Scholarly article on topic 'Effects of salt-induced reversible self-association on the elution behavior of a monoclonal antibody in cation exchange chromatography'

Effects of salt-induced reversible self-association on the elution behavior of a monoclonal antibody in cation exchange chromatography Academic research paper on "Chemical sciences"

CC BY-NC-ND
0
0
Share paper
Academic journal
Journal of Chromatography A
OECD Field of science
Keywords
{"Monoclonal antibody (mAb)" / "Cation exchange chromatography (CEX)" / "Reversible self-association (RSA)" / "Peak splitting" / "Mobile-phase modifier."}

Abstract of research paper on Chemical sciences, author of scientific article — Haibin Luo, Nathaniel Macapagal, Kelcy Newell, Adrian Man, Arun Parupudi, et al.

Abstract Some monoclonal antibodies (mAbs) are reported to display concentration-dependent reversible self-association (RSA). There are multiple studies that investigate the effect of RSA on product characteristics such as viscosity, opalescence, phase separation and aggregation. This work investigates the effects of RSA on a bind-and-elute mode cation exchange chromatography (CEX) unit operation. We report a case study in which the RSA of an IgG2 (mAb X) resulted in significant peak splitting during salt gradient elution in CEX. Multiple factors including resin type, load challenge, residence time and gradient slope were evaluated and demonstrated little effect on the peak splitting of mAb X. It was determined that high NaCl concentrations in combination with high protein concentrations induced mAb X to form one RSA species that binds more strongly to the column, resulting in a large second elution peak. The finding of NaCl-induced RSA suggested that lower NaCl elution concentrations and different types of salts could mitigate RSA and thus eliminate peak splitting. Different salts were tested, showing that chaotropic salts such as CaCl2 reduced the second elution peak by inducing less RSA. The addition of a positively charged amino acid (such as 50mM histidine) into the CEX elution buffer resulted in elution at lower NaCl concentrations and also effectively reduced peak splitting. However, experiments that were intended to reduce salt concentration by increasing the elution buffer pH did not significantly mitigate peak splitting. This is because higher pH conditions also increase RSA. This work identifies salt-induced RSA as the cause of peak splitting of a mAb in CEX and also provides solutions to reduce the phenomenon.

Academic research paper on topic "Effects of salt-induced reversible self-association on the elution behavior of a monoclonal antibody in cation exchange chromatography"

ELSEVIER

Contents lists available at ScienceDirect

Journal of Chromatography A

journal homepage www.elsevier.com/locate/chroma

Effects of salt-induced reversible self-association on the elution behavior of a monoclonal antibody in cation exchange chromatography

Haibin Luoa, Nathaniel Macapagal3, Kelcy Newell3, Adrian Mana, Arun Parupudib Yiming Lib, Yuling Lia *

a Purification Process Sciences, Medimmune LLC, One Medimmune Way, Gaithersburg, MD 20878, USA b Analytical Biotechnology Science and Strategy, Medimmune LLC, One Medimmune Way, Gaithersburg, MD 20878, USA

ABSTRACT

Some monoclonal antibodies (mAbs) are reported to display concentration-dependent reversible self-association (RSA). There are multiple studies that investigate the effect of RSA on product characteristics such as viscosity, opalescence, phase separation and aggregation. This work investigates the effects of RSA on a bind-and-elute mode cation exchange chromatography (CEX) unit operation. We report a case study in which the RSA of an IgG2 (mAb X) resulted in significant peak splitting during salt gradient elution in CEX. Multiple factors including resin type, load challenge, residence time and gradient slope were evaluated and demonstrated little effect on the peak splitting of mAb X. It was determined that high NaCl concentrations in combination with high protein concentrations induced mAb X to form one RSA species that binds more strongly to the column, resulting in a large second elution peak. The finding of NaCl-induced RSA suggested that lower NaCl elution concentrations and different types of salts could mitigate RSA and thus eliminate peak splitting. Different salts were tested, showing that chaotropic salts such as CaCl2 reduced the second elution peak by inducing less RSA. The addition of a positively charged amino acid (such as 50 mM histidine) into the CEX elution buffer resulted in elution at lower NaCl concentrations and also effectively reduced peak splitting. However, experiments that were intended to reduce salt concentration by increasing the elution buffer pH did not significantly mitigate peak splitting. This is because higher pH conditions also increase RSA. This work identifies salt-induced RSA as the cause of peak splitting of a mAb in CEX and also provides solutions to reduce the phenomenon.

© 2014 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/3.0/).

CrossMark

ARTICLE INFO

Article history:

Received 5June 2014

Received in revised form 13 August 2014

Accepted 14 August 2014

Available online 20 August 2014

Keywords:

Monoclonal antibody (mAb) Cation exchange chromatography (CEX) Reversible self-association (RSA) Peak splitting Mobile-phase modifier.

