the impact of cell adaptation to serum-free conditions on the glycosylation profile of a monoclonal...

ResearchPap

er

New Biotechnology �Volume 00, Number 00 �December 2012 RESEARCH PAPER

The impact of cell adaptation to serum-free conditions on the glycosylationprofile of a monoclonal antibodyproduced by Chinese hamster ovary cellsAna Rita Costa1, Joanne Withers2, Maria Elisa Rodrigues1, Niaobh McLoughlin2,Mariana Henriques1,, Rosario Oliveira1, Pauline M. Rudd2 and Joana Azeredo1

1 IBB-Institute for Biotechnology and Bioengineering, Centre of Biological Engineering, University of Minho, Campus de Gualtar 4710-057, Braga, Portugal2NIBRT Dublin-Oxford Glycobiology Laboratory, National Institute for Bioprocessing Research and Training, Fosters Avenue, Mount Merrion, Blackrock, Co. Dublin, Ireland

N-glycosylation is one of the most crucial parameters affecting the biological activity of therapeutic

monoclonal antibodies (mAbs), and should therefore be closely monitored and controlled to guarantee a

consistent and high-quality product in biopharmaceutical processes. In the present work, the effect of

the time-consuming step of gradual cell adaptation to serum-free conditions on the glycosylation profile

of a mAb produced by CHO-K1 cells was evaluated. High-performance liquid chromatography analysis

revealed important changes in mAb glycosylation patterns in all steps of serum reduction. These changes

could be grouped in two distinct phases of the process of adaptation: middle (2.5 to 0.15% serum) and

final (0.075 and 0% serum). For intermediate levels of serum, a desirable increase of galactosylation and

decrease of fucosylation, but an undesirable increase in sialylation were observed; while the inverse was

obtained at the final stages of adaptation. These divergences may be related to the reduction of serum

supplementation, to variations in the levels of cell density and viability achieved at these stages, and to

the natural shift of the cell growth mode during adaptation from adherent to suspended. The divergent

glycan profiles obtained in this study demonstrate a strong influence of the adaptation process on mAb

glycosylation, suggesting that control and monitoring of product quality should be implemented at the

early stages of process development.

IntroductionImmunoglobulin G (IgG) monoclonal antibodies (mAbs) are cur-

rently one of the most successful drug classes in the biopharma-

ceutical industry [1–4]. For the production of these therapeutic

antibodies, mammalian cells are the preferred system mainly

owing to their innate ability to perform post-translational proces-

sing that closely resembles that found in humans [1,5–7]. Of the

post-translational modifications, glycosylation is one of the most

crucial for protein quality [7]. Particularly, the glycosylation of the

Fc region of the IgG (containing a bi-antennary N-glycan at

the highly conserved asparagine 297 residue in each of the

CH2 domains [8–10]) plays essential roles in structural integrity,

Please cite this article in press as: Costa, A.R. et al., The impact of cell adaptation to serum-frhamster ovary cells, New Biotechnol. (2013), http://dx.doi.org/10.1016/j.nbt.2012.12.002

Corresponding author: Henriques, M. ([email protected])

1871-6784/$ - see front matter � 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.nbt.2012.1

effector functions, immunogenicity, plasmatic clearance, solubi-

lity, and resistance toward proteases [7,11–20].

These functions and, therefore, the therapeutic efficacy of the

mAb are also dependent on the specific glycoforms present [1,21],

which are highly variable owing to the different types of mono-

saccharides available and their combinational pairing [11,22]. For

example, the levels of terminal galactose, core fucose, sialic acid,

bisecting N-acetylglucosamine (GlcNAc), and high mannose spe-

cies are known important glycosylation elements with impact on

antibody effector functions [23–26]. Indeed, higher levels of galac-

tosylation [27–30] and bisecting GlcNAc [11,31], as well as reduced

core fucosylation [24,32,33] and sialylation [34–37], have been

shown to enhance the clinical efficacy of mAbs that exert their

therapeutic effect by antibody-dependent cellular cytotoxicity

(ADCC) or complement dependent cytotoxicity (CDC) mediated

ee conditions on the glycosylation profile of a monoclonal antibody produced by Chinese

2.002 www.elsevier.com/locate/nbt 1

http://dx.doi.org/10.1016/j.nbt.2012.12.002

mailto:[email protected]


RESEARCH PAPER New Biotechnology � Volume 00, Number 00 �December 2012

NBT-572; No of Pages 10

Research

Pap

er

killing. This is particularly relevant for core fucosylation, whose

absence can improve ADCC activity by up to 100-fold [24]. Addi-

tionally, the presence of high mannose structures has been asso-

ciated with reduced ADCC and CDC [26], and with higher

clearance rates [27].

Affecting the glycan profile of glycosylated therapeutics are the

host cell expression system [20,38,39], environmental factors such

as dissolved oxygen levels [40,41], pH [42,43], and temperature

[44,45], as well as cell culture media and mode of culture (e.g.

batch, fed-batch, and continuous) [38,46–50].

Concerning the cell culture media, this is typically a mixture of

50–100 chemically defined components and a handful of unde-

fined components that are sometimes supplemented [7]. Of these,

serum is the most common for the positive effects on mammalian

cell growth. However, owing to regulatory and safety issues, serum

and animal-derived components need to be avoided in the bio-

pharmaceutical industry [51,52], where chemically defined and

serum-free media are considered fundamental to the manufacture

of highly purified products of consistent quality [28]. Conse-

quently, mammalian cells that typically grow in serum-containing

media need to go through a process of adaptation to culture in

serum-free conditions, usually consisting of progressive reductions

of serum concentration in the medium, and additional supple-

mentation with commercial serum substitutes, defined proteins,

or small molecules such as amino acids or trace elements [53].

However, serum-free adaptation is very time-consuming and can

affect cell growth and product yield, as well as product quality (e.g.

glycosylation profile) [53–56].

To better understand the effects of such medium changes on

product glycosylation, the present work reports an analysis of the

glycosylation profile of a mAb produced by CHO-K1 cells during

the different steps encompassed in a gradual process of adaptation

to serum-free culture, involving the test of different combinations

of supplements (insulin and trace elements). The potential impact

of variations of the mAb glycosylation pattern on its biological

activity is also discussed.

Materials and methodsCell adaptation to serum-free conditionsCell line

A CHO-K1 cell line (ATCC CCL-61) previously transfected for the

production of a recombinant human mAb was used [57]. Cultures

were initiated with cells at passage 16, using Dulbecco’s modified

Eagle’s medium (DMEM, Sigma–Aldrich) supplemented with 10%

fetal bovine serum (FBS, Sigma–Aldrich) and 1� hypoxanthine-

aminopterin-thymidine (HAT, Sigma–Aldrich), at 378C and 5% CO2.

Adaptation to serum-free conditionsThe mAb-producing CHO-K1 cells were gradually adapted from

growth in the serum-containing DMEM medium to growth in

serum-free conditions. For this, cells were seeded into 25 cm2

culture flasks at a viable density of 4 � 105 cells/mL, using the

culture conditions described above, with additional supplementa-

tion with the following combinations of insulin and trace ele-

ments (all from Sigma–Aldrich): C1 – rhInsulin, copper sulfate,

zinc sulphate, sodium selenite, and ferrous ammonium sulfate; C2

– rhInsulin, copper sulfate, zinc sulfate, sodium selenite, ferrous

ammonium sulfate, ammonium metavanadate, nickel chloride,


2 www.elsevier.com/locate/nbt

and stannous chloride; and C3 – rhInsulin, ammonium metava-

nadate, nickel chloride, and stannous chloride. These supplements

were tested due to their potential positive effect on cell survival

during the process of adaptation.

The serum percentage supplemented in the medium was then

reduced stepwise from the initial 10% to the stages of 5, 2.5, 1.25,

0.625, 0.31, 0.15, 0.075 and 0%. At the stage of 0.625% serum,

the basal culture medium was gradually switched from DMEM to

the chemically defined and serum-free EXCELL CHO DRHF-

medium (Sigma–Aldrich). According to the information pro-

vided by the manufacturer, the formulation of EXCELL includes

inorganic salts, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic

acid (HEPES), sodium bicarbonate, essential and non-essential

amino acids, vitamins, recombinant human insulin, plant

hydrolysates, trace elements and other organic compounds.

The remaining steps of adaptation were performed using EXCELL

as the growth medium. Cell concentration and viability was

evaluated at each step using the trypan blue (Sigma–Aldrich)

exclusion method and a hemocytometer.

N-glycosylation analysisDuring the process of adaptation to serum-free conditions, several

samples were taken for the assessment of the glycosylation profile

of the mAb produced, as described in Table 1. Sampling was

performed when renewing the culture medium (every two days)

and consisted of the removal of all culture supernatant while cells

were growing adherently and of half the culture supernatant when

cells started growing suspended; samples were frozen immediately

after being collected and were all analyzed for mAb glycosylation

at the same time.

IgG purificationBefore IgG purification, samples were concentrated to volumes of

200 mL using 15 mL spin concentrators of 5 kDa molecular weight

cutoff (Agilent Technologies). Then, the IgG from the concen-

trated samples was purified with Protein A spin plates (Thermo

Scientific) according to manufacturer instructions, with slight

modifications. Briefly, the wells of a spin plate were equilibrated

with phosphate buffered saline (PBS, pH 7.2). Samples were diluted

twofold in PBS, added to the plates, and the plates incubated for

30 min with moderate agitation. The resin was washed with PBS,

and the purified IgG eluted (0.5 M acetic acid, pH 2.5) and neu-

tralized (1 M ammonium bicarbonate, pH 7–8.5). The elutions

containing the purified antibody were dried overnight.

In gel block immobilizationThe dried samples were reduced with a solution containing

13.88 mM sodium dodecyl sulfate (SDS), 12.5 mM Tris (pH 6.6)

and 0.05 M dithiothreitol (DTT) for 15 min at 658C; and alkylated

with 100 mM iodoacetamide for 30 min in the dark. The gel blocks

were formed by adding a mixture of 22.5 mL Protogel 30%, 11.25 mL

1.5 mM Tris, 1 mL 35 mM SDS, and 1 mL 34.7 mM ammonium

peroxisulphate solution (APS), and finally adding 1 mL tetramethy-

lethylenediamine (TEMED) to the samples, letting set for 15 min.

N-glycan release and fluorescent labelingEach gel block was cut into small pieces (�2 mm2), washed alter-

nately with acetonitrile and 20 mM sodium bicarbonate (pH 7),





TABLE 1

Description of the samples taken at different stages of the process of adaptation of CHO-K1 cells to growth in serum-free conditions, forthe assessment of the N-glycosylation profile of the monoclonal antibody produced

Stage of serum Medium Days of culture Sampling day Combinations of supplements Sample

10% DMEM 1 (0–1) 1 None D 10%

5% DMEM 2 (1–3) 3 C1, C2, C3 –

2.5% DMEM 9 (3–12) 12 C1 C1 D 2.5%

C2 C2 D 2.5%

C3 C3 D 2.5%

1.25% DMEM 8 (12–20) – C1, C2, C3 –

0.625% DMEM 4 (20–24) – C1, C2 –24 C3 C3 D 0.625%

75% DMEM: 25% EXCELL 2 (24–26) – C1, C2, C3 –

50% DMEM: 50% EXCELL 2 (26–28) – C1, C2, C3 –

25% DMEM: 75% EXCELL 2 (28–30) – C1, C2, C3 –EXCELL 36 (30–66) 66 C1 C1 E 0.625%

C2 C2 E 0.625%

C3 C3 E 0.625%

0.31%a EXCELL 101 (66–167) – C1, C2b –86 C3 C3 E 0.31% i

167 C3 E 0.31% f

0.15% EXCELL 27 (167–194) 194 C3 C3 E 0.15%

0.075% EXCELL 43 (194–237) 237 C3 C3 0.075%

0% EXCELL 43 (237–280) 280 C3 C3 E 0%a This stage was the most time-consuming, so samples at the initial (i) and final (f ) phases were assessed.b Cells growing in combinations C1 and C2 died at this stage.

ResearchPap

er

and dried in a vacuum centrifuge. For the release of N-linked

glycans, the dried gel pieces were incubated for 15 min in a mix

of 2 mL of PNGase F (Prozyme) and 48 mL of 20 mM sodium

bicarbonate (pH 7), and a further 12–16 hours at 378C after adding

extra 50 mL of 20 mM sodium bicarbonate. The supernatant was

then removed (13,000 rpm, 5 min), and the samples subjected to a

series of washes with water and acetonitrile, with a 15-min sonica-

tion for each wash. The supernatants obtained were collected and

dried in a vacuum centrifuge. The dried samples were redissolved

in 20 mL formic acid (1%), incubated for 40 min at room tempera-

ture and dried in a vacuum centrifuge.

The released N-glycans were labeled for fluorescent detection

using the LudgerTagTM 2-aminobenzamide (2AB) Glycan Labeling

Kit (Ludger), according to manufacturer instructions. Briefly, 5 mL

of the 2AB labeling mix were added to each dried glycan sample,

incubated for 30 min at 658C, vortexed, spun down, and incubated

for further 90 min.

Excess 2AB label from labeled N-glycans was cleaned-up using

Normal Phase 1 PhyNexus tips (PhyNexus). Briefly, samples were

diluted in 95 mL water and 900 mL acetonitrile. The PhyNexus tips

were prepared by a series of ten uptake washes with 95% acetoni-

trile (v/v in water), 20% acetonitrile (v/v in water) and again 95%

acetonitrile. Samples were loaded into the PhyNexus tips through

ten in-out cycles, followed by ten uptake washes with 95% acet-

onitrile. The glycans were then eluted with five uptake rounds of

20% acetonitrile, and the elutions collected and dried in a vