Evaluation of adsorption selectivity of immunoglobulins M, A and G and purification of immunoglobulin M with mixed-mode resins

Accepted Manuscript Title: Evaluation of adsorption selectivity of Immunoglobulins M, A and G and purification of immunoglobulin M with mixed-mode resins Authors: Ying-Di Luo, Qi-Lei Zhang, Shan-Jing Yao, Dong-Qiang Lin PII: DOI: Reference:

S0021-9673(17)31778-8 https://doi.org/10.1016/j.chroma.2017.12.018 CHROMA 359074

To appear in:

Journal of Chromatography A

Received date: Revised date: Accepted date:

9-11-2017 7-12-2017 8-12-2017

Please cite this article as: Luo Y-D, Zhang Q-L, Yao S-J, Lin D-Q, Evaluation of adsorption selectivity of Immunoglobulins M, A and G and purification of immunoglobulin M with mixed-mode resins, Journal of Chromatography A (2010), https://doi.org/10.1016/j.chroma.2017.12.018 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Evaluation of adsorption selectivity of Immunoglobulins M, A and G and purification of immunoglobulin M with

SC R

Ying-Di Luo, Qi-Lei Zhang, Shan-Jing Yao, Dong-Qiang Lin*

IP T

mixed-mode resins

Key Laboratory of Biomass Chemical Engineering of Ministry of Education, College

U

of Chemical and Biological Engineering, Zhejiang University, Hangzhou 310027,

A

N

China

ED

Prof. Dong-Qiang Lin

M

*Corresponding author:

College of Chemical and Biological Engineering

PT

Zhejiang University

CC E

Hangzhou 310027, China Fax: +86-571-87951982

A

E-mail: D.-Q. Lin ([email protected])

Highlights

► Adsorption selectivity of IgM and IgA with mixed-mode resins was studied. ► Four mixed-mode resins with different ligands were evaluated 1

► Adsorption selectivity could be improved by the control of pH. ► Monoclonal IgM was separated from cell broth with mixed-mode resin. ►IgM was purified directly from human serum with two mixed-mode resins

IP T

combined together.

ABSTRACT

IgA and IgG on four

SC R

This study investigated adsorption selectivity of IgM,

mixed-mode resins with the functional ligands of 4-mercatoethyl-pyridine (MEP),

U

2-mercapto-1-methylimidazole (MMI), 5-aminobenzimidazole (ABI) and

N

tryptophan-5-aminobenzimidazole (W-ABI), respectively. IgM purification processes

A

with mixed-mode resins were also proposed. All resins showed typical pH-dependent

M

adsorption, and high adsorption capacity was found at pH 5.0~8.0 with low

ED

adsorption capacity under acidic conditions. Meanwhile, high selectivity of IgM/IgA and IgM/IgG was obtained with ABI-4FF and MMI-4FF resins at pH 4.0~5.0, which

PT

was used to develop a method for IgM, IgA and IgG separation by controlling loading

CC E

and elution pH. Capture of monoclonal IgM from cell culture supernatant with ABI-4FF resins was studied and high purity (~99%) and good recovery (80.8%) were obtained. Moreover, IgM direct separation from human serum with combined

A

two-step chromatography (ABI-4FF and MMI-4FF) was investigated, and IgM purity of 65.2% and a purification factor of 28.3 were obtained after optimization. The antibody activity of IgM was maintained after purification. The results demonstrated that mixed-mode chromatography with specially-designed ligands is a promising way 2

to improve adsorption selectivity and process efficiency of IgM purification from complex feedstock.

Keywords: Mixed-mode chromatography, adsorption selectivity, human

IP T

immunoglobulin M, human serum, cell culture supernatant

SC R

1. Introduction

Immunoglobulins are glycoproteins with antibody activity that constitute 15%-20%

U

of total serum proteins [1]. Five immunoglobulin classes are distinguished and named

N

IgG, IgA, IgM, IgD and IgE with serum concentrations of 8~16 mg/mL, 1~4 mg/mL,

A

0.5~2 mg/mL, ~0.4 mg/mL and 10~400 ng/mL, respectively [2-4]. Molecular

M

structures of IgG, IgA and IgM are showed in Fig. S1 (supporting information). Each

ED

IgG, IgA and IgM molecule is made up of two heavy chains (CH, C C-chains , respectively) and two light chains. IgG molecule shows ‘Y’ shape with two disulfide

PT

bridges in hinge region and one glycosylation site between CH2 and CH3 [5]. IgA is

CC E

chiefly monomeric in human serum with ~90 % IgA1 which is ‘T-shaped’ due to the extended hinge region with three to six O-linked glycosylation sites [6]. IgM dominantly exists in human serum as a cyclic pentamer with molecular weight of

A

~950 kDa, which is a mushroom-shaped molecule with a central protruding region called J chain. Its heavy chains are covalently linked in CCCdomains by disulfide bridges upon assembly [7]. Most antibody therapeutics are IgG-based drugs due to their effectiveness in activating human immune defense and extended plasma 3

half-life [8]. Nevertheless, IgM has attracted great attention in biopharmaceutical industry and showed an important role in eradicating upcoming tumor cells. It can be used for tumor therapy [9,10], combating microbial infections [11,12] and vaccine adjuvant [13] or preventive vaccine [14].

IP T

IgM purification is much more difficult than IgG because of its size and complexity. The industrial processes of IgG purification from human serum, ascites or

SC R

cell culture supernatants are mainly focused on Protein A or Protein G affinity chromatography. Traditional methods for IgM purification are based on the

U

combination of precipitation [15-18] and chromatography which includes

N

size-exclusion chromatography [16,17,19], ion-exchange chromatography [20-23],

A

hydrophobic interaction chromatography [21,24], thiophilic chromatography [25] and

M

hydroxyapatite chromatography [26,27]. Affinity chromatography with special

ED

ligands like mannan binding protein [28], snowdrop bulb lectin [29], C1q [30] and Protein L [31] was also used to purify IgM. However, those ligands from natural

PT

source are complex, expensive and have poor stability. Currently, biominetic peptide

CC E

ligands like tetrameric ligand TG19318 [32] and hexameric ligand HWRGWV [33,34] are viable alternatives to natural binding partners for low cost and easy synthesis, but they have weak affinity to distinguish IgG, IgA, IgM, IgE or IgD [35-37]. A fourteen

A

peptide ligand which mimicked the human polymeric immunoglobulin receptor (hpIgR) exhibited good specificity to IgM with negligible affinity for IgG, IgE and IgA. However, the structure is too complex and could be easily hydrolyzed by NaOH

4

solution [38,39]. Great efforts are needed to develop new practicable ligands and improve binding selectivity for IgM. Mixed-mode chromatography (MMC) is new type of bioseparation method that utilizes more than one form of interactions between the stationary phase and solutes in

IP T

the feed [40]. Mixed-mode ligands usually combine different functional groups with multiple interactions to proteins, including hydrophobic interactions, electrostatic

SC R

interactions and hydrogen bonding [41]. Those specially-designed small-molecular

ligands of MMC have unique properties to facilitate separation process. The binding

U

selectivity of MMC can be improved and high adsorption capacity is often found with

N

high ligand density. In addition, salt-tolerant adsorption property of MMC can offer

A

flexibility in feedstock pretreatment without the need of dilution or salt addition [42].

