Multimodal charge-induction chromatography for antibody purification

G Model

ARTICLE IN PRESS

CHROMA-357149; No. of Pages 7

Journal of Chromatography A, xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Multimodal charge-induction chromatography for antibody purification Hong-Fei Tong a , Dong-Qiang Lin a,∗ , Wen-Ning Chu a , Qi-Lei Zhang a , Dong Gao b , Rong-Zhu Wang a , Shan-Jing Yao a a Key Laboratory of Biomass Chemical Engineering of Ministry of Education, College of Chemical and Biological Engineering, Zhejiang University, Hangzhou 310027, China b Key Laboratory of Synthetic and Natural Functional Molecular Chemistry of Ministry of Education, Institute of Modern Separation Science, Northwest University, Shaanxi Key Laboratory of Modern Separation Science, Xi’an 710068, China

a r t i c l e

i n f o

Article history: Received 6 October 2015 Received in revised form 9 December 2015 Accepted 17 December 2015 Available online xxx Keywords: Hydrophobic charge-induction chromatography Ligand design Antibody Purification

a b s t r a c t Hydrophobic charge-induction chromatography (HCIC) has advantages of high capacity, salt-tolerance and convenient pH-controlled elution. However, the binding specificity might be improved with multimodal molecular interactions. New ligand W-ABI that combining tryptophan and 5-amino-benzimidazole was designed with the concept of mutimodal charge-induction chromatography (MCIC). The indole and benzimidazole groups of the ligand could provide orientated mutimodal binding to target IgG under neutral pH, while the imidazole groups could induce the electrostatic repulsion forces for efficient elution under acidic pH. W-ABI ligand was coupled successfully onto agarose gel, and IgG adsorption behaviors were investigated. High affinity to IgG was found with the saturated adsorption capacity of 70.4 mg/ml at pH 7, and the flow rate of mobile phase showed little impact on the dynamic binding capacity. In addition, efficient elution could be achieved at mild acidic pH with high recovery. Two separation cases (IgG separation from albumin containing feedstock and monoclonal antibody purification from cell culture supernatant) were verified with high purity and recovery. In general, MCIC with the speciallydesigned ligand is an expanding of HCIC with improved adsorption selectivity, which would be a potential alternative to Protein A-based capture for the cost-effective purification of antibodies. © 2015 Elsevier B.V. All rights reserved.

1. Introduction The increasing market potential of monoclonal antibodies (mAbs) has promoted the revolution of antibody production techniques [1,2]. With the rapid advancement of upstream processing, antibody titers have been improved to 5–13 g/L [3,4], which leads the downstream processing to be a bottleneck of mAb production [5]. The development of more economical bioseparation technologies with high efficiency has been one of research hotspots. Protein A affinity chromatography is the most widespread technique to capture antibody for large-scale production [6]. However, there are still some limitations, such as high cost, ligand leakage, and harsh elution conditions [7]. Researchers tried to develop new ligands with low cost and high affinity as alternatives to Protein A. Naik et al [8]. immobilized tryptophan to separate IgG, but

∗ Corresponding author at: College of Chemical and Biological Engineering, Zhejiang University, Hangzhou 310027, China. Fax: +86 571 87951982. E-mail address: [email protected] (D.-Q. Lin).

the addition of PEG 600 in the buffer was required to improve adsorption selectivity. Histidine was used as a ligand to bind IgG at neutral pH and elute by salt addition [9–11]. However, due to the relatively simple structure, the binding specificity was limited. Moreover, some biomimetic synthetic ligands with more complex structure were developed, such as triazine ligands [12–15], peptide mimetic ligands [16–19] and new synthetic ligands based on the multi-component Ugi reaction [20–22]. A series of linear Fcbinding hexamer peptides and cyclic peptides were also reported [23–27]. D2 AAG ligand composed by amino acids and a synthetic aromatic acid was designed to improve mAb binding selectivity [28]. The specificities of these peptide ligands are generally comparable with Protein A, but they are also expensive and complicated to synthesize. Furthermore, high affinity leads to the difficulty on elution that often needs harsh pH or special additives. Therefore, novel ligands with both perfect binding and mild elution are still under development. Hydrophobic charge-induction chromatography (HCIC) was introduced in 1998 as a novel technology for protein purification, especially for antibody [29]. MEP HyperCel with 4-mercaptoethyl-

http://dx.doi.org/10.1016/j.chroma.2015.12.047 0021-9673/© 2015 Elsevier B.V. All rights reserved.

Please cite this article in press as: H.-F. Tong, et al., Multimodal charge-induction chromatography for antibody purification, J. Chromatogr. A (2015), http://dx.doi.org/10.1016/j.chroma.2015.12.047