1. Introduction

Monoclonal antibodies (mAbs) have emerged as a rapidly growing class of therapeutic since the mid-1990s [1]. More than 20 modern antibody-based therapeutic agents have been approved worldwide [2], and 500 additional products are currently in clinical development [1]. Although mAbs share certain structural similarities, development of commercially viable mAb pharmaceuticals is sometimes complicated by their unpredictable solution behaviors [3,4]. For example, some mAbs display reversible or irreversible self-association behavior [5-16]. Reversible self-association (RSA) is a unique solution property in which native, reversible oligomeric species are formed as a result of non-covalent intermolecular

* Corresponding author. Tel.: +301 398 4845. E-mail address: Liyu@Medimmune.com (Y. Li).

interactions [5]. Due to the multiple-domain structure and highly variable complementarity determining region (CDR) of mAbs, the nature of mAb RSA has been found to be often differing case-by-case. For instance, different mAb regions are responsible for the RSA behavior of different mAbs: Kanai et al. identified Fab-Fab interaction as the primary source of self-association [6], Bethea et al. implicated some exposed charge residues in the CDR [17] and Nishi and coworkers highlighted the Fc domains [8]. Many studies have showed that mAb RSA is concentration-dependent: with no or low propensity to form at low concentrations (<10 mg/mL), and enhanced attractive intermolecular interactions at high concentrations (>30 mg/mL), resulting in a high propensity to self-associate and form more stable self-associative species [5,11-13]. RSA can be related to viscosity [6,7,15,16], opalescence [9], phase separation [8] and aggregation [17]. Many studies focused on the impact of RSA on high-concentration mAb formulations, while the impact on purification chromatography receives little attention.

http://dx.doi.org/10.1016/j.chroma.2014.08.048

0021-9673/© 2014 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/3.0/).

Bind-and-elute mode cation exchange chromatography (CEX) is widely used as mAb purification polishing step due to its ability to remove product- and process-related impurities [14,18-20]. Protein biophysical properties can significantly affect the binding and eluting behavior of a protein in CEX. For example, Voitl et al. [14] observed pure human serum albumin bound to the CEX column in two different binding conformations. Gillespie et al. [21] reported that resin surface-mediated denaturation of an unstable mAb led to two distinct elution peaks during CEX chromatography. The strong cation exchangers currently in use can easily provide a dynamic binding capacity of >50 mg mAb/mL resin to enable more cost-efficient production processes [18]. Protein-protein interactions are strengthened at such high on-column protein concentrations [5], with the potential to induce RSA during CEX purification; this may lead to unexpected binding and eluting behaviors in CEX, resulting in protein loss or other issues. This study reports the RSA of an IgG2 (mAb X), which resulted in significant elution peak splitting on a strong CEX chromatography during linear salt gradient elution. In the elution step, NaCl in the elution buffer induced mAb X to form a reversibly self-associative species that bound strongly to the CEX column and was co-eluted with aggregates, resulting in low product yield for the elution product at the desired purity. Different factors were experimentally evaluated and the major factors affecting peak splitting and RSA were identified. As a result, approaches to mitigate RSA caused peak splitting in the CEX step were established.

2. Materials and methods

2.1. Chemicals, recombinant protein and cation exchanger resins

Buffers and cleaning solutions were prepared in house. All the chemicals used in this study were obtained fromJ.T. Baker (Phillipsburg, NJ, USA). mAb X was expressed in Chinese hamster ovary (CHO) cells and produced by Medimmune (Gaithersburg, MD, USA). The mAb X used in this study is a fully human IgG2 monoclonal antibody composed of two identical heavy chains and two identical light chains, with an overall molecular weight of 147 kDa. The light and heavy chains have Kyte-Doolitle hydropathicity index values at -0.447 and -0.348, respectively. It has 12 intra-chain and 6 interchain disulfide bonds. The experimental pi (isoelectric point) is 8.3 as determined by clEF (capillary isoelectric focusing). Unless noted otherwise, the protein feed material (the process material) was purified by two purification steps. The capture step was performed on MabSelect SuRe (GE Healthcare, Piscataway, NJ, USA) protein A affinity chromatography media. The MabSelect SuRe elution pool underwent a titration to pH 3.6 followed by neutralization to pH 4.5 then was further purified by a polishing CEX step. Capto SP ImpRes resin was obtained from GE Healthcare sciences (GE Healthcare, Piscataway, NJ, USA); Fractogel EMD SO3-(M) and COO-(M) resins were obtained from EMD Biosciences (Gibbstown, NJ, USA); POROS HS50 resin was purchased from Applied Biosystems (Grand Island, NY, USA).

2.2. Chromatography instrumentation and operations

Laboratory-scale chromatography experiments were performed using a GE Healthcare AKTA Explorer 100 using Unicorn software version 5.2 (GE Healthcare, Piscataway, NY, USA). CEX resin was packed into 1.15 cm inner diameter (ID) Vantage columns (Millipore, Billerica, MA, USA) to a bed height of approximately 12 cm and operated at a residence time of 5 min for different columns and runs. The CEX columns were pre-equilibrated with 3 column volumes (CV) of50 mM sodium acetate pH 4.5 (unless mentioned otherwise). The pH and conductivity of the column effluent

were monitored through built-in ÄKTA probes to ensure that the resin was properly equilibrated. Before the purified CEX pools for the first and second peaks were re-purified, the materials were buffer exchanged into 50 mM sodium acetate pH 4.5 prior to loading. After loading, the column was washed with 3 CV of 50 mM sodium acetate, pH 4.5. In the experiments at pH 5.5 and pH 6.0 using a NaCl gradient elution, the column was washed with 3 CV of 50 mM sodium acetate, pH 5.5 and 3 CV of 50 mM sodium phosphate pH 6.0, respectively. The protein bound on the column was eluted over a 20 CV linear salt gradient from 0 to 500 mM sodium chloride in 50 mM sodium acetate pH 4.5 unless mentioned otherwise. The elution was fractionated in half-CV fractions based on A280 collection criteria of >100 mAu. The absorbance of the protein was monitored at A280 by built-in ÄKTA probes. In-line pH and conductivity were also monitored for all test runs. All runs were carried out at room temperature (21-23 °C). Cleaning in-place was conducted using 3 CV in-house solution (50 mM sodium acetate, 1 M sodium chloride, pH 5.0) followed by 1 N sodium hydroxide. The columns were stored in 20% ethanol after each run. The protein concentration was measured using a NanoDrop 2000 (Thermo Fisher Scientific, Wilmington, DE. USA).