M

MMC has already showed excellent performance on the purification of IgG antibodies

ED

[43-45]. For IgM purification, Chen et al. [46] compared several methods including MMC with ABx resin, anion exchange chromatography with Mono Q

and size

PT

exclusion chromatography with Superose-6 and. The ABx resin combines weak

CC E

anionic, cationic and hydrophobic interactions on a silica gel support and the results indicated that ABx had high selectivity to IgM and could separate IgM from ascites fluids with a IgM purity of 99%. However, the ABx resin also binds IgG and IgA,

A

which limits its application in separating IgM from complex feedstock containing IgG and IgA. In our previous work, a series of MMC resins with multimodal ligands including 2-mercapto-1-methyl-imidazole (MMI) [47], 5-aminobenzimidazole (ABI) [48] and 5

tryptophan-5-aminobenzimidazole (W-ABI) [49] were developed for antibody separation. The adsorption selectivity of hIgG and Fab/Fc fragments was evaluated and the W-ABI resin was used to separate hIgG and Fc fragment with high efficiency under optimized pH conditions [50]. In the present work, MMC was investigated for

IP T

IgM separation using MEP HyperCel (commercial resin) and three laboratory-developed resins (with MMI, ABI and W-ABI ligands, respectively). The

SC R

adsorption selectivity of human IgM, IgA and IgG of these four mixed-mode resins was evaluated and potential process conditions for IgM separation was analyzed.

N

U

Moreover, IgM purification from cell culture media and human serum was studied.

M

2.1. Materials

A

2. Materials and methods

ED

MEP HyperCel was purchased from Pall Corporation (East Hills, NY, USA). MMI and ABI were purchased from J&K Scientific Ltd (Beijing, China), and

PT

tryptophan was from Aladdin (Shanghai, China). Bestarose 4FF was obtained from

CC E

Bestchrom Bio-Technology Co., Ltd (Shanghai, China). Human IgM and IgA were purchased from Athens Research & Technology, Inc (Georgia, USA), which have molecular weights of ~ 950 and 160 kDa, respectively. Human IgG for intravenous

A

injection (hIgG, electrophoresis purity 95%, MW=150 kDa) was purchased from Green Cross (Chinese) Biological Products Co., Ltd (Huainan, China). Other chemicals were all analytical grade and used as received.

6

The monoclonal human IgM supernatant was expressed in hybridoma cells and was generously provided by Vazyme Biotech Co., Ltd (Nanjing, China), and the titer was about 0.5 mg/mL. The human serum was kindly provided by a local hospital

IP T

(Hangzhou, China) and IgM concentration in the serum was ~ 1.1 mg/mL.

2.2 Preparation of mixed-mode resins

SC R

The mixed-mode resins with ligands MMI, ABI or W-ABI were prepared

according to the methods reported by Lu et al. [51], Yan et al. [48] and Tong et al. [49].

U

Briefly, 10 g Bestarose 4FF gel beads were activated by allyl bromide and brominated

N

with N-bromosuccinimide, and then mixed with 3 molar excess (over the allyl group)

A

of MMI, ABI or tryptophan. The resins coupled with MMI or ABI were obtained after

M

washing and named MMI-4FF and ABI-4FF, respectively. The beads with

ED

tryphtophan ligand were further coupled with ABI to prepare W-ABI resin named W-ABI-4FF. The ligand densities (determined by pH titration method [17, 18, 31]) of

PT

MMI-4FF, ABI-4FF and W-ABI-4FF were 150, 124, 55 mol/g gel, respectively. All

CC E

resins were stored in 20% (v/v) ethanol at 4 °C.

A

2.3. Adsorption in column IgM and IgA adsorption in the column were investigated as our previous work

[51] with ÄKTA Prime Plus system (GE Healthcare, Uppsala, Sweden). A 5 mm I.D. column (Tricorn 5/100, GE Healthcare, Uppsala, Sweden) was packed with 1.1 mL Protein A, MEP HyperCel, MMI-4FF, ABI-4FF or W-ABI-4FF resins. Acetate buffer 7

(pH 3.0, 0.1 M and pH 4.0, 5.0, 20 mM), sodium phosphate buffer (pH 6.0, 7.0 and 8.0, 20 mM) and Tris–HCl buffer (pH 8.9, 20 mM) were used as the liquid phases. The column was equilibrated with 10 column volumes (CV) equilibration buffer under a flow rate of 0.5 mL/min. IgM or IgA solution (1 mg/mL) was loaded onto the

IP T

column at different pH values (pH 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 8.9). The loading volume was 100 L. Unbound proteins were then washed with equilibration buffer.

SC R

The absorbed protein was eluted with 10 CV 0.1 M acetic acid (pH 3.0) under a flow

rate of 0.5 mL/min and regenerated with 10 CV 0.1 M NaOH under a flow rate of 0.3

Breakthrough peak area ×100 Total peak area

(1)

A

N

PA (%) = 100 -

U

mL/min. The adsorption proportion (PA) of protein was calcated as follows,

The selectivity index (SI) of protein adsorption was defined as, (2)

10PA_IgM% = 10PA_IgG%

(3)

M

10PA_IgM% = 10PA_IgA%

ED

SI_IgM/IgA

SI_IgA/IgG

10PA_IgA % = 10PA_IgG %

(4)

CC E

PT

SI_IgM/IgG

2.4. Protein elution in column

A

The elution of IgM in the column were also investigated with the ÄKTA Prime

Plus system and Tricorn column (1.1 mL MMI-4FF or ABI-4FF resins). IgM solution (pH 4.5, 1.0 mg/mL) was first loaded onto the pre-equilibrated column. After washing to baseline, the adsorbed proteins were eluted by different eluents (pH 2.2 and 3.0

8

with different additives). The column was then regenerated with 0.1 M NaOH and re-equilibrated with the equilibrium buffer for next analysis. The elution proportion (PE) of protein was calculated as follows. Elution peak area ×100 Total peak area-Breakthrough peak area

SC R

2.5. Monoclonal hIgM purification from cell culture supernatant

(5)

IP T

PE (%) =

The monoclonal hIgM separation from hybridoma cell culture supernatant (pH 7.0~7.2) was studied using the same ÄKTA Prime Plus system and Tricon column as

U

described in Section 2.3. The column was packed with 1.1 mL ABI-4FF resin. The

A

N

cell culture supernatant (containing 0.5 mg/mL IgM) was filtered through a 0.22 m

M

microporous membrane, and then 0.5~1.0 mL supernatant was loaded onto the ABI-4FF column at 0.5 mL/min. The column was washed by the equilibrium buffer

ED

(pH 7.0, 20 mM), and IgM was then eluted with 10 CV glycine buffer (0.1 M

PT

Gly-HCl, pH 2.2) with 1.0 M arginine at 1 mL/min. The elution volume was about 2~3 mL. The flow-through and elution fractions were collected and analyzed by

CC E

SEC-HPLC. The recovery (R) of monoclonal IgM was calculated as following, Elution amount (IgM peak × volume) ×100% Feed amount (IgM peak × volume)

(7)

A

R (%) =

2.6. IgM purification from human serum The human serum was adjusted to pH 4.5~5.0 at room temperature with 1.0 M acetic acid. After centrifugation at 12000 g under 4 oC, the precipitate was removed and the suspension was collected. The suspension containing HSA (46 g/L), IgG (12.3 9

g/L), IgA (2.2 g/L) and IgM (1.1 g/L) was filtered through a 0.22 m syringe filter for IgM separation. Chromatographic separation was performed with ÄKTA Prime Plus system. Two Tricorn 5/100 columns were packed with ABI-4FF and MMI-4FF resins, respectively.

IP T

The ABI-4FF column was first equilibrated with the equilibrium buffer (20 mM acetate buffer, pH 4.5) and then 2 mL serum sample was injected at 0.5 mL/min. IgM

SC R

mostly flowed through the column. After washing by the equilibrium buffer, HSA was eluted by sodium bicarbonate buffer (pH 9.0, 0.1 M), and then IgA was eluted by

U

glycine-HCl buffer (pH 2.2, 0.1 M). 0.1 M NaOH was used for regeneration and clean

N

in place (CIP). Meanwhile, the MMI-4FF column was use to separate IgM from the

A

flow-through fraction during ABI-4FF column separation. The MMI-4FF column was

M

equilibrated with acetate buffer (pH 4.5, 20 mM). 2 mL sample was loaded under 0.5

ED

mL/min. After washing with acetate buffer (pH 4.5, 20 mM) and sodium bicarbonate buffer (pH 9.0, 0.1 M), IgM was eluted with 10 CV glycine-HCl buffer (pH 2.2, 0.1

PT

M) with 1.0 M arginine at 1 mL/min. The elution volume was about 2~3 mL. The

CC E

chromatographic runs were monitored on-line at 280 nm. The fractions of breakthrough and elution were collected and analyzed by SEC-HPLC. The IgM purity was evaluated by the SEC-HPLC results. The samples were also sent to Adicon

A

Clinical Laboratories (Hangzhou, China) to determine the concentrations of IgG, IgA, IgM and HSA via ELISA assays. The recovery (R) of IgM from serum was calculated as following, R (%) =

Elution amount (IgM Conc. × volume) ×100% Feed amount (IgM Conc. × volume)

(8) 10

2.7. SEC-HPLC analysis The analytical SEC-HPLC was performed on the LC-3000 HPLC system (Beijing Chuangxintongheng Science & Technology, Beijing, China) with a TSKgel

IP T

G3000 SWXL column (7.8 mm × 30.0 cm, TOSOH, Tokyo, Japan). Sodium phosphate buffer (0.1 M) containing 1% isopropanol (pH 7.0) was used as the mobile

SC R

phase, and the flow rate was 0.5 mL/min. All the runs were monitored on-line at 280 nm. The purity of IgM was defined as the proportion of the peak area of IgM to the

N

U

total peak areas.

A

2.8. IgM activity test

M

The biological activity of IgM separated from human serum was assayed with an

ED

ELISA kit (SenBeiJia Biological Technology Co., Nanjing, China).

PT

3. Results and discussion

CC E

The ligand structures of the four resins used are compared in Fig. S2 (supporting information). MEP, MMI and ABI ligands have nitrogen heterocyclic rings (pyridine, imidazole and benzimidazole, respectively), which show the combination effects of

A

hydrophobic and electrostatic interactions. Target proteins can be bound by hydrophobic interactions under neutral pH, while the positively-charged nitrogen atoms at acidic condition can induce electrostatic repulsion for protein elution. Compared with ABI, the W-ABI ligand has one more tryptamine group. The indole 11

group of tryptamine might provide the orientated multimodal binding to target IgG [49]. The adsorption selectivity of IgG and Fab/Fc fragments was investigated in our previous work with the same four MMC resins. The results showed that the adsorption were obviously pH-dependent. Under the optimized loading and elution

IP T

pH conditions, IgG and Fc could be separated with high purity and recovery. It is important to investigate the influence of pH on the IgM adsorption capacity and

SC R

selectivity.

U

3.1. Adsorption behaviors of IgM and IgA on the four resins

N

IgM and IgA adsorption on MEP HyperCel, MMI-4FF, ABI-4FF and

A

W-ABI-4FF columns at pH 3.0 ~ 8.9 were investigated and compared in Figs. 1 and

M

2.

ED

Fig. 1 shows that IgM adsorption was pH dependent as that of IgG [50]. High IgM adsorption capacity was found at pH 5.0~8.0 for all resins, and pH 5.0 was the

PT

best pH tested for IgM adsorption (more than 90%). Adsorption capacity significantly

CC E

decreased when pH changed from 5.0 to 4.0, (especially for ABI-4FF), which was ~ 40% for MEP HyperCel, MMI-4FF and W-ABI-4FF, and only 20% for ABI-4FF at pH 3.0~4.0. Under alkaline conditions, IgM adsorption decreased obviously for MEP

A

HyperCel, ABI-4FF and W-ABI-4FF. However, MMI-4FF still kept high IgM adsorption (~90%) at pH 8.9. In general, ABI-4FF showed the strongest sensitivity at acidic pH and W-ABI-4FF was most sensitive at basic pH.

12

Our previous results showed that MEP HyperCel, MMI-4FF and ABI-4FF adsorbed only 10%~25% IgG at pH 5.0, and hardly any adsorption of IgG at pH 4.0. The difference between IgG and IgM adsorption might be related to the pKa of the ligands and the isoelectric point (pI) of the proteins. The pKa of MEP, MMI and ABI

IP T

ligands are around 4.8, 6.5 and 5.2, respectively [50]. The pI value of normal serum IgM is about 4.5~6.5 [52], which is more acidic than IgG (7.0~8.1). At pH 3.0~4.0,

SC R

IgM would be less positively charged than IgG that results in weaker electrostatic repulsion between IgM and positively-charged ligands. In addition, the

U

hydrophobicity of the ethyl-pyridine group of the MEP ligand (log P=1.71) and the

N

tryptophan-5-aminobenzimidazole group of the W-ABI ligand (log P =1.90) might be

A

higher than the methyl-imidazole group of the MMI ligand (log P=0.08) and the

M

benzimidazole group of the ABI ligand (log P=1.26). Hence, the adsorption of IgM on

ED

MEP HyperCel and W-ABI-4FF was relatively strong at pH 3.0. Fig. 2 shows that IgA has similar trends as IgM since the pI value of IgA is about

PT

4.5~6.8 [52,53]. High adsorption of IgA (more than 90%) was also found at pH

CC E

5.0~8.0 with MEP HyperCel, MMI-4FF and ABI-4FF. For W-ABI-4FF, IgA adsorption decreased gradually when pH was higher than 7.0 and almost no IgA was adsorbed at pH 8.9. MEP HyperCel also showed a sharp reduction on IgA adsorption

A

at pH 8.9 (less than 10%), while MMI-4FF and ABI-4FF still kept high adsorption ability even at pH 8.9. All four resins showed decrease of IgA adsorption at pH < 5.0 with different response profiles. MEP HyperCel showed strong acid-induced desorption at pH 4.0 and ABI-4FF had the lowest decrease. Compared to IgM, lower 13

IgA adsorption capacity was found at pH 3.0 for all four resins tested, which indicates that the binding force between IgA and the resins might be weaker than that of IgM. It might be explained by the molecular structures shown in Fig. S1. IgM consists of five identical subunits and might own more binding sites than IgA.