G Model CHROMA-357149; No. of Pages 7

ARTICLE IN PRESS H.-F. Tong et al. / J. Chromatogr. A xxx (2015) xxx–xxx

2

pyridine (MEP) as ligand is the typical HCIC resin, which has been used for antibody separation [30–32]. Other HCIC ligands, such as mercapto-methyl-imidazole, mercapto-benzimidazole, 2mercaptoimidazole, histamine, and 5-aminoindole, have also been reported in literatures [33–36]. Generally, HCIC resins could bind target proteins through hydrophobic interactions at neutral pH, and elute proteins effectively by electrostatic repulsion between protein and charged ligands at acidic pH. In addition, the salt-tolerant adsorption property can offer the flexibility in feedstock preparation without the need of dilution or salt addition [31]. However, due to simple chemical structure with only one functional group, the binding selectivity of these ligands was limited. Therefore, better ligand design with multimodal molecular interactions to improve the specificity and expand HCIC to multimodal charge-induction chromatography (MCIC) is necessary. The ideal MCIC ligand should be highly specific with high capacity, and the elution process could be achieved easily under weak acidic conditions. In our previous work [37], the recognition mode between IgG-Fc domain and some natural Fc-specific ligands (e.g. Protein A, Protein G, and Fc receptor) were evaluated with molecular simulation. The results revealed that Met252, Ile253, Asn434, His435, and Tyr436 are the key residues of Fc and two binding modes based on tryptophan or tyrosine were constructed. Therefore, tryptophan was introduced in this work as a functional group of new ligand to enhance IgG-Fc affinity. 5-amino-benzimidazole (ABI) was selected as the charge-induced group to facilitate protein elution. Novel resin with tryptophan-ABI (W-ABI) ligand was then prepared. The static adsorption behaviors, dynamic binding capacity and elution properties were investigated. Moreover, MCIC with W-ABI resin was applied for IgG separation from albumin containing feedstock and mAb purification from cell culture supernatant.

2. Materials and methods 2.1. Materials Crossed-linked 4% agarose gel (Bestarose 4FF) was obtained from Bestchrom Bio-Technology Co., Ltd. (Shanghai, China). MEP HyperCel was purchased from Pall Life science (East Hills, NY, USA), and HiTrap rProtein A FF from GE Healthcare (Uppsala, Sweden). ABI was purchased from J&K Scientific Ltd. (Beijing, China), and tryptophan from Aladdin (Shanghai, China). Bovine serum ␥-globulin (bIgG, electrophoresis purity 99.0%) was purchased from Merck KGaA (Darmstadt, Germany). Bovine serum albumin (BSA) was obtained from Sigma (Milwaukee, WI, USA). Human immunoglobulin G for intravenous injection was purchased from Boya Biopharmaceutical Ltd. (Jiangxi, China). CHO cell culture supernatant (CCS) containing 9.6 mg/ml mAb-1 or 0.85 mg/ml mAb-2 was provided by a local biotechnology company. Other chemicals are of analytical grade.

MD simulations were performed using Amber 11 software with amber ff10 force field at pH 7. The geometric structure of W-ABI was optimized with Firefly v.8.0.1 [39,40]. The atomic charges of W-ABI were obtained with the RESP formalism using RED-vIII.5 tools [41]. The temperature was controlled at 300 K with a Langevin dynamics algorithm and a collision frequency of 2 ps−1 . The pressure was controlled by a weak coupling Berendsen scheme. The SHAKE algorithm was used for all covalent bonds involving hydrogen using a 2 fs time step. The non-bonded cutoff was set as 12 Å and long-range electrostatic interactions were evaluated using the Particle Mesh Ewald (PME). Detailed MD procedures could refer to our previous work [37]. 2.3. Preparation of W-ABI resin 10 g Bestarose 4FF gel beads were mixed with 5 ml allyl bromide and 2.5 g sodium hydroxide in 10 ml 20% (v/v) dimethyl sulfoxide solution, and the mixture was agitated at 150 rpm and 25 ◦ C for 24 h. The allyl-activated gels were brominated with 1.2 molar excess of N-bromosuccinimide over the allyl groups in 50% acetone at 150 rpm and 25 ◦ C for 1 h. After reaction, the gels were washed with deionized water. Three molar excess of tryptophan molecules over the allyl groups were added into the brominated gels in 1 M carbonate buffer (pH 10) at 150 rpm and 25 ◦ C for 8 h. The gels were successively washed with 30% (v/v), 70% (v/v), 100% (v/v) acetone and dioxane. 10 ml dioxane, 1.2 g N-hydroxysuccinimide (NHS) and 1.2 g N, N-Dicyclohexylcarbodiimide (DCC) were added to the gels, which was followed by shaking at 150 rpm and 25 ◦ C for 8 h to activate the carboxyl group of tryptophan. The NHS-activated gels were washed sequentially using 70% (v/v), 30% (v/v) acetone and deionized water. 10 ml deionized water and 200 mg ABI was added into the NHS-activated gels, and the reaction was carried out at 150 rpm and 25 ◦ C for 8 h. The remaining epoxy groups were blocked by 10 ml 50% (v/v) ethanolamine at 25 ◦ C for 2 h. Finally, the gels were washed extensively with deionized water and stored in 20% (v/v) ethanol at 4 ◦ C. 2.4. Adsorption equilibrium experiments The adsorption isotherms of human IgG on W-ABI resin were measured using the procedure described in our previous work [42]. The effects of pH at the range of 4.0–8.9 were focused. 20 mM acetate buffer (pH 4.0 and pH 5.0), 20 mM sodium phosphate buffer (pH 6.0, pH 7.0 and pH 8.0) and Tris–HCl buffer (pH 8.5 and pH 8.9) were used as the liquid phases. In addition, for MEP HyperCel the adsorption isotherms at pH 7–8.5 was also studied, and rProtein A FF was tested at pH 7–8 as recommended in the product instruction. The Langmuir equation was used to fit the experimental data with two parameters, the saturated adsorption capacity Qm and the apparent dissociation constant Kd .

