protein–protein interactions indicated by negative B22 values.30

In this work, the precipitating effect of LiBr on both proteins is estimated based on the second osmotic virial coefficient B22 also determined via CG-MALS measurements.

The results of this work elucidate that our hybrid shortcut calculation could be efficiently applied to ATPS containing the DA LiBr in addition to NaCl.

The osmotic virial coefficients B22 and B23 represent the protein–protein and protein–solute interactions, respectively and are key parameters for the protein-phase behavior within ATPS. In combination with our hybrid shortcut calculation and knowledge of the mutual influence between DAs and phase formers, a selection of suitable ATPS is provided in an early stage of downstream processing development.

Section snippets

Materials

The proteins of this work are IgG (CAS: 9007-83-4) and HSA (CAS: 70024-90-7), both purchased from Sigma-Aldrich (Steinheim, Germany). Potassium phosphate dibasic (K2HPO4, CAS: 7758-11-4), sodium phosphate monobasic (NaH2PO4, CAS: 7558-80-7), sodium chloride (NaCl, CAS: 7647-14-5), and trisodium citrate dihydrate (C6H5Na3O7·2H2O, CAS: 6132-04-3) were delivered from VWR BDH Prolabo (Leuven, Belgium). PEG (CAS: 25322-68-3) with a molecular weight of 2000 Da was obtained from Merck (Darmstadt,

Calculation of Partition Coefficients as a Function of B23

As shown in our previous publication,25 the protein partition coefficient, K2m, within salt-PEG ATPS can be calculated based on Equation 1:lnK2m=i=soluten2·M2·Mi·B23,i·(mimi)1000

Herein, M2 is the molar mass of the protein, and Mi the molar mass of the solutes of the ATPS (phase-forming components and possible DA). B23,i represents the cross virial coefficient between the protein and the solute i. The solute molality difference between the top (′) and bottom phase (″) is described by mimi

Second Osmotic Virial Coefficients B22

B22 values of IgG and HSA were measured at pH 7 and 298.15 K as function of the LiBr concentration to access the influence of LiBr on the protein–protein interactions in solution. B22 measurements were performed by CG-MALS as described previously. The results are illustrated in Figure 1a (IgG) and in Figure 1b (HSA). In addition, B22 values of both proteins as function of the phase formers (PEG2000 and citrate) and DA (NaCl) concentration, taken from previous publication,25 are included in

Estimating Precipitating Effect of LiBr Based on B22

According to the partitioning results already shown, LiBr is an efficient DA allowing for a selective purification of IgG from HSA. To enable an efficient and continuous ATPE within the downstream processing of pharmaceutical proteins, protein precipitation within the ATPS has to be avoided. To quantify the precipitating effect of LiBr on both proteins within the ATPS, estimation strategies have to be available. As mentioned in the Introduction section, protein precipitation in solution is

Conclusion

In this work, a new DA (LiBr) was found allowing for the selective purification of the target protein IgG from the impurity protein HSA within the 14.5% (w/w) citrate and 8% (w/w) PEG2000 ATPS at pH 7 and 298.15 K. Using 20% (w/w) LiBr allowed for an IgG yield YIgG of 74.6% and a HSA yield of YHSA of 3.6%. Thereby, the IgG precipitation was reduced from 8.8% within the initial ATPS (0% LiBr) to 3.4% applying 20% (w/w) LiBr. HSA was not precipitated in any ATPS investigated within this work.

Acknowledgments

The authors would like to thank Florian Worm for his help during the experiments. Financial support from the Ministry of Innovation, Science and Research of North Rhine-Westphalia in the frame of CLIB-Graduate Cluster Industrial Biotechnology, contract number: 314-108 001 08, is gratefully acknowledged by the authors. The authors like to acknowledge the financial support from the German Science Foundation (Leibniz Award to GS).

References (36)

  • A. Schmidt et al.

    Partitioning and purification of α-amylase in aqueous two-phase systems

    Enzyme Microb Technol

    (1994)
  • M. Herhut et al.

    Modeling and prediction of protein solubility using the second osmotic virial coefficient

    Fluid Phase Equilibria

    (2016)
  • F. Bonneté et al.

    Second virial coefficient: variations with lysozyme crystallization conditions

    J Cryst Growth

    (1999)
  • C. Kress et al.

    Osmotic virial coefficients as access to the protein partitioning in aqueous two-phase systems

    J Pharm Sci

    (2015)
  • J. Asenjo et al.

    Model for predicting the partition behaviour of proteins in aqueous two-phase systems

    J Chromatogr A

    (1994)
  • J.A. Asenjo et al.

    Aqueous two-phase systems for protein separation: a perspective

    J Chromatogr A

    (2011)
  • International Market Analysis Research and Consulting Group

    Global Biopharmaceutical Market Report & Forecast (2012-2017)

    (2012)
  • G. Walsh

    Biopharmaceutical benchmarks 2014

    Nat Biotechnol

    (2014)
  • Cited by (7)

    • Purification of chimeric amine dehydrogenase using a tailor-made aqueous two-phase system - A case study

      2021, Journal of Molecular Liquids
      Citation Excerpt :

      The partitioning of the biomolecule and impurities within the ATPS can be influenced by either the choice of the phase-former or by the addition of a displacement agent (e.g. neutral salt such as sodium chloride (NaCl) or lithium bromide (LiBr)) [11,12,24,25]. NaCl was used in several publications to displace the target biomolecule into the preferred phase (separation from impurities) obtaining biomolecule yields higher than 95% [11,12]. However, ATPE never made it into commercial processes due to a lack of knowledge about the selection of suitable ATPS [20,26], resulting in a high effort with respect to labor and time to identify appropriate ATPS for the purification of the target molecule.

    • Rapid identification of tailor-made aqueous two-phase systems for the extractive purification of high-value biomolecules

      2020, Journal of Molecular Liquids
      Citation Excerpt :

      However, purification of biomolecules using ATPE is only possible by selecting a suitable ATPS for the target biomolecule ensuring its physical stability and activity is retained, simultaneously avoiding aggregation/precipitation. In the last years, several academic research groups [11,17,20,22–24] investigated the ATPE of biomolecules (e.g. immunoglobulin G; IgG) within a salt-polyethylene glycol (PEG) ATPS and showed that a selective partitioning of the target biomolecule to the PEG-rich phase was possible, obtaining a yield higher than 95% [20,21]. However, methods applied were still mainly based on experiments, with the potential to identify ATPS that still perform below capabilities [17,25].

    • Solubilization of proteins in aqueous two-phase extraction through combinations of phase-formers and displacement agents

      2017, European Journal of Pharmaceutics and Biopharmaceutics
      Citation Excerpt :

      In order to overcome this drawback and estimate the protein partitioning, hybrid shortcut calculations have been applied recently [20,21]. Within previous works [20–22] the displacement ability of NaCl allowing for a selective immunoglobulin G (IgG) purification from human serum albumin (HSA) within polyethylene glycol (PEG)-salt (citrate, phosphate) ATPS could be efficiently estimated based on the cross virial coefficient B23. Furthermore the precipitating effect of one solute (phase-former and displacement agent) on IgG and HSA within salt (citrate, phosphate)-PEG2000 ATPS could be sufficiently estimated using the second osmotic virial coefficient B22.

    View all citing articles on Scopus

    This article contains supplementary material available from the authors by request or via the Internet at http://dx.doi.org/10.1016/j.xphs.2016.06.006.

    View full text