IP T

In summary, the four resins tested showed similar pH-dependent adsorption of IgG, IgM and IgA. However, the adsorption of IgM and IgA can be performed under

acidic condition should be used for IgM and IgA elution.

U

Fig. 1

SC R

more acidic environment with a relative broad pH range than IgG. Therefore, more

M

3.2. Specificity of adsorption

A

N

Fig. 2

ED

In order to better evaluate the adsorption specificity of IgM, IgA and IgG with different resins, an selectivity index (SI) as defined in Eqs. (2)-(4) was proposed. The

CC E

selectivity.

PT

results are shown in Fig. 3. The points far away from 1 means better separation

Fig 3. shows that the SI values of MEP HyperCel, MMI-4FF and ABI-4FF were

near 1 at pH 5.0~8.0 for IgM/IgA, and pH 6.0~8.0 for IgM/IgG and IgA/IgG, which

A

means the adsorption selectivity was low under those conditions. On the contrary, the SI values far away from 1 at specific pH indicated high adsorption specificity. (1) For IgM/IgA, the highest selectivity was found at pH 4.0 for ABI-4FF (IgA adsorption) and pH 8.9 for MEP HyperCel (IgM adsorption). (2) For IgM/IgG, MMI-4FF and 14

ABI-4FF showed the highest selectivity at pH 5.0. MMI-4FF and ABI-4FF could adsorb IgM efficiently, but hardly adsorb any IgG at pH 5.0. (3) For IgA/IgG, the SI values of ABI-4FF at pH 4.0, ABI-4FF and MMI-4FF at pH 5.0, MEP HyperCel and W-ABI-4FF at pH 8.9 exhibited better selectivity than other conditions.

IP T

In general, it is feasible to improve adsorption specificity and fractionate IgM, IgA and IgG by pH control due to the pH-dependent adsorption of mixed-mode resins.

SC R

The key is to choose right resins and corresponding pH conditions. The results

indicate that ABI-4FF and MMI-4FF are two promising resins for the separation of

U

IgM, IgA and IgG.

A

N

Fig. 3

M

3.3 Elution optimization of IgM with MMI-4FF

ED

Elution recovery is another important factor for chromatographic separation applications. For mixed-mode resins used in the present work, there are N-containing

PT

heterocycle in the functional ligands that would be positively charged under acidic

CC E

conditions when the solution pH is less than the pKa of the N-based groups. Therefore, electrostatic repulsion between the positively-charged ligands and the protein molecules under acidic conditions can be used to improve elution efficiency. Based on

A

the adsorption properties as shown in Fig. 1, the elution pH 3.0 were tested for the separation of IgM with MMI-4FF, but the recovery was only 25%. Therefore, the elution conditions were optimized to improve recovery.

15

Glycine buffer (pH 2.2, 0.1 M Gly-HCl) was first used and the elution recovery was improved to 40%, but still too low for practicable applications. The results indicate that the electrostatic repulsion between the positively-charged IgM and MMI ligands was not enough to elute protein due to strong binding of IgM on mixed-mode

IP T

resins. Therefore, some additives including urea, ethylene glycol, cyclodextrin (CD), arginine (Arg) and guanidinium chloride (GdmCl) were investigated to

SC R

enhance the elution process. The results were shown in Fig. 4. Urea with

concentration as high as 1.0~2.0 M could only increase the recovery by ~ 8%.

U

Ethylene glycol as a water structure breaker that can destroy hydrogen bonds [54]

N

showed limited improvement to 54.8% at high concentration (50% v/v), but the

A

concentration of 50% ethylene glycol is too high for practicable application.

M

Jia et al. [55] reported that cyclodextrin could form reversible, noncovalent

ED

inclusion complex with immobilized MEP ligand and induce the adsorbed IgG to detach from the MEP ligand over a wide pH range. In the present study,

PT

cyclodextrin was also tested

and its effect was obvious. 15 mM cyclodextrin

CC E

addition can enhance the recovery from 40% to 70.5%. However, higher cyclodextrin concentration could not further improve IgM elution. Arginine is a basic amino acid that suppresses aggregation of proteins with no

A

denaturation effects [56]. It has been used as an effective eluent to purify antibodies at neutral pH with MEP HyperCel [57], Capto adhere [58] and Capto MMC [59] resins. Kameda et al. [60] reported that alkyl group of the arginine side chain could primarily interact with the pyridine ring of the MEP ligand through hydrophobic interactions, 16

and then reduce the attraction between antibodies and MEP ligand that leads to efficient dissociation of antibodies from MEP HyperCel. As showed in Fig. 4, arginine could also promote the desorption of IgM from MMI-4FF resin. 0~0.5 M arginine significantly enhance IgM elution and the elution recovery increased slowly

IP T

with concentration increase at 0.5~1.0 M. 1.0 M arginine addition could result in the

further increase of Arg concentration was not studied.

SC R

elution recovery of ~75%. The solubility of Arg at 25 oC is 182 g/L (1.04 M), so

Guanidinium chloride (GmdCl) with guanidinium group as arginine is a strong

U

protein denaturant. The results show that the elution efficiency could be improved

N

with the increase of GmdCl concentration. IgM elution recovery with 2 M GmdCl

A

was 60.5%, which was 15% less than that of 1 M Arg. In addition, SEC analysis of

M

the eluted IgM at 1.0 M GmdCl showed an extra monomeric IgM peak in the elution

ED

(Fig. S3 in supporting information), which means GmdCl above 1.0 M would break

monomers.

PT

the inner disulfide bond of pentameric IgM and results in the degradation to

CC E

Therefore, arginine seemed to be the best elution additives and 1.0 M Arg addition at pH 2.2 glycine buffer was used for the separation of IgM in the further studies.

A

Fig. 4

3.4. Purification of monoclonal IgM from cell culture supernatant

17

Fig. 1 indicates that MEP HyperCel, MMI-4FF, ABI-4FF and W-ABI-4FF could all be used for IgM purification. The suitable separation conditions were: sample loading at pH 5.0 ~7.0 and IgM elution at acidic conditions with optimized additives. Considering the elution efficiency, ABI-4FF was selected for monoclonal IgM

IP T

purification from hybridoma cell culture supernatant (pH 7.0~7.2) with IgM titer of ~0.5 mg/mL. In order to simplify the separation process, raw feedstock without

SC R

pretreatment was loaded onto the ABI-4FF column and 1.0 M Arg in 0.1 M Gly-HCl buffer (pH 2.2) was used for IgM elution.