2.2. Molecular simulation

2.5. Dynamic binding capacity

The crystal structure of binding complex between Fc portion of IgG and B domain of Protein A was obtained from Protein Data Bank (PDB ID: 1FC2, http://www.rcsb.org/pdb/) and used as the model for studying the interactions between W-ABI ligand and Fc. Molecular docking with Autodock vina 1.1.2 (http://vina.scripps. edu/) was employed for preliminary search of binding pose [38]. Polar hydrogen and Kollman United Atomic Charges were added to the protein structure, and Gasteiger charges were added to W-ABI ligand. According to the docking affinity energy, the best docked complex was chosen for further molecular dynamics (MD) simulation.

The IgG dynamic binding capacity of W-ABI at various linear flowrates was determined with Tricon 5/100 column (GE Healthcare, Uppsala, Sweden) through frontal breakthrough experiments. 3.5 mg/ml human IgG for intravenous injection (pH 7) was used as the feedstock. The protein concentration was monitored on-line at 280 nm with an UV detector (WellChrom fast scanning spectrophotometer K-2600, KNAUER, Berlin, Germany). The dynamic binding capacity for MEP HyperCel and rProtein A FF at different pH and flowrate was also studied. The dynamic adsorption capacity at 10% breakthrough (Q10% ) was calculated following the method published previously [43].

Please cite this article in press as: H.-F. Tong, et al., Multimodal charge-induction chromatography for antibody purification, J. Chromatogr. A (2015), http://dx.doi.org/10.1016/j.chroma.2015.12.047

G Model CHROMA-357149; No. of Pages 7

ARTICLE IN PRESS H.-F. Tong et al. / J. Chromatogr. A xxx (2015) xxx–xxx

2.6. Protein elution behavior in the column

HPLC. The purity and recovery of mAb-2 were obtained for each cycle.

The retention factor and recovery were determined by isocratic elution with ÄKTA explorer 100 (GE Healthcare, Uppsala, Sweden). 4 ml resin was packed in C10/10 column and 0.5 ml protein samples (5 mg/ml) were loaded. The proteins were eluted with different elution buffers followed by clean-in-place (CIP) with 0.5 M NaOH. The liquid flowrate was 1 ml/min, and the protein concentration was monitored at 280 nm. The un-retained volume (V0 ) of protein was determined with 0.1 M NaOH. The retention factor (KV ) and elution recovery (R) can be calculated as: KV = R=

VR − V0 V0

AR × 100% AR + ACIP

3

(1) (2)

where VR is the elution volume, AR is the peak area of elution, and ACIP is the peak area of CIP. 2.7. Purification of IgG from protein mixture A protein mixture of bovine IgG (1 mg/ml) and BSA (5 mg/ml) was used as the mimetic serum feedstock. 4 ml resin was packed into C10/10 column and the column was pre-equilibrated with 4 column volumes (CVs) of 20 mM sodium phosphate buffer (pH 7) or sodium phosphate buffer containing 0.1 M NaCl (pH 7). 10 ml protein mixture was loaded and washed with 5 CVs of equilibration buffer to remove unbound proteins. Those bound proteins were eluted with 20 mM acetate buffer (pH 4.0) and the column was regenerated with 4 CVs of 0.1 M NaOH. The flowrate was 1 ml/min, and the chromatographic run was monitored on-line at 280 nm. The collected fractions were analyzed with SDS-PAGE and SEC-HPLC. 2.8. Purification of mAb from cell culture supernatant The chromatographic separation of mAb-1 from CHO cell culture supernatant (containing 9.6 mg/ml mAb-1) was performed on ÄKTA explorer 100 system using Tricorn 5/5 column with 1 ml W-ABI resin. The flow rate was 0.5 ml/min. The column was preequilibrated with 4 CVs of 20 mM sodium phosphate buffer (pH 7), and 10–40 ml supernatant was loaded without any pre-treatment. The column was then washed with the equilibration buffer, and eluted with 20 mM acetate buffer (pH 3.6–5.5). The fractions collected were analyzed with SDS-PAGE, SEC-HPLC and ELISA. 1 ml MEP HyperCel in a Tricorn 5/5 column was tested to purify mAb-1 from 10 ml same feedstock for comparison under the optimized conditions. 20 mM sodium phosphate buffer (pH 7) and 20 mM acetate buffer (pH 5.0) were used as the equilibrium buffer and the elution buffer, respectively. HiTrap rProtein A FF (1 ml) was also tested with similar procedure, but 0.1 M citrate buffer (pH 3.0) was used for the elution and the eluate was collected into 2 ml-tube containing 100 ␮l Tris–HCl buffer (2.0 M, pH 8).