2.3. Analytical HPLC

Analytical size-exclusion chromatography (HPSEC) was performed using a TSK-GEL G3000SWXL column obtained from Tosoh Bioscience (King of Prussia, PA, USA) with an Agilent HPLC system HP1200 from Agilent Technologies (Santa Clara, CA, USA). The column was equilibrated at a flow rate of 0.5 mL/min with 78 mM sodium phosphate, 0.5 M sodium chloride, pH 7.4. All collected fractions were injected and eluted isocratically with the same buffer. The eluted protein was monitored by UV absorbance at 280 nm. Samples were 0.22 |im filtered and injected at a mass load of 0.25 mg. All injected volumes were less than 0.7% of the column volume. Samples used to obtain the levels of reversible self-association (RSA) were prepared in appropriate buffers at the desired protein concentrations. These samples were injected at a fixed volume of 20 | l but within a max load mass of 1 mg. The monomer, dimer, aggregate and RSA percentage were estimated by integrating the analytical chromatograms.

2.4. Dynamic light scattering (DLS) analysis

mAb X was buffer exchanged and prepared in the desired solution and concentration, then filtered with a 0.2 | m filter. Each sample was spun at 2000 rpm for 2 min to remove any potential air bubbles before DLS analysis. The hydrodynamic radius for mAb X was analyzed using a high-throughput 384-well plate DynaPro DLS instrument (Wyatt Technology, Santa Barbara, CA, USA) equipped with a 633 nm laser. The scattered light was monitored at 173° to the incident beam and autocorrelation functions were generated using a digital autocorrelator. The autocorrelation data were fitted to cumulant analysis as the samples showed very low polydisper-sity index (<0.1). The hydrodynamic radius was calculated using Stokes-Einstein equation [5].

3. Results and discussion

3.1. mAb X exhibited two large elution peaks on Capto SP ImpRes during NaCl linear gradient elution

Capto SP ImpRes (strong CEX) was used in the mAb X purification process to remove high-molecular-weight (HMW) aggregate and recover monomer. HP-SEC analysis indicated that the feed material had 90% monomer and 10% aggregate. The feed material was

Fig. 1. mAb X exhibited two large elution peaks during NaCl linear gradient elution from Capto SP ImpRes chromatography. (A) Elution profile of mAb X over a 20 CV linear elution gradient (0-500 mM NaCl). mAb X was loaded at a load challenge of30mg/mL resin. The equilibration and elution buffer used 50 mM NaAc, pH 4.5, as the buffer system. The eluate was fractionated on a half column volume basis after the elution began. Each fraction was analyzed by HP-SEC. The black solid line represents the A280 trace. The circles and triangles represent the concentrations of monomers and aggregates, respectively. The monomer purities for the first and second peaks were 98.2 and 72.3%, respectively. The second peak contained 25% of the total monomer from the feed. (B) Re-purification of the proteins collected in the first and second peak proteins led to a similar elution profile as the process material. The first and second peak proteins were loaded at load challenges of 18 and 11 mg/mL, respectively. Other conditions remained the same.

loaded at 30 g/L resin and was eluted using a 0-500 mM NaCl linear gradient elution (LGE). Surprisingly, mAb X exhibited two large elution peaks (Fig. 1A, solid curve) rather than a main peak followed by a small one as expected. The eluate was collected into

half column volume fractions. Each fraction was analyzed to obtain protein concentration and monomer purity. The first elution peak contained mainly monomers as expected, while the large second eluting peak contained high portion of monomer (72%) (Fig. 1A, circles). In the second elution peak, the monomer was eluted in the early part; aggregates were eluted in the later part (Fig. 1A, triangles) and only contributed to 28% of the second peak. Since this unexpected large second elution peak included a large amount of monomer, separation by peak cutting resulted in significantly lower yield for the product with the desired monomer purity.

These results led to a question as to whether the monomer observed in the first peak was different than that observed in the second peak. If they are different species, re-purification on this column should result in different elution profiles. To test this hypothesis, the first and second peaks were separately pooled, conditioned and re-purified by the same column using the same conditions. Their elution profiles are shown in Fig. 1B. Interestingly, the re-purification of the proteins from the first peak caused re-distribution into two elution peaks (Fig. 1B, dash curve). Furthermore, the re-purification of the proteins from the second peak in Fig. 1A also resulted in peak distribution (Fig. 1B, dash-dot curve). Further re-purification ofthese eluate fractions consistently demonstrated similar two-elution-peak pattern (data not shown). These results indicated that monomers in the first and second elu-tion peaks are the same species and the cause for the peak splitting is likely reversible in nature.

3.2. mAb X exhibited reversible self-association (RSA) under the elution condition

Dimitrova and Mody recently reported that this mAb X displayed reversible self-association (RSA) under some formulation conditions and the RSA could be detected by HP-SEC [22]. The reversibility of the peak-splitting inspired us to investigate RSA as a possible cause. The load challenge in Fig. 1A is 30 mg/mL; and the local concentration of mAb X is expected to be higher on the Capto SP ImpRes column. To explore if mAb X forms RSA under the conditions experienced on the column, mAb X was prepared separately in equilibration buffer (50 mM NaAc, pH4.5) and in elution buffer (50 mM NaAc, 500 mM NaCl, pH 4.5) at multiple protein concentrations (10,20,30,40 and 50 mg/mL). These samples were then analyzed by HP-SEC. In the equilibration buffer which does not contain NaCl, the five samples yielded similar sizing profiles (Fig. 2A). Similar percentages of monomers, dimers and high-molecular-weight (HMW) aggregates were observed for the five samples. In the elution buffer containing 500 mM NaCl, mAb X at 10mg/ml exhibited a similar sizing profile. However, at concentrations above 20 mg/mL, a new peak appeared between the monomer peak and the dimer peak (Fig. 2B). This new peak represents reversible self-association (RSA) species of mAb X [22]. Meanwhile, the size of the new peak increased with protein concentration, suggesting that

Table 1

Effects of salt species on RSA and the elution profile.

Salt species RSA level in 150 mM cation (%) Monomer amount in the second peak (%)a Salt concentration for eluting the first peak (mM)b RSA level in salt concentration for eluting the first peak (%)

Potassium chloride 27.6 29.3 178 26.5

Sodium sulfate 20.3 22.8 88 16.7

Sodium chloride 24.4 24.8 151 19.1

Sodium acetate 20.1 19.9 169 15.6

Sodium citrate 19.0 18.2 49 17.1

Arginine hydrochloride 11.8 10.7 146 13.3

Calcium chloride 11.2 7.3 40 7.6

Magnesium chloride 14.4 8.5 38 4.6

a Monomers in the second peak was reported as the percentage of the monomers in the load material. b The salt concentration obtained at the max height of the first elution peak.

Monomer -50 mg/ml

40 mg/ml

-30 mg/ml

1 -20 mg/ml

I -10 mg/ml

HMW III/ 1

agg- Dimer 1 I

■ ] i 1 i 1 i 1 i I i I i

Time (min)

Monomer

-50 mg/ml

RSAX 1 - -40 mg/ml

— 30 mg/ml

20 mg/ml

1 \ lit -10 mg/ml

Dimer 1 \l a 1

HMW r J \ 1

I////71

8 9 Time (min)

:n 2 rs es

10 8 6 4 2 0

In Equilibration Buffer I In Elution Buffer

Il 11 I

mAb X Concentration (ing/mL)

Fig. 2. mAb X exhibited reversible self-association (RSA) in the elution buffer. HP-SEC profiles of a range of mAb X concentrations in the equilibration buffer (A) and the elution buffer (B). (C) Hydrodynamic radii of mAb X at different concentrations in the equilibration buffer and the elution buffer. Hydrodynamic radius was calculated using Stokes-Einstein equation based on dynamic light scattering analysis. The compositions of the equilibration and elution buffers were 50 mM NaAc, pH 4.5, and 50 mM NaAc, 500 mM NaCl, pH 4.5, respectively.

the RSA of mAb X in the elution buffer is protein concentration-dependent. This is similar to many other reported RSA observations [5,11-13]. Dynamic light scattering was also utilized to study association behavior of mAb X in the equilibration buffer and elution buffer, respectively. As showed in Fig. 2C, mAb exhibited greater hydrodynamic radius in the elution buffer than in the equilibration buffer; meanwhile, hydrodynamic radius of mAb X also gradually

increased with protein concentration. These results clearly indicate the presence of RSA in the elution buffer. Since the major difference between the equilibration buffer and elution buffer is that the latter contains 500 mM NaCl, the RSA of mAb X is induced by NaCl. Therefore, mAb X likely formed RSA during the elution step on the column. RSA has been found to play a role in peak splitting in anion exchange chromatography for several proteins such as apolipopro-tein [19,23] and bovine serum albumin [24], and therefore, this NaCl-induced RSA may have resulted in the peak splitting of mAb X on the Capto SP ImpRes column.

3.3. The role of elution salt species and a plausible mechanism

Different salt species are reported to affect self-association behaviors of mAbs in solution [27-29]. Therefore, if the peak splitting of mAb X in CEX is associated with formation of RSA, using a salt that induces less RSA should mitigate peak splitting. To test this hypothesis, the RSA levels of mAb X in seven additional salts (chosen from the Holfmeister series and included komostropes and chaotropes) were tested and compared. We noticed that the monomer species eluted at 150 mM NaCl in Fig. 1A; thus for simplicity the same concentration of 150 mM for the cation was used for different salts in evaluating their effect on RSA. Compared to the absence of salt, all tested salts induced RSA (Table 1), but different salts resulted in different levels of RSA. Chaotropic salts, such as calcium chloride (CaCl2) and magnesium chloride (MgCl2), led to less RSA. The seven salts were also used to elute mAb X on Capto SP ImpRes and evaluated their effects on elution profile. The results indicated that salt species has a significant impact on the peak splitting of mAb X. The results are summarized in Table 1 by reporting the level of monomer in the second elution peak. For instance, potassium chloride (KCl) resulted in the greatest amount of monomer in the second elution peak, while CaCl2 resulted in the smallest amount of monomer in the second elution peak and therefore the largest first peak.

Interestingly, the impact of salt species on RSA formation and peak splitting shared a similar trend. This trend is also consistent with their order in the Holfmeister series, i.e. chaotropic salt results in less RSA, as well as less monomer in the second elution peak. It was noted that the concentration to trigger elution differed from salt to salt. Therefore, the RSA levels under these appropriate salt concentrations were also determined (Table 1) to better demonstrate their correlation. A plotting of the RSA levels against second peak sizes for different salts showed a positive linear correlation (Fig. 3), i.e., the more RSA, the greater the second elution peak. These results demonstrated the peak splitting observed above is likely resulted from the formation of RSA species during elution phase. A plausible explanation for the observations is: (1) the elu-tion salt induces the formation of RSA species; (2) like aggregates, the RSA species binds more strongly to the column than monomer and remains bound on the column when monomer is eluted; (3) the RSA species is eluted at a slightly higher salt concentration (likely close to that of aggregate); (4) the new species dissociated into monomer (due to dilution) in the mobile phase.

3.4. The role of elution pH

As discussed above, RSA was induced by the presence of NaCl in the elution buffer. Therefore, an elution condition that results in the elution of bound proteins at a lower NaCl concentration should mitigate RSA and thus reduce the size of the second elution peak. One way to achieve this is eluting protein by pH gradient. A pH gradient from pH 4.5 (50 mM sodium acetate) to pH 8.0 (50 mM sodium phosphate) was tested. However, this pH gradient elu-tion also resulted in significant peak splitting (three elution peaks, Fig. 4A). The first and second peaks mainly contained monomer, 100

Fig. 3. Amount of monomer in the second elution peak plotted against the level of RSA in the corresponding elution buffers show direct correlation. The monomer amount was expressed as a percentage of the total monomer in the load. The RSA levels were obtained by integrating the HP-SEC curves. See more details in Table 1.

and 98%, respectively. The tiny third eluting peak mainly contained HMW aggregates. Another way of reducing elution NaCl concentration is NaCl gradient elution at a higher pH (such as pH 5.5 and pH 6.0, respectively, Fig. 4B). As expected, at the higher elution pH levels, mAb X monomer was eluted at lower salt concentrations. However, the peak splitting was still severe by giving three elution peaks. Although the aggregate peak was smaller, a new shoulder peak that mainly contained monomer appeared after the first elution peak. While these conditions resulted in lower salt concentrations needed for the occurrence of protein elution, peak splitting was not mitigated. One possibility is that pH also affects RSA formation. mAb X RSA at different pH levels with or without NaCl (Fig. 4C) showed the mAb X RSA is pH-dependent. At pH levels close to its pi, mAb X is less positively charged and thus inter-molecular electrostatic repulsion is weaker. The results that mAb X tended to form more RSA at pHs closer to its pi explained why either the pH gradient elution or the salt gradient elution at higher pH were not effective to mitigate peak splitting.

3.5. The role of mobile-phase modifier

Aside from higher pH for elution, the addition of a positively charged amino acid (as mobile phase modifier) into the elution buffer should enable protein elution at lower NaCl concentration by modifying protein-CEX interactions [30]. Therefore, we carried out several Capto SP ImpRes runs with 50 mM arginine, histidine or lysine as additives in the buffers (maintaining a constant buffer pH). NaCl was used as the elution salt. As shown in Fig. 5A, the presence of these amino acids resulted in protein elution at lower NaCl concentrations and also reduced the second elution peak, compared to the runs without additive. The magnitude of these effects ranked as follows: His>Lys>Arg. We tested whether these amino acids exerted their effects by inhibiting RSA as previously reported in literature [31,32]. Fig. 5B showed that His, Lys or Arg alone could not effectively inhibit RSA of mAb X at 50 mM concentration. As expected, lower RSA levels were observed in the elution conditions with amino acid (Fig. 5C) mainly because the elution occurred at lower NaCl concentrations. The effects of these amino acids on reducing peak splitting were concentration-dependent. Monomer/aggregate separation ability of Capto SP ImpRes was not significantly affected when the concentration is below 100 mM. Histidine exhibited the greatest effect on reducing peak splitting among these amino acids at a concentration of 50 mM. Histidine is

Fig. 4. Neither pH gradient elution nor NaCl gradient elution from Capto SP ImpRes chromatography at higher pH reduced the second elution peak because increasing pH increased mAb X RSA. (A) Elution profile of mAb X for pH linear gradient elution. The pH gradient was achieved by mixing 50 mM NaAc, pH 4.5 and 50 mM sodium phosphate, pH 8.0. The monomer purities forthe first, second and third peaks were 100,98.3 and 23.5%, respectively. (B) Elution profile of mAbX by NaCl lineargradient elution at pH 4.5,5.5 and 6.0. The buffers consisted of 50 mM NaAc pH for pH 4.5 and 5.5, and 50 mM sodium phosphate for pH 6.0. The NaCl gradient extended from 0 to 500 mM NaCl over 20 column volumes and other elution conditions remained the same. Monomer purities forthe elution peaks: pH 6.0,98.3% (first peak), 90.8% (second peak) and 32.1% (third peak); pH 5.5,97.5% (first peak), 82.7% (second peak) and 61.2% (third peak); pH 4.5,98.2% (first peak) and 72.3% (second peak). (C) The effects ofpH and NaCl on RSA. All samples were prepared at 50 mg/mL in the corresponding solutions.