U

SEC-HPLC analysis of the fractions collected during IgM separation is shown in

N

Fig. 5. The results show that the purity of IgM was only 17.4% in the cell culture

A

supernatant, and the majority of the impurities were low-molecule weight fragments .

M

It was interesting that almost all impurities flowed through the ABI-4FF column, and

ED

IgM showed high purity of 99.7% in the elution pool. Tscheliessnig et al. [61] combined anion exchange chromatography and size exclusion chromatography to

PT

reach a IgM purity of 98%. Tscheliessnig et al. [62] also used PEG precipitation and

CC E

anion exchange chromatography to get the IgM purity of 84%~95% . Garbonell et. al. [33] developed the biomimic ligand hexamer peptide HWRGWV and used it for purification of IgM from human B lymphocyte cell culture supernatant, and the purity

A

of 95% was obtained with one step chromatography. Therefore, MMC with ABI-4FF resin showed a promising potential for IgM purification with high selectivity. It was found that the recovery of IgM elution with ABI-4FF was influenced by loading amount due to the relative low dynamic capacity. With 0.5 mg IgM loading 18

(1.0 mL feedstock) at the flow rate of 0.5 mL/min, the IgM recovery was about 50%, and certain amount of IgM could be found in the flow-through factions. When the loading was reduced to 0.25 mg IgM, the elution recovery increased to 80.8%. Due to the large molecule weight of IgM, the dynamic capacity on the resins was always low.

IP T

Same trends were also found for the biomimic ligand HWRGWV [33], thiophilic ligand 2-mercapto-pyridine [25] and synthetic affinity peptide TG19318 [32]. The

SC R

hydrodynamic diameter of IgM is about 24 nm [63]. Since the pore size to molecule

ratio is recommended as 10:1 for non-hindered diffusive transport [64], pores of 240

U

nm would be ideal for the adsorption of IgM. However, the mean pore size of the

N

matrix (Bestarose 4FF) is about 90 nm [65]. In order to improve dynamic capacity

A

and productivity, new matrices with large pores should be tested for large-scale

M

purification of IgM.

ED

Fig. 5

PT

3.5. Purification of IgM from human serum

CC E

Human serum is a complicated raw material that can provide more than 20 protein biologicals for treating bleeding, immunological and metabolic disorder diseases [66]. Traditional serum protein fractionation was mainly focused on albumin

A

and immunoglobulins via cold ethanol precipitation [67]. As one of major immunoglobulin fractions, the average level of IgM in human serum is around 1.0~1.5 mg/mL [68]. Unfortunately, IgM is currently discarded during the production process of intravenous immunoglobulin (IVIg). The reason might due to the difficulty 19

of serum IgM purification with low IgM concentration and high concentration of HSA and other immunoglobulins. MMC was tried to separate IgM directly from human serum in this study. Main protein components in the serum included HSA (46.0 g/L), IgG (12.3 g/L), IgA (2.2

IP T

g/L) and IgM (1.1 g/L). Based on the analysis in Section 3.2, two resins, ABI-4FF and MMI-4FF showed high adsorption selectivity of IgM, IgA and IgG, which could be

SC R

used to purify IgM from human serum. A two-step chromatographic separation

process was designed as shown in Fig. 6. The ABI-4FF resin was used as the first

U

separation step for removing IgA at pH 4.5, and IgM and IgG were in the

N

flow-through fraction. The MMI-4FF resin was then used as the second step for

A

capturing IgM at pH 4.5 from the flow-through fraction during the first step

M

separation. IgG was in the flow-through solution, and IgM was adsorbed and later

ED

eluted.

PT

Fig. 6

CC E

SEC-HPLC analysis of different factions is compared in Fig. 7. The purity of IgM in the feedstock was only about 2.3%. The main contents in the serum were HSA (60.5%), IgG (27.3%), IgA and other impurities (9.9%). Under the optimized loading

A

conditions, a great majority (more than 90% in the loading) of IgG and HSA flowed through the ABI-4FF and MMI-4FF columns. Some amount of IgM could also be found in the flow-through fractions of both ABI-4FF and MMI-4FF column. The pI of HSA (4.7~5.5) [69] is overlapped with that of IgM, and HSA could also adsorb onto 20

the mixed-mode resins as reported in our previous work [70]. Therefore, large amount of HSA in the feedstock would compete the binding site with IgM molecules at the loading pH of 4.5. The other reason might be the low dynamic capacity of IgM caused by the restricted diffusion of the large IgM molecules in resins. It could be found that

IP T

IgA and IgM were enriched in the elution fractions of Step 1# with ABI-4FF column and Step 2# with MMI-4FF column, respectively. The purity of IgM in the elution

SC R

pool was 46.8% and the purification factor was 20.3. Fig. 7 shows that HSA

concentration decreased significantly comparing with feedstock, but there was still

U

HSA (about 38.5%) in the elution which needed to be further removed.

A

N

Fig. 7

M

Our previous work indicated that HSA could be adsorbed with high capacity on

ED

ABI and MMI-based resins at pH 5.0 and the binding capacity reduced significantly at basic conditions. Moreover, Fig. 1~2 show that IgM and IgA could still be adsorbed

PT

on ABI-4FF and MMI-4FF resins at pH 8.9. Therefore, the separation process was

CC E

modified with additional washing at pH 9.0 before the elution of IgM. The process modification is also shown in Fig. 6 (dash lines). The results of SEC-HPLC analysis are compared in Fig. 8 and Table 1. It could be found that the main component in the

A

washing factions at pH 9.0 was HSA for both columns. The purity of IgM in the final elution pool increased to 65.2%, and the content of HSA decreased from 38.5% to 11.2%. The purification factor of IgM reached to 28.3 after modification. The overall recovery was 22%, and the productivity was about 0.24 mg IgM/mL serum. The 21

results reveals that mixed-mode resins own a better selectivity of IgM, IgA and IgG than the ligands reported in the literature. For the peptide ligand HWRGWV, the purity of IgM was only 37.3% even from Cohn fraction II/III with cold ethanol precipitation of human serum. The purity of IgM was improved to 46% by adding

IP T

pretreatment with caprylic acid or the combination of caprylic acid and polyethylene glycol precipitation, but the elution pool still contained 24% IgG [71]. The Protein A

SC R

mimetic peptide ligand TG19318 could purify IgM from mouse sera after a

preliminary Protein A adsorption step, but the impurities were mainly IgA and

U

retained IgG [32].

N

In order to test the influences of elution condition on the structure and biological

A

activity of IgM purified. The final elution pool in 1.0 M Arg at pH 2.2 was

M

concentrated and stored at 4 oC for 7 days, and then analyzed by SEC-HPLC. No

ED

other peak or degraded IgM could be found. In addition, the biological activity of IgM was also tested by ELISA. The IgM activity in the feedstock, first step fraction and

PT

final elution fraction was 20.6±3.1, 25.3±0.5, 22.0±2.4 U/mg, respectively.

CC E

Therefore, the activity of IgM could be maintained after the separation with ABI-4FF and MMI-4FF columns, which means that the separation conditions did not cause IgM denaturation or loss of antigen binding capacity. Hence, the process with new

A

mixed-mode resins provided a convenient, efficient and economic method for IgM purification directly from human serum. Fig. 8 Table 1 22

4. Conclusions

Four mixed-mode resins named MEP HyperCel, MMI-4FF, ABI-4FF and W-ABI-4FF were used to evaluate the adsorption selectivity of IgM, IgA and IgG.

IP T

High adsorption capacity was obtained around neutral pH, which declined under

SC R

acidic conditions for all four resins tested. However, some differences could be found for IgM, IgA and IgG. IgM showed a relative wider pH range with higher adsorption

and lower IgA adsorption was found at pH 3.0 for all four resins tested. An adsorption

U

selectivity index was introduced to evaluate the separation efficiency. The results

A

N

indicated that the control of pH was a feasible way to improve adsorption specificity

M

and fractionate IgM, IgA and IgG, and ABI-4FF and MMI-4FF showed high selectivity of IgM/IgA and IgM/IgG at pH 4.0~5.0. The elution conditions were

ED

optimized, and two application cases of IgM purification were tested. Separation of

PT

monoclonal IgM from hybridoma cell supernatant with ABI-4FF was achieved with high purity (~99%) and good recovery (80.8%). IgM was further separated directly

CC E

from human serum, and a two-step chromatographic separation process was designed with ABI-4FF and MMI-4FF subsequently. High purification factor of 28.3 was

A

obtained and IgM purity reached 65.2% after optimization. Moreover, antibody activity of IgM was maintained after the separation process. The results demonstrated that mixed-mode chromatography with specially-designed ligands can provide a convenient, efficient and economic approach for IgM purification directly from complex feedstock, which is promising for industrial applications. 23

Acknowledgements This work was financially supported by the National Natural Science Foundation of China and the International Science& Technology Cooperation Program of China.

SC R

Nanjing, China) for providing the monoclonal human IgM supernatant.

IP T

The authors would also want to thank Dr. Bo Tang (Vazyme Biotech Co., Ltd,

References

U

[1] N.L. Anderson, N.G. Anderson, The human plasma proteome, Mol. Cell

N

Proteomics 1 (2002) 845-867.

A

[2] J.W. Stoop, P.C. Sander, R.E. Ballieux, Serum immunoglobulin levels in healthy

M

children and adults, Clin. Exp. Immunol. 4 (1969) 101-112.

ED

[3] R.H. Buckley, S.A. Fiscus, Serum IgD and IgE concentrations in immunodeficiency diseases, J. Clin. Invest. 55 (1975) 157-165.

PT

[4] F. Listi, G. Candore, M.A. Modica, M. Russo, G. Di Lorenzo, M.

CC E

Esposito-Pellitteri, G. Colonna-Romano, A. Aquino, M. Bulati, D. Lio, C. Franceschi, C. Caruso, A study of serum immunoglobulin levels in elderly

A

persons that provides new insights into B cell immunosenescence, Ann. N. Y. Acad. Sci. 1089 (2006) 487-495.

[5] G. Vidarsson, G. Dekkers, T. Rispens, IgG subclasses and allotypes: from structure to effector functions, Frontiers in immunology 5 (2014) 1-16.

24

[6] J.M. Woof, M.W. Russell, Structure and function relationships in IgA, Mucosal Immunol. 4 (2011) 590-597. [7] V.B. Klimovich, IgM and its receptors: structural and functional aspects, Biochemistry Mosc. 76 (2011) 534-549.

antibody therapeutics, Nat. Biotechnol. 25 (2007) 1421-1434.

IP T

[8] W.A. Marasco, J. Sui, The growth and potential of human antiviral monoclonal

immunotherapy, Immunotherapy 1 (2009) 241-248.

SC R

[9] H.P. Vollmers, S. Brändlein, Natural human immunoglobulins in cancer

U

[10] L. Rasche, J. Duell, I.C. Castro, V. Dubljevic, M. Chatterjee, S. Knop, F. Hensel,

N

A. Rosenwald, H. Einsele, M.S. Topp, S. Brandlein, GRP78-directed

A

immunotherapy in relapsed or refractory multiple myeloma results from a phase

M

1 trial with the monoclonal immunoglobulin M antibody PAT-SM6,

ED

Haematologica 100 (2015) 377-384.

[11] M.P. Horn, A.W. Zuercher, M.A. Imboden, M.P. Rudolf, H. Lazar, H. Wu, N.

PT

Hoiby, S.C. Fas, A.B. Lang, Preclinical in vitro and in vivo characterization of

CC E

the fully human monoclonal IgM antibody KBPA101 specific for Pseudomonas aeruginosa serotype IATS-O11, Antimicrob. Agents Chemother. 54 (2010) 2338-2344.

A

[12] M.R. Ehrenstein, C.A. Notley, The importance of natural IgM: scavenger, protector and regulator, Nat. Rev. Immunol. 10 (2010) 778-786. [13] L.L. Ilag, Immunoglobulin M as a vaccine adjuvant, Med. Hypotheses 77 (2011) 473-478. 25

[14] Q.E. Yang, IgM, not IgG, a key for HIV vaccine, Vaccine 27 (2009) 1287-1288. [15] A.H. Tatum, Large scale recovery of biologically active IgM (95% pure) from human plasma obtained by therapeutic plasmapheresis, J. Immunol. Methods 158 (1993) 1-4.

IP T

[16] G. Coppola, J. Underwood, G. Cartwright, M.T.W. Heam, High-performance liquid chromatography of amino acids, peptides and proteins: XCIII. Comparison

SC R

of methods for the purification of mouse monoclonal immunoglobulin M autoantibodies, J. Chromatogr. A 476 (1989) 269-290.

U

[17] F. Steindl, A. Jungbauer, E. Wenisch, G. Himmler, H. Katinger, Isoelectric

N

precipitation and gel chromatography for purification of monoclonal IgM,

A

Enzyme Microb. Technol. 9 (1987) 361-364.

M

[18] F. Perosa, R. Carbone, S. Ferrone, F. Dammacco, Purification of human

ED

immunoglobulins by sequential precipitation with caprylic acid and ammonium sulphate, J. Immunol. Methods 128 (1990) 9-16.

PT

[19] E. Johnson, L. Miribel, P. Arnaud, K.Y. Tsang, Purification of IgM monoclonal

CC E

antibody from murine ascitic fluid by a two-step column chromatography procedure, Immunol. Lett. 14 (1987) 159-165.

A

[20] S.H. Mahassni, D.G. Klapper, R.G. Hiskey, Purification of a murine IgM monoclonal antibody, Hybridoma 28 (2009) 189-197.