2.10. Assays Protein impurities were evaluated with SDS-PAGE using 8% nonreducing resolving gel or 10% reducing resolving gel. The purity and recovery of IgG were determined by SEC-HPLC with a G3000SWXL column (7.8 mm × 30 mm, TOSOH Bioscience) on a HPLC system (Beijing Chuangxintongheng Science & Technology Co., Ltd, Beijing, China). CHO host cell proteins (HCPs) was determined using CHO HCP ELISA Kit 3 G F015 from Cygnus Technologies (Southport, NC, USA). 3. Results and discussion 3.1. Rational design of new ligand Molecular interactions between the consensus binding site (CBS) on the Fc domain of IgG and seven natural Fc-specific ligands were evaluated in our previous work [37]. It was found that tryptophan and tyrosine on those ligands have significant contribution on the ligand-Fc binding, which was defined as TRP mode and TYR mode. The average energy contribution of tryptophan was higher than tyrosine, thus the TRP mode was preferentially considered for the design of new ligands. For TRP mode, there are two key residues that participate in the Fc recognition. Tryptophan as the main residue could provide pi–pi interaction with His435, and hydrophobic force with Met252 and Ile253. Except tryptophan, there is a second important residue acted as the auxiliary group to improve the specificity of ligand to IgG, which could be hydrophobic, aromatic or weak acid residues. Therefore, new ligands would need at least two functional parts, i.e. tryptophan and an additional group. In order to introduce the charge-induction effects, ABI was selected from a serial of heterocyclic compounds by molecular docking and experimental evaluation. The benzimidazole group of ABI is similar to the indole ring of tryptophan, which could act as the auxiliary group for TRP mode to enhance the recognition of Fc. Moreover, the pKa of ABI is 6.5, which means ABI has non-charge when pH above 6.5, and will be positive-charged under acidic pH. The electrostatic repulsion would prompt the departing between ABI and IgG. A new ligand W-ABI was designed by combining tryptophan and ABI (shown in Fig. 1). The results of molecular simulation with docking and MD indicated that W-ABI can bind to the CBS of Fc domain with high binding energies. W-ABI fit into the cave among the residues of Met252, Ile253, Asn434, His435 and Tyr436 on CBS. The benzimidazole group of ABI closely contacted with Asn434, His435 and Ile253, and the indole group of tryptophan interacted with Tyr436 and Met252. In addition, hydrogen bonds were formed between benzimidazole and His435 and pi–pi interactions existed between indole group and Tyr436. Therefore, W-ABI could provide the coordinated multimodal molecular interactions to IgG, which can mimic Fc-specific ligands with favorable recognition on CBS.

2.9. Reusing ability test 3.2. IgG adsorption onto the W-ABI resin The reuse test of W-ABI resin was performance with series of chromatographic separation of mAb-2 from CHO cell culture supernatant (containing 0.85 mg/ml mAb-2). ÄKTA explorer 100 system and Tricorn 5/5 column with 1 ml W-ABI resin was used. The flow rate was 0.5 ml/min. The column was pre-equilibrated with 20 mM sodium phosphate buffer (pH 7), and 10 ml supernatant was loaded. The column was then washed with the equilibration buffer, and eluted with 20 mM acetate buffer. Finally the column was regenerated with 0.1 M NaOH for next separation cycle. Totally 100 cycles were tested. All fractions were collected and analyzed with SEC-

The adsorption capacity of IgG onto the W-ABI resin was evaluated by batch experiments around neutral pH. The adsorption isotherms and the correlated Qm and Kd are shown in Fig. 2. The results show that the maximum capacity was at pH 7. At pH 6, Qm was 33.6 mg/ml, which raised to 70.4 mg/ml at pH 7 and decreased slightly to 66.8 mg/ml at pH 8. This can be explained by the protonation status of the W-ABI ligand and IgG. The pKa of benzimidazole is 6.5 and the isoelectric point (pI) of human IgG is about 7–9.5 [44]. At pH 7–8, both IgG and ligand had few net

Please cite this article in press as: H.-F. Tong, et al., Multimodal charge-induction chromatography for antibody purification, J. Chromatogr. A (2015), http://dx.doi.org/10.1016/j.chroma.2015.12.047

G Model CHROMA-357149; No. of Pages 7

ARTICLE IN PRESS H.-F. Tong et al. / J. Chromatogr. A xxx (2015) xxx–xxx

4

Fig. 1. Molecular structure of W-ABI ligand (a) and the interaction mode of W-ABI and IgG on the CBS (b). Ligands shown as ball and sticks, and hot spots of CBS shown as lines.

Fig. 2. Adsorption isotherms (a), Qm and Kd (b) of human IgG on W-ABI resin at different pHs. Symbols represent experimental data, and solid lines represent the fitting results with Langmuir equation. () pH 6.0; () pH 7.0; (䊏) pH 8.0.

charges that can favor the binding of IgG to W-ABI resin through multimodal interactions, including hydrophobic and pi–pi interactions. At pH 6, both the W-ABI ligand and IgG have positive charges, so the charge-induced electrostatic repulsion could significantly reduce the adsorption capacity of IgG on the W-ABI resin. The values of Kd were in the range of 0.19–1.82 mg/ml (corresponding 1.27–12.1 × 10−6 M), which were slightly lower than that of synthetic ligands (HWRGWV, HYFKFD, and HFRRHL) (10−5 –10−6 M) [24] but higher than that of Protein A ligand (10−7 –10−8 M) [6]. In addition, the effects of pH on the adsorption of IgG on typical HCIC resin MEP HyperCel and Protein A-based resin rProtein A FF were also tested (data not shown). It was found that the adsorption capacity of IgG on MEP HyperCel was high around neutral pH and had no significant difference at the range of pH 7.0–8.5, and rProtein A FF also showed the highest adsorption capacity at pH 7.0–7.5. Therefore, for three resins tested, neutral pH around 7.0 would be the suitable condition for IgG adsorption. 3.3. Dynamic binding capacity The dynamic binding capacities of three resins (W-ABI, MEP HyperCel and rProtein A FF) were measured at pH 7.0. The breakthrough curves of IgG on W-ABI column at different linear flowrates are shown in Fig. 3. The DBC at 10% breakthrough (Q10% ) was 29.6–23.1 mg/ml at the linear flowrate of 76.5–306 cm/h, which corresponds to the residence time of 7.8–1.95 min with the bed height of 9.8 cm. The comparison with typical HCIC resin MEP HyperCel and Protein A-based resin rProtein A FF are listed in Table 1. It was found that DBC of W-ABI was higher than that of rProtein A FF, and also better than that of MEP HyperCel at high flowrate.