Fig. 5. Addition of positively charged amino acids as mobile-phase modifiers enables elution at lower NaCl concentration with a lower level of RSA. (A) The elution profile of mAb X on Capto SP ImpRes by NaCl LGE with or without mobile-phase modifiers. All the elution conditions were buffered with 50 mM NaAc, pH 4.5. The LGE was from 0 to 500 mM NaCl over 20 CV. Monomer purity for each peak: no modifier, 98.2% (first peak) and 72.3% (second peak); Arg, 97.6% (first peak), 64.1% (second peak); Lys, 96.4% (first peak), 53.9% (second peak); His, 95.3% (first peak), 44.0% (second peak). (B) RSA levels of mAb X with or without modifier in 50 mM NaAc, 150 mM NaCl, pH 4.5. Arg, Lys and His did not inhibit RSA formation. (C) RSA levels of mAb X under the elution conditions. In (B) and (C), mAb X was prepared at 50 mg/mL in the corresponding solutions. The RSA levels were obtained by HP-SEC. Decreasing the NaCl concentration necessary for protein elution was the key to these modifiers' reduction of RSA.

Fig. 6. Elution profiles of mAb X from different CEX columns. All tests were carried out under the same conditions: a 3 CV equilibration of 50 mM NaAc, pH 4.5; 30 mg/mL as the load mass; a 3 CV wash of 50 mM NaAc, pH 4.5; elution with a 20 CV 0-500 mM NaCl LGE in 50 mM NaAc, pH 4.5. The monomer purity for each peak: Fractogel SO3-(M), 97.5% (first peak) and 55.7% (second peak); Poros HS50, 96.9% (first peak) and 63.7% (second peak); Capto SP ImpRes, 98.2% (first peak) and 72.3% (second peak); Fractogel COO-(M), 90.3%.

used as an excipient in many protein formulations due to its stabilizing effects [33], making histidine a good candidate for mitigating RSA-induced peak splitting on CEX.

3.6. The role of CEX resin type

These results demonstrated that the peak splitting of mAb X on Capto SP ImpRes is associated with it RSA behavior. To evaluate if CEX resin-type affects mAb X's peak splitting, two additional strong cation exchangers (Fractogel SO3-(M) and Poros HS50) and a weak cation exchanger (Fractogel COO-(M)) were tested using similar conditions. As shown in Fig. 6, mAb X also exhibited peak splitting on Fractogel SO3-(M) and Poros HS50, and the two second peaks contained large amounts of monomer. MAb X was eluted at a slightly lower NaCl from Fractogel SO3-(M) and Poros HS50, resulting in slightly smaller second elution peaks, consistent with our findings above. On Fractogel COO-(M), mAb X exhibited an extremely broad peak, which was due to RSA of mAb X. Therefore, RSA of mAb X significantly affected its elution behavior not only on strong cation exchangers, but also on weak cation exchangers. While screening of CEX resins for mAb X purification may be helpful for other purification needs, it did not offer any effective solutions to the peak splitting issue.

3.7. The role of load challenge

As RSA depends on protein concentration, lower load challenges should reduce the second elution peak. Lower load challenges were tested and the results showed load challenge had no significant impact on the peak splitting (Fig. 7A). The profile of two elution peaks remained similar for load challenges as low as 3.3 mg/mL resin. Therefore, using lower load challenge does not mitigate the peak splitting of mAb X on Capto SP ImpRes. One explanation is that mAb X did not equally distribute on the column. The resins in the upper layer of the column absorbed more proteins until they are saturated than the resins in the bottom layer. Since the loading condition is the same, the local mAb X concentrations on the column are likely similar under these tested load challenges.

Fig. 7. Evaluation ofthe effects of load challenge (A), residence time (B) and gradient steepness (C) on elution profile in Capto SP ImpRes. All tests used the same load material and solutions: 50 mM NaAc, pH 4.5, for equilibration and wash; 50 mM NaAc, 500 mM NaCl, pH 4.5, as the 100% B solution for LGE. Different load challenges, residence time and gradient steepness were achieved by altering injection volume, flow rate and linear gradient elution length, respectively.

3.8. The role of residence time

To test whether residence time has an effect, the residence time was extended from 5 min to 25 min for the loading and elution steps, respectively. As shown in Fig. 7B, extending the residence time of the loading step to 25 min did not have a significant effect. In contrast, extending the residence time of the elution step to 25 min

lead to a slightly larger second elution peak. Therefore, residence time during elution has some impact and should be controlled as needed.

3.9. The role of elution gradient steepness

The effects of the gradient slope were also evaluated by comparing 20 CV and 40 CV NaCl linear gradient elutions (the residence time remained as 5 min). As shown in Fig. 7C, these two gradients led to similar peak distribution, demonstrating that the elution gradient steepness had no significant effect on the peak splitting of mAb X.

4. Conclusion

This study reports the peak splitting for an IgG2 (mAb X) in CEX is due to RSA induced by NaCl contained in the elution buffer. The RSA formation of mAb X was proportional to the NaCl concentration with higher NaCl resulting in more RSA. This makes the RSA of mAb X novel in its type because most reports indicate that mAb RSA is suppressed by high NaCl concentration [6,11,25,26]. Our hypothesis for the observations of mAb X on Capto SP ImpRes is that NaCl in the elution buffer induces the formation of reversibly self-associated species; the RSA species binds more strongly to the CEX column, resulting in the large second elution peak; after being eluted, the reversibly self-associated species dissociate into monomers possibly due to dilution in the mobile phase.

We investigated the effects of a variety of experimental conditions on the peak splitting behavior of mAb X. Salt species has a significant effect on the peak splitting on CEX of mAb X as well as on RSA formation. Chaotropes such as CaCl2 resulted in smaller second elution peaks by inducing less RSA. Because mAb X exhibited more RSA at higher pH, pH gradient elution or salt gradient elution at higher pH did not mitigate peak splitting. The addition of positively charged amino acids into CEX buffers led to elution at lower salt concentration and resulted in smaller second peaks. The modifiers by themselves did not significantly inhibit RSA at the tested concentration in solution. Instead, their major contribution was likely through modification of protein-CEX resin interactions, shifting protein elution to lower NaCl concentration which resulted in less RSA upon elution. Among the amino acids tested here, histidine showed the greatest effect. Histidine is also used as an excipient in many protein formulations due to its protein stabilizing effects, making it a good candidate for the mitigation of peak splitting on CEX. Furthermore, the RSA of mAb X not only resulted in peak splitting on Capto SP ImpRes, but also on other strong cation exchangers such as Fractogel SO3-(M) and Poros HS50. The RSA of mAB X led to an extremely broad elution peak on the weak cation exchangers Fractogel COO-(M). Therefore, screening different CEX resins for mAb X purification may not be a good solution for this problem. Residence time, load challenge and gradient slope did not have any significant effect.

Those RSA behaviors that are suppressed by salt are believed to be driven by electrostatic attractions [6,11,25,26]. In contrast, based on our results, the salt-induced RSA of mAb X is likely driven by hydrophobic interactions although a convincing argument and mechanistic understanding would need further investigation. A characterization study for mAb X indicates that mAb X has high surface hydrophobicity that likely plays a role in its RSA behavior (unpublished results) [34]. A mAb exhibiting hydrophobic-interaction-driven RSA behaviors can have abnormal CEX elution profile complicating its process development. Our work sheds light on understanding the impacts of RSA on elution behaviors during CEX chromatography and provides solutions to reduce the negative impact. Furthermore, the ideas identified here to

prevent mAb X RSA formation have potential to solve RSA relevant issues.

Acknowledgements

We would like to thank Denial Callahan and Justin Weaver from the Medimmune Purification Process Sciences Department and Sophia Levitskaya from the Analytical Biotechnology Department for their contributions to this work. We also thank William Wang, Kripa Ram and Alan Hunter for critical review and suggestions.

References

[1] G. Walsh, Modern antibody-based therapeutics, Biopharmaceutical 17 (2004) 18-22.

[2] P. Chames, M. Van Regenmortel, E. Weiss, D. Baty, Therapeutic antibodies: successes, limitations and hopes for the future, Br. J. Pharmacol. 157 (2009) 220-233.

[3] W. Wang, S. Singh, D.L. Zeng, K. King, S. Nema, Antibody structure, instability, and formulation, J. Pharm. Sci. 96 (2007) 1-26.

[4] S.J. Shire, Z. Shahrokh, J. Liu, Challenges in the development of high protein concentration formulations, J. Pharm. Sci. 93 (2004) 1390-1402.

[5] R. Esfandiary, D.B. Hayes, A. Parupudi, J. Casas-Finet, S. Bai, H.S. Samra, A.U. Shah, H.A. Sathish, Systematic multitechnique approach for detection and characterization of reversible self-association during formulation development of therapeutic antibodies, J. Pharm. Sci. 102 (2013) 62-72.

[6] S. Kanai, J. Liu, T.W. Patapoff, S.J. Shire, Reversible self-association of a concentrated monoclonal antibody solution mediated by Fab-Fab interaction that impacts solution viscosity, J. Pharm. Sci. 97 (2008) 4219-4227.

[7] J. Liu, M.D. Nguyen, J.D. Andya, S.J. Shire, Reversible self-association increases the viscosity of a concentrated monoclonal antibody in aqueous solution, J. Pharm. Sci. 94 (2005) 1928-1940.

[8] H. Nishi, M. Miyajima, H. Nakagami, M. Noda, S. Uchiyama, K. Fukui, Phase separation ofan IgG1 antibody solution under a low ionic strength condition, Pharm. Res. 27 (2010) 1348-1360.

[9] B.A. Salinas, H.A. Sathish, S.M. Bishop, N. Harn, J.F. Carpenter, T.W. Randolph, Understanding and modulating opalescence and viscosity in a monoclonal antibody formulation, J. Pharm. Sci. 99 (2010) 82-93.

[10] C. Simm, C.H. Luan, E. Weiss, T. O'Halloran, High-throughput screen for identifying small molecules that target fungal zinc homeostasis, PLoS One 6 (2011) 1-9.

[11] S.V. Sule, J.K. Cheung, V. Antochshuk, A.S. Bhalla, C. Narasimhan, S. Blaisdell, M. Shameem, P.M. Tessier, Solution pH that minimizes self-association of three monoclonal antibodies is strongly dependent on ionic strength, Mol. Pharm. 9 (2012) 744-751.

[12] S.V. Sule, C.D. Dickinson, J. Lu, C.K. Chow, P.M. Tessier, Rapid analysis of antibody self-association in complex mixtures using immunogold conjugates, Mol. Pharm. 10(2013)1322-1331.

[13] S.V. Sule, M. Sukumar, W.F.t. Weiss, A.M. Marcelino-Cruz, T. Sample, P.M. Tessier, High-throughput analysis of concentration-dependent antibody self-association, Biophys. J. 101 (2011) 1749-1757.

[14] A. Voitl, A. Butte, M. Morbidelli, Behavior of human serum albumin on strong cation exchange resins: I. experimental analysis, J. Chromatogr. A 1217 (2010) 5484-5491.

[15] S. Yadav, J. Liu, S.J. Shire, D.S. Kalonia, Specific interactions in high concentration antibody solutions resulting in high viscosity, J. Pharm. Sci. 99 (2010) 1152-1168.

[16] I.E. Zarraga, R. Taing, J. Zarzar, J. Luoma, J. Hsiung, A. Patel, F. Lim, High shear rheology and anisotropy in concentrated solutions of monoclonal antibodies, J. Pharm. Sci. 102 (2013) 2538-2549.

[17] D. Bethea,S.J.Wu,J. Luo, L. Hyun, E.R. Lacy,A.Teplyakov, S.A.Jacobs, K.T. O'Neil, G.L. Gilliland, Y. Feng, Mechanisms of self-association of a human monoclonal antibody CNT0607, Protein Eng. Des. Sel. 25 (2012) 531-537.

[18] P. Gagnon, Technology trends in antibody purification, J. Chromatogr. A 1221 (2012)57-70.

[19] A.K. Hunter, E.J. Suda, J.T. Herberg, K.E. Thomas, R.E. Shell, M.E. Gustafson, S.V. Ho, Separation of recombinant apolipoprotein A-l(Milano) modified forms and aggregates in an industrial ion-exchange chromatography unit operation, J. Chromatogr. A 1204 (2008) 42-47.

[20] P.A. Marichal-Gallardo, M.M. Alvarez, State-of-the-art in downstream processing of monoclonal antibodies: process trends in design and validation, Biotechnol. Prog. 28 (2012) 899-916.

[21] R. Gillespie, T. Nguyen, S. Macneil, L. Jones, S. Crampton, S. Vunnum, Cation exchange surface-mediated denaturation ofan aglycosylated immunoglobulin (lgG1), J. Chromatogr. A 1251 (2012) 101-110.

[22] M. Dimitrova, N. Mody, Antibody formulations, PCT/US2011/042838 (2012) 1-31.

[23] T.E. Bankston, G. Carta, Apolipoprotein A-l(Milano) anion exchange chromatography: self association and adsorption equilibrium, Biotechnol. J. 5 (2010) 1028-1039.

[24] A.K. Hunter, G. Carta, Effects of bovine serum albumin heterogeneity on frontal analysis with anion-exchange media, J. Chromatogr. A 937 (2001) 13-19.

[25] T.M. Scherer, J. Liu, S.J. Shire, A.P. Minton, Intermolecular interactions of lgG1 monoclonal antibodies at high concentrations characterized by light scattering, J. Phys. Chem. B 114 (2010) 12948-12957.

[26] T.M. Scherer, Cosolute effects on the chemical potential and interactions of an lgG1 monoclonal antibody at high concentrations, J. Phys. Chem. B 117 (2013) 2254-2266.

[27] P. Arosio, B. Jaquet, H. Wu, M. Morbidelli, On the role of salt type and concentration on the stability behavior of a monoclonal antibody solution, Biophys. Chem. 168-169(2012) 19-27.

[28] W.G. Lilyestrom, S.J. Shire, T.M. Scherer, lnfluence ofthe cosolute environment on lgG solution structure analyzed by small-angle X-ray scattering, J. Phys. Chem. B 116 (2012) 9611-9618.

[29] R. Majumdar, P. Manikwar, J.M. Hickey, H.S. Samra, H.A. Sathish, S.M. Bishop, C.R. Middaugh, D.B. Volkin, D.D. Weis, Effects of salts from the Hofmeister series on the conformational stability, aggregation propensity, and local flexibility of an lgg1 monoclonal antibody, Biochemistry 52 (19) (2013) 3376-3389.

[30] M.A. Holstein, S. Parimal, S.A. McCallum, S.M. Cramer, Mobile phase modifier effects in multimodal cation exchange chromatography, Biotechnol. Bioeng. 109(2012) 176-186.

[31] T. Arakawa, D. Ejima, K. Tsumoto, N. Obeyama, Y. Tanaka, Y. Kita, S.N. Timasheff, Suppression of protein interactions by arginine: a proposed mechanism ofthe arginine effects, Biophys. Chem. 127 (2007) 1-8.

[32] T. Arakawa, K Tsumoto, K. Nagase, D. Ejima, The effects of arginine on protein binding and elution in hydrophobic interaction and ion-exchange chromatography, Protein Expr. Purif. 54 (2007) 110-116.

[33] A. Al-Hussein, H. Gieseler, lnvestigation of histidine stabilizing effects on LDH during freeze-drying, J. Pharm. Sci. 102 (3) (2013) 813-826.

[34] S. Levitskaya, J. Hope, J. Casas-Finet, A. Polozova, Z. Wei, P. Cash, Detection and analytical characterization ofreversible self-association ofan lgG2 monoclonal antibody, manuscript in preparation.