[21] I. Tomoea, I. Titlestadb, K. Kejlingb, K. Erba, H.J. Ditzelac, J.C. Jenseniusd, Pilot scale purification of human monoclonal IgM (COU-1) for clinical trials, J. Immunol. Methods 205 (1997) 11-17. 26

[22] E. McCarthy, G. Vella, R. Mhatre, Y.P. Lim, Rapid purification and monitoring of immunoglobulin M from ascites by perfusion ion-exchange chromatography, J. Chromatogr. A 743 (1996) 163-170. [23] P. Brne, A. Podgornik, K. Bencina, B. Gabor, A. Strancar, M. Peterka, Fast and

monolithic columns, J. Chromatogr. A 1144 (2007) 120-125.

IP T

efficient separation of immunoglobulin M from immunoglobulin G using short

SC R

[24] D. Roggenbuck, U. Mark, S.T. Kiessig, G. Schoenherr, S. Jahn, T. Porstmann, Purification and immunochemical characterization of a natural human

U

polyreactive monoclonal IgM antibody, J. Immunol. Methods 167 (1994)

N

207-218.

A

[25] A. Bergenstrühle, A. Heijbel, One step purifications of IgM and IgY with HiTrap

M

purification columns, Amersham Biosciences Life Science News 3 (1999) 1-2.

ED

[26] L.H. Stanker, M. Vanderlaan, H.J. Salinas, One-step purification of mouse monoclonal antibodies from ascites fluid by hydroxylapatite chromatography, J.

PT

Immunol. Methods 76 (1985) 157-169.

CC E

[27] A.J. Henniker, K.F. Bradstock, Purification of two murine monoclonal antibodies of the IgM class by hydroxylapatite chromatography, Biomed. Chromatogr. 7 (1993) 121-125.

A

[28] J.R. Nevens, Affinity chromatographic purification of immunoglobulin M antibodies utilizing immobilized mannan binding protein, J. Chromatogr. A 597 (1992) 247-256.

27

[29] N. Shibuya, J.E. Berry, I.J. Goldstein, One-step purification of murine IgM and human α2-macroglobulin by affinity chromatography on immobilized snowdrop bulb lectin, Arch. Biochem. Biophys. 267 (1988) 676-680. [30] A. Nethery, R.L. Raison, S.B.E. Smith, Single-step purification of

IP T

immunoglobulin M on C1q-Sepharose, J. Immunol. Methods 126 (1990) 57-60. [31] B.H.K. Nilson, L. logdberg, W. Kastern, L. Bjorck, B. Akerstrom, Purification of

SC R

antibodies using protein L-binding framework structures in the light chain variable domain, J. Immunol. Methods 164 (1993) 33-40.

U

[32] G. Palombo, A. Verdoliva, G. Fassina, Affinity purification of immunoglobulin

N

M using a novel synthetic ligand, J. Chromatogr. B 715 (1998) 137-145.

A

[33] Z. Liu, P.V. Gurgel, R.G. Carbonell, Affinity chromatographic purification of

M

human immunoglobulin M from human B lymphocyte cell culture supernatant,

ED

Biochem. Eng. J. 70 (2013) 63-70.

[34] Z. Liu, P.V. Gurgel, R.G. Carbonell, Effects of peptide density and elution pH on

PT

affinity chromatographic purification of human immunoglobulins A and M, J.

CC E

Chromatogr. A 1218 (2011) 8344-8352. [35] H. Yang, P.V. Gurgel, R.G. Carbonell, Hexamer peptide affinity resins that bind the Fc region of human immunoglobulin G, J. Pept. Res. 66 (2006) 120-137.

A

[36] G. Palombo, S. De Falco, M. Tortora, G. Cassani, G. Fassina, A synthetic ligand for IgA affinity purification, J. Mol. Recognit. 11 (1998) 243-246.

28

[37] G. Palombo, M. Rossi, G. Cassani, G. Fassina, Affinity purification of mouse monoclonal IgE using a protein a mimetic ligand (TG19318) immobilized on solid supports, J. Mol. Recognit. 11 (1998) 247-249. [38] S. Gautam, K.C. Loh, Human pIgR mimetic peptidic ligand for affinity

IP T

purification of IgM mimetic Part I: Ligand design and binding mechanism, Sep. Purif. Technol. 102 (2013) 173-179.

SC R

[39] S. Gautam, K.C. Loh, Human pIgR mimetic peptidic ligand for affinity

purification of IgM Part II: Ligand binding characteristics, Sep. Purif. Technol.

U

102 (2013) 43-49.

N

[40] Y. Yang, X. Geng, Mixed-mode chromatography and its applications to

A

biopolymers, J. Chromatogr. A 1218 (2011) 8813-8825.

M

[41] G. Zhao, X.Y. Dong, Y. Sun, Ligands for mixed-mode protein chromatography:

ED

Principles, characteristics and design, J. Biotechnol. 144 (2009) 3-11. [42] S.C. Burton, D.R.K. Harding, Hydrophobic charge induction chromatography

PT

salt independent protein adsorption and facile elution with aqueous buffers, J.

CC E

Chromatogr. A 814 (1998) 71-81. [43] W. Schwartz, D. Judd, M. Wysocki, L. Guerrier, E. Birck-Wilson, E. Boschetti,

A

Comparison of hydrophobic charge induction chromatography with affinity chromatography on protein A for harvest and purification of antibodies, J. Chromatogr. A 908 (2001) 252-263.

29

[44] K. Kallberg, H.O. Johansson, L. Bulow, Multimodal chromatography: An efficient tool in downstream processing of proteins, Biotechnol. J. 7 (2012) 1485-1495. [45] I.F. Pinto, M.R. Aires-Barros, A.M. Azevedo, Multimodal chromatography

IP T

debottlenecking the downstream processing of monoclonal antibodies, Pharm. Bioprocess 3 (2015) 263-279.

SC R

[46] F.M. Chen, G.S. Naeve, A.L. Epstein, Comparison of mono Q, superose-6, and ABx fast protein liquid chromatograhy for the purification of IgM Monoclonal

U

antibodies, J. Chromatogr. A 444 (1988) 153-164.

N

[47] D.Q. Lin, H.F. Xia, L.P. Wang, Z.J. Chen, S.J. Yao, Preparation and evaluation of

M

Chem. Res. 47 (2008) 9566-9572.

A

cellulose adsorbents for hydrophobic charge induction chromatography, Ind. Eng.

ED

[48] J. Yan, Q.L. Zhang, H.F. Tong, D.Q. Lin, S.J. Yao, Hydrophobic charge-induction resin with 5-aminobenzimidazol as the functional ligand:

PT

Preparation, protein adsorption and immunoglobulin G purification, J. Sep. Sci.

CC E

38 (2015) 2387-2393.

[49] H.F. Tong, D.Q. Lin, W.N. Chu, Q.L. Zhang, D. Gao, R.Z. Wang, S.J. Yao,

A

Multimodal charge-induction chromatography for antibody purification, J. Chromatogr. A 1429 (2016) 258-264.

[50] Y.D. Luo, Q.L. Zhang, X.M. Yuan, W. Shi, S.J. Yao, D.Q. Lin, Selectivity evaluation and separation of human immunoglobulin G, Fab and Fc fragments with mixed-mode resins, J. Chromatogr. B 1040 (2017) 105-111. 30

[51] H.L. Lu, D.Q. Lin, D. Gao, S.J. Yao, Evaluation of immunoglobulin adsorption on the hydrophobic charge-induction resins with different ligand densities and pore sizes, J. Chromatogr. A 1278 (2013) 61-68. [52] F. Chiodi, A. Siden, E. Osby, Isoelectric focusing of monoclonal

IP T

immunoglobulin G, A and M followed by detection with the avidin-biotin system, Electrophoresis 6 (1985) 124-128.

SC R

[53] R.C. Monteiro, L.H. Mecarelli, M.C. Roque-Barreira, L.H. Noel, J. Berger, P.

Lesavre, Charge and size of mesangial IgA in IgA nephropathy, Kidney Int. 28

U

(1985) 666-671.

N

[54] A. El-Kak, S. Manjini, M.A. Vijayalakshmi, Interaction of immunoglobulin G

M

604 (1992) 29-37.

A

with immobilized histidine: mechanistic and kinetic aspects, J. Chromatogr. A

ED

[55] J. Ren, P. Yao, Y. Cao, J. Cao, L. Zhang, Y. Wang, L. Jia, Application of cyclodextrin-based eluents in hydrophobic charge-induction chromatography:

PT

elution of antibody at neutral pH, J. Chromatogr. A 1352 (2014) 62-68.

CC E

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

A

[57] T. Arakawa, Y. Kita, H. Sato, D. Ejima, MEP chromatography of antibody and Fc-fusion protein using aqueous arginine solution, Protein Expr. Purif. 63 (2009) 158-163.

31

[58] A. Hirano, T. Arakawa, T. Kameda, Effects of arginine on multimodal anion exchange chromatography, Protein Expr. Purif. 116 (2015) 105-112. [59] A. Hirano, T. Arakawa, T. Kameda, Interaction of arginine with Capto MMC in multimodal chromatography, J. Chromatogr. A 1338 (2014) 58-66.

IP T

[60] A. Hirano, T. Maruyama, K. Shiraki, T. Arakawa, T. Kameda, Mechanism of protein desorption from 4-mercaptoethylpyridine resins by arginine solutions, J.

SC R

Chromatogr. A 1373 (2014) 141-148.

[61] J. Lee, A. Tscheliessnig, A. Chen, Y.Y. Lee, G. Adduci, A. Choo, A. Jungbauer,

U

Adaptation of hybridomas to protein-free media results in a simplified two-step

N

immunoglobulin M purification process, J. Chromatogr. A 1216 (2009)

A

2683-2688.

M

[62] A. Tscheliessnig, D. Ong, J. Lee, S. Pan, G. Satianegara, K. Schriebl, A. Choo, A.

ED

Jungbauer, Engineering of a two-step purification strategy for a panel of monoclonal immunoglobulin M directed against undifferentiated human

PT

embryonic stem cells, J. Chromatogr. A 1216 (2009) 7851-7864.

CC E

[63] R.R. Akhouri, S. Goel, H. Furusho, U. Skoglund, M. Wahlgren, Architecture of human IgM in complex with P. falciparum Erythrocyte membrane protein, Cell Rep. 14 (2016) 723-736.

A

[64] A. Jungbauer, Chromatographic media for bioseparation, J. Chromatogr. A 1065 (2005) 3-12. [65] L. Hagel, M. Ostberg, T. Andersson, Apparent pore size distributions of chromatography media, J. Chromatogr. A 743 (1996) 33-42. 32

[66] T. Burnouf, Modern plasma fractionation, Transfus. Med. Rev. 21 (2007) 101-117. [67] T. Burnouf, Current status and new developments in the production of plasma derivatives, ISBT Sci. Ser. 11 (2016) 18-25.

IP T

[68] J.L. Fahey, E.M. Mckelvey, Quantitative determination of serum immunoglobulins in antibody-agar plates, J. Immunol. 94 (1965) 84-90.

SC R

[69] G. Va, A. Glu, Isoelectric fractions of healthy human serum albumin and their ability to bind bilirubin, Ukr. Biokhim. Zh. 50 (1978) 735-738.

U

[70] Q.C. Wu, D.Q. Lin, W. Shi, Q.L. Zhang, S.J. Yao, A mixed-mode resin with

N

tryptamine ligand for human serum albumin separation, J. Chromatogr. A 1431

A

(2016) 145-153.

M

[71] Z. Liu, P.V. Gurgel, R.G. Carbonell, Purification of human immunoglobulins A,

ED

G and M from Cohn fraction II/III by small peptide affinity chromatography, J.

A

CC E

PT

Chromatogr. A 1262 (2012) 169-179.

33

Figure Captions

A

N

U

SC R

IP T

Fig. 1. Adsorption of IgM on the four resins at different loading pH values.

A

CC E

PT

ED

M

Fig. 2. Adsorption of IgA on the four resins at different loading pH values.

34

Fig. 3. Selectivity index of IgM/IgA (a), IgM/IgG (b) and IgA/IgG (c) of the four

A

CC E

PT

ED

M

A

N

U

SC R

IP T

resins under different loading pH values.

35

36

A ED

PT

CC E

IP T

SC R

U

N

A

M

Fig. 4. Effects of urea, ethlylene glycol, -cyclodextrin and arginine on the elution

A

CC E

PT

ED

M

A

N

U

SC R

IP T

proportion of IgM.

37

Fig. 5. SEC-HPLC analysis for IgM purification from cell culture supernatant with ABI-4FF resin. Loading 0.5 mL at pH 7.0 and elution with 1.0 M Arg at pH

A

CC E

PT

ED

M

A

N

U

SC R

IP T

2.2.

38

Fig. 6. Process scheme of IgM purification from human serum. The dash lines

A

CC E

PT

ED

M

A

N

U

SC R

IP T

describe process modification with washing at pH 9.0 to remove HSA.

39

Fig. 7. SEC-HPLC analysis of the factions of IgM purification from human serum using a two-step chromatographic process. First step: ABI-4FF column, loading at pH 4.5 and elution at pH 2.2. Second step: MMI-4FF column,

A

CC E

PT

ED

M

A

N

U

SC R

IP T

loading at pH 4.5 and elution with 1M Arg at pH 2.2.

40

Fig. 8. SEC-HPLC analysis of the factions of IgM purification from human serum using a two-step chromatographic process with the additional washing. First step: ABI-4FF column, loading at pH 4.5, washing at pH 9.0 and elution at pH 2.2. Second step: MMI-4FF column, loading at pH 4.5, washing at pH 9.0 and

CC E

PT

ED

M

A

N

U

SC R

IP T

elution with 1M Arg at pH 2.2.

Table 1. Purity and recovery of IgM separated from human serum with ABI-4FF and

A

MMI -4FF resins with or without additional washing step.

IgM Recovery

IgM Purity

(%)

(%)

Purification Factor

41

100

2.3±1.2

First step

77.4±3.6

5.2±1.2

Second step without washing

33.6±3.4

46.8±4.5

20.3 ± 3.1

Second step with washing

28.2±1.6

65.2±3.2

28.3±2.2

A

CC E

PT

ED

M

A

N

U

SC R

IP T

Feedstock

42