Fig. 3. Breakthrough curves of hIgG on W-ABI resin at different linear flowrates.

Table 1 Dynamic binding capacity of IgG on different resins. Resin

Capacity

W-ABI

Q10% Q10% /Q∗ Q10% Q10% /Q∗ Q10% Q10% /Q∗

MEP HyperCel rProA FF

U (cm/h) 76.5

153

306

29.6 0.493 46.4 0.542 15.6 0.390

27.5 0.458 27.7 0.324 12.1 0.303

23.1 0.384 17.3 0.202 9.9 0.248

Note: W-ABI, Q* = 60.1 mg/ml; MEP, Q* = 85.5 mg/ml; rProA, Q* = 40 mg/ml.

Please cite this article in press as: H.-F. Tong, et al., Multimodal charge-induction chromatography for antibody purification, J. Chromatogr. A (2015), http://dx.doi.org/10.1016/j.chroma.2015.12.047

G Model

ARTICLE IN PRESS

CHROMA-357149; No. of Pages 7

H.-F. Tong et al. / J. Chromatogr. A xxx (2015) xxx–xxx

5

Fig. 4. Retention factor (a) and recovery (b) of proteins on W-ABI column at different pHs.

In order to better evaluate the dynamic adsorption performance, Q10% /Q∗ was compared. For W-ABI, Q10% /Q∗ decreased only 22.1% as the flowrate increased from 76.5 to 306 cm/h. In contrast, rProtein A FF and MEP HyperCel showed decrease of 36.4% and 62.7%, respectively. The results indicate that W-ABI would be more suitable for high velocity applications, which certainly improves the process efficiency of chromatographic separation for large-scale production. 3.4. Protein elution behavior in the column Electrostatic repulsion between ligands and proteins is the main driving force for the protein elution during HCIC, which is one of the most important characteristics of charge-induction chromatography. With hIgG and BSA as the model proteins, the retention factor KV and the recovery R of protein on W-ABI were evaluated at pH 3.6–5.5 and the results are shown in Fig. 4. BSA could be eluted when the pH was lower than 4.0. For hIgG, the elution pH was relative higher due to the enhanced repulsion between ligand and IgG. The effective elution of IgG from W-ABI resin could be achieved as pH below 5.5. Compared to other HCIC ligands for IgG separation, this elution pH condition was relatively mild [18,22,26,28]. 3.5. Purification of IgG from protein mixture The W-ABI resin was used to separate IgG from a serummimicked protein mixture (1 mg/ml bovine IgG and 5 mg/ml BSA). Sometimes the albumin-containing media is applied for cell culture of mAb production. IgG can also be separated from natural resources, such as serum, whey and ascites fluid. For that, serum albumin would be main impurities during the separation of IgG. After preliminary optimization, the loading pH was chosen as pH 7.0 and the elution pH was pH 4.0. As shown in Fig. 1S (Supporting information), almost all albumins flowed through the column, and no IgG was detected in the breakthrough fractions. There was only a small amount of BSA found in the elution fraction. The results of SEC-HPLC analysis indicated the IgG purity was 95.2%. Moreover, 0.1 M NaCl addition in the loading buffer could further enhance the binding selectivity of IgG, and the purity was improved to 98.4%. IgG recovery was at the range of 88.1–89.9%, which was higher than that of MEP HyperCel with similar feedstock [42,45]. 3.6. Purification of mAb from cell culture supernatant The W-ABI resin was further challenged to purify mAb-1 from CHO cell culture supernatant. The influence of elution pH on the

Fig. 5. Chromatographic separation of mAb-1 from cell culture supernatant with WABI resin under different elution pHs. (a) Chromatogram: loading at pH 7; elution at pH 3.6–5.5; breakthrough (BT). (b) non-reduced SDS-PAGE. Protein maker (M), feedstock (F), breakthrough fraction (B), elution factions at pH 3.6 (E1), pH 4.0 (E2), pH 4.5 (E3), pH 5.0 (E4) and pH 5.5 (E5), mAb standard (Std).

purity and recovery of mAb were studied. Fig. 5 shows that almost all impurities flow through the column and there was no mAb detected in breakthrough fraction, as also shown in Fig. 2S (Supporting information). Elution at pH 3.6 or 4.0 had high recovery with the purity of about 93%, but mAb was partly degraded under serious acidic condition as shown in SDS-PAGE (Fig. 5). The recoveries of mAb under elution at pH 4.5–5.0 were about 88% with high purity over 97%. For the elution at pH 5.5, the purity could be improved, but the recovery was only 37.5%. Therefore, elution at pH 4.5–5.0 was suitable, and pH 5.0 was the optimal condition for high purity.

Please cite this article in press as: H.-F. Tong, et al., Multimodal charge-induction chromatography for antibody purification, J. Chromatogr. A (2015), http://dx.doi.org/10.1016/j.chroma.2015.12.047

G Model

ARTICLE IN PRESS

CHROMA-357149; No. of Pages 7

H.-F. Tong et al. / J. Chromatogr. A xxx (2015) xxx–xxx

6

Table 2 Operation conditions and chromatographic separation results of mAb from cell culture supernatant with different resins. Group

Resin

Loading (ml)

Elutiona

Breakthrough IgG (%)

Purity (%)

Yield (%)

1 2 3 4 5 6 7 8

W-ABI W-ABI W-ABI W-ABI W-ABI W-ABI MEP HyperCel rProA FF

10 10 10 10 10 20 10 10

pH 3.6 pH 4.0 pH 4.5 pH 5.0 pH 5.5 pH 5.0 pH 5.0 pH 3.0

NDb ND ND ND ND ND ND 0.30%

93.7 93.8 97.6 98.9 98.5 99.2 96.7 99.9

89.6 94.8 92.8 88.6 37.5 87.6 86.1 89.3

a b

Citrate buffer, 100 mM. No detection.

Fig. 7. Recovery (white) and purity (grey) of mAb-2 with W-ABI separation for each 20 cycles.

for W-ABI resin was similar to MAbsorbent A2 P with biomimetic ligand [46]. 3.7. Reuse test

Fig. 6. Purification of mAb-1 from cell culture supernatant with different resins. Chromatogram: breakthrough (BT), (b) Non-reduced SDS-PAGE: Protein maker (M), feedstock (F), W-ABI BT (B1), W-ABI elution (E1), W-ABI clean-in-place (CIP), mAb standard (Std), MEP HyperCel BT (B2), MEP HyperCel elution (E2), rProtein A BT (B3), rProtein A elution (E3).

The reuse ability of W-ABI resin was test with the separation of mAb-2 from CHO cell culture supernatant. The purity of recovery of mAb-2 for each 20 cycles are shown in Fig. 7. During total 100 cycles, the purities were quite stable and kept at the range of 96–98%. The recoveries were also as high as 98% for the first 60 cycles, and then reduced slightly to 91% during 80–100 cycles. The results demonstrated that new resin can be reused more than 100 cycles without the performance degradation on the separation purity and recovery of antibody, which would be suitable for the industrial applications. 4. Conclusions

In addition, when the sample loading volume increased from 10 to 20 ml, no mAb was found in the breakthrough fractions. The mAb purity was as high as 99.2% with a recovery of 87.6%. In order to better evaluate the performance of W-ABI resin, two typical resins, MEP HyperCel and rProtein A FF, were tested with same feedstock under the optimized separation conditions, and the results are compared in Fig. 6, Fig. 3S (Supporting information) and Table 2. It was found that all three resins could purify mAb with high efficiency. The performance of W-ABI was better than that of MEP HyperCel, and comparable with that of rProtein A FF. It should be noted that the elution pH for rProtein A FF was as low as 3, which may lead to unfavorable aggregation or degradation of mAb. In contrast, the elution for W-ABI could be achieved at pH 5, which was better for mAb purification. The removal of HCP was also evaluated and compared. The content of HCP with rProtein A FF was the lowest (1055 ppm), followed by W-ABI (4909 ppm) and MEP HyperCel (7057 ppm). The log reduction value (LRV) for the three resins were 1.43, 0.76 and 0.60, respectively. The performance of HCP removal

MCIC was proposed as the extending of HCIC to improve the binding specificity with multimodal molecular interactions. A novel multimodal charge-induction ligand W-ABI was designed based on molecular simulation of the binding mode between ligand and Fc fragment of IgG. W-ABI resin was prepared by coupling tryptophan and 5-amino-benzimidazole onto agarose gels. The adsorption experiments showed that the W-ABI resin has high affinity to IgG with the saturated adsorption capacity of 70.4 mg/ml at pH 7.0. The influence of flow rate on DBC was less than that of other resins compared, which indicated that W-ABI could be used for high velocity processes. The elution studies demonstrated that IgG could be eluted effectively at mild acidic condition (pH 4.5–5.0) with the charge-induction effect. In addition, W-ABI was used successfully to purify mAb from cell culture supernatant. The purity and recovery of mAb were 98.9 and 88.6%, respectively. The HCP removal performance was satisfactory, which was more efficient than MEP HyperCel. The elution pH could be set at pH 5, which was higher than that of rProtein A FF and more suitable for mAb purification.

Please cite this article in press as: H.-F. Tong, et al., Multimodal charge-induction chromatography for antibody purification, J. Chromatogr. A (2015), http://dx.doi.org/10.1016/j.chroma.2015.12.047

G Model CHROMA-357149; No. of Pages 7

ARTICLE IN PRESS H.-F. Tong et al. / J. Chromatogr. A xxx (2015) xxx–xxx

In summary, new resin for MCIC combined multimodal molecular interactions to improve binding selectivity and charge-induction effects for convenient elution, which is promising for cost-effective and large-scale purification of antibodies. Acknowledgments This work was supported by the National Natural Science Foundation of China and the Zhejiang Provincial Natural Science Foundation of China. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chroma.2015.12. 047. References [1] J.M. Reichert, Marketed therapeutic antibodies compendium, Mabs 4 (2012) 413–415. [2] J.G. Elvin, R.G. Couston, C.F. van der Walle, Therapeutic antibodies: market considerations, disease targets and bioprocessing, Int. J. Pharm. 440 (2013) 83–98. [3] A.A. Shukla, J. Thommes, Recent advances in large-scale production of monoclonal antibodies and related proteins, Trends Biotechnol. 28 (2010) 253–261. [4] F. Li, N. Vijayasankaran, A.Y. Shen, R. Kiss, A. Amanullah, Cell culture processes for monoclonal antibody production, Mabs 2 (2010) 466–479. [5] P. Gagnon, Technology trends in antibody purification, J. Chromatogr. A 1221 (2012) 57–70. [6] S. Hober, K. Nord, M. Linhult, Protein A chromatography for antibody purification, J. Chromatogr. B 848 (2007) 40–47. [7] D. Low, R. O’Leary, N.S. Pujar, Future of antibody purification, J Chromatogr. B 848 (2007) 48–63. [8] A.D. Naik, M. Raina, A.M. Lali, AbSep-an amino acid based pseudobioaffinity adsorbent for the purification of immunoglobulin G, J. Chromatogr. A 1218 (2011) 1756–1766. [9] A. el-Kak, M.A. Vijayalakshmi, Study of the separation of mouse monoclonal antibodies by pseudobioaffinity chromatography using matrix-linked histidine and histamine, J. Chromatogr. 570 (1991) 29–41. [10] E. Nedonchelle, O. Pitiot, M.A. Vijayalakshmi, A preliminary study for isolation of catalytic antibodies by histidine ligand affinity chromatography as an alternative to conventional protein A/G methods, Appl. Biochem. Biotechnol. 83 (2000) 287–294, discussion 294–285, 297–313. [11] Z. Du, S. Zhang, C. Zhou, M. Liu, G. Li, l-histidine functionalized multi-walled carbon nanotubes for on-line affinity separation and purification of immunoglobulin G in serum, Talanta 99 (2012) 40–49. [12] R. Li, V. Dowd, D.J. Stewart, S.J. Burton, C.R. Lowe, Design, synthesis, and application of a protein A mimetic, Nat. Biotechnol. 16 (1998) 190–195. [13] S.F. Teng, K. Sproule, A. Husain, C.R. Lowe, Affinity chromatography on immobilized biomimetic ligands. Synthesis, immobilization and chromatographic assessment of an immunoglobulin G-binding ligand, J. Chromatogr. B Biomed. Sci. Appl. 740 (2000) 1–15. [14] A.C. Roque, M.A. Taipa, C.R. Lowe, An artificial protein L for the purification of immunoglobulins and fab fragments by affinity chromatography, J. Chromatogr. A 1064 (2005) 157–167. [15] A.C. Roque, M.A. Taipa, C.R. Lowe, Synthesis and screening of a rationally designed combinatorial library of affinity ligands mimicking protein L from Peptostreptococcus magnus, J. Mol. Recognit. 18 (2005) 213–224. [16] G. Fassina, A. Verdoliva, G. Palombo, M. Ruvo, G. Cassani, Immunoglobulin specificity of TG19318: a novel synthetic ligand for antibody affinity purification, J. Mol. Recognit. 11 (1998) 128–133. [17] G. Fassina, A. Verdoliva, M.R. Odierna, M. Ruvo, G. Cassini, Protein A mimetic peptide ligand for affinity purification of antibodies, J. Mol. Recognit. 9 (1996) 564–569. [18] A. Verdoliva, F. Pannone, M. Rossi, S. Catello, V. Manfredi, Affinity purification of polyclonal antibodies using a new all-D synthetic peptide ligand: comparison with protein A and protein G, J. Immunol. Methods 271 (2002) 77–88. [19] D. Moiani, M. Salvalaglio, C. Cavallotti, A. Bujacz, I. Redzynia, G. Bujacz, F. Dinon, P. Pengo, G. Fassina, Structural characterization of a protein A mimetic peptide dendrimer bound to human IgG, J. Phys. Chem. B 113 (2009) 16268–16275. [20] J.M. Haigh, A. Hussain, M.L. Mimmack, C.R. Lowe, Affinity ligands for immunoglobulins based on the multicomponent Ugi reaction, J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 877 (2009) 1440–1452.

7

[21] G. El Khoury, C.R. Lowe, A biomimetic protein G affinity adsorbent: an Ugi ligand for immunoglobulins and Fab fragments based on the third IgG-binding domain of protein G, J. Mol. Recognit. 26 (2013) 190–200. [22] J. Qian, G. El Khoury, H. Issa, K. Al-Qaoud, P. Shihab, C.R. Lowe, A synthetic protein G adsorbent based on the multi-component Ugi reaction for the purification of mammalian immunoglobulins, J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 898 (2012) 15–23. [23] H. Yang, P.V. Gurgel, R.G. Carbonell, Hexamer peptide affinity resins that bind the Fc region of human immunoglobulin G, J. Pept. Res. 6 (2005) 120–137. [24] H. Yang, P.V. Gurgel, R.G. Carbonell, Purification of human immunoglobulin G via Fc-specific small peptide ligand affinity chromatography, J. Chromatogr. A 1216 (2009) 910–918. [25] H. Yang, P.V. Gurgel, D.K. Williams Jr, B.G. Bobay, J. Cavanagh, D.C. Muddiman, R.G. Carbonell, Binding site on human immunoglobulin G for the affinity ligand HWRGWV, J. Mol. Recognit. 23 (2010) 271–282. [26] A.D. Naik, S. Menegatti, P.V. Gurgel, R.G. Carbonell, Performance of hexamer peptide ligands for affinity purification of immunoglobulin G from commercial cell culture media, J. Chromatogr. A 1218 (2011) 1691–1700. [27] S. Menegatti, M. Hussain, A.D. Naik, R.G. Carbonell, B.M. Rao, mRNA display selection and solid-phase synthesis of Fc-binding cyclic peptide affinity ligands, Biotechnol. Bioeng. 110 (2013) 857–870. [28] L.N. Lund, P.E. Gustavsson, R. Michael, J. Lindgren, L. Norskov-Lauritsen, M. Lund, G. Houen, A. Staby, P.M. St Hilaire, Novel peptide ligand with high binding capacity for antibody purification, J. Chromatogr. A 1225 (2012) 158–167. [29] S.C. Burton, D.R. Harding, Hydrophobic charge induction chromatography: salt independent protein adsorption and facile elution with aqueous buffers, J. Chromatogr. A 814 (1998) 71–81. [30] E. Boschetti, Antibody separation by hydrophobic charge induction chromatography, Trends Biotechnol. 20 (2002) 333–337. [31] L. Guerrier, P. Girot, W. Schwartz, E. Boschetti, New method for the selective capture of antibodies under physiolgical conditions, Bioseparation 9 (2000) 211–221. [32] W. Schwart, D. Judd, M. Wysocki, L. Guerrier, E. Birck-Wilson, E. Boschetti, Comparison of hydrophobic charge induction chromatography with affinity chromatography on protein A for harvest and purification of antibodies, J. Chromatogr. A 908 (2001) 251–263. [33] H.F. Xia, D.Q. Lin, L.P. Wang, Z.J. Chen, S.J. Yao, Preparation and evaluation of cellulose adsorbents for hydrophobic charge induction chromatography, Ind. Eng. Chem. Res. 47 (2008) 9566–9572. [34] G. Zhao, G. Peng, F. Li, Q. Shi, Y. Sun, 5-Aminoindole, a new ligand for hydrophobic charge induction chromatography, J. Chromatogr. A 1211 (2008) 90–98. [35] Q.H. Shi, F.F. Shen, S. Sun, Studies of lysozyme binding to histamine as a ligand for hydrophobic charge induction chromatography, Biotechnol. Prog. 26 (2010) 134–141. [36] W. Phottraithip, D.Q. Lin, F. Shi, S.J. Yao, New hydrophobic charge-induction resin with 2-mercaptoimidazole as the ligand and its separation characteristics for porcine IgG, Biotechnol. Bioprocess Eng. 18 (2013) 1169–1175. [37] H.F. Tong, D.Q. Lin, Q.L. Zhang, R.Z. Wang, S.J. Yao, Molecular recognition of Fc-specific ligands binding onto the consensus binding site of IgG: insights from molecular simulation, J. Mol Recognit. 27 (2014) 501–509. [38] O. Trott, A.J. Olson, AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading, J. Comput. Chem. 31 (2010) 455–461. [39] A. Alex Granovsky Firefly version 8.0, www . [40] M.W. Schmidt, K.K. Baldridge, J.A. Boatz, S.T. Elbert, M.S. Gordon, J.H. Jensen, S. Koseki, N. Matsunaga, K.A. Nguyen, S. Su, T.L. Windus, M. Dupuis, J.A. Montgomery, General atomic and molecular electronic structure system, J. Comput. Chem. 14 (1993) 1347–1363. [41] P. Cieplak, W.D. Cornell, C. Bayly, P.A. Kollman, Application of the multimolecule and multiconformational RESP methodology to biopolymers: charge derivation for DNA, RNA, and proteins, J. Comput. Chem. 16 (1995) 1357–1377. [42] H.F. Tong, D.Q. Lin, X.M. Yuan, S.J. Yao, Caprylate as the albumin-selective modifier to improve IgG purification with hydrophobic charge-induction chromatography, J. Chromatogr. A 1285 (2013) 88–96. [43] H.F. Xia, D.Q. Lin, S.J. Yao, Chromatographic performance of macroporous cellulose-tungsten carbide composite beads as anion-exchanger for expanded bed adsorption at high fluid velocity, J. Chromatogr. A 1195 (2008) 60–66. [44] C. Prin, M.C. Bene, B. Gobert, P. Montagne, G.C. Faure, Isoelectric restriction of human immunoglobulin isotypes, Biochim. Biophys. Acta 1243 (1995) 287–289. [45] H.F. Tong, D.Q. Lin, X.M. Yuan, S.J. Yao, Enhancing IgG purification from serum albumin containing feedstock with hydrophobic charge-induction chromatography, J. Chromatogr. A 1244 (2012) 116–122. [46] S. Ghose, B. Hubbard, S.M. Cramer, Evaluation and comparison of alternatives to protein A chromatography mimetic and hydrophobic charge induction chromatographic stationary phases, J. Chromatogr. A 1122 (2006) 144–152.

Please cite this article in press as: H.-F. Tong, et al., Multimodal charge-induction chromatography for antibody purification, J. Chromatogr. A (2015), http://dx.doi.org/10.1016/j.chroma.2015.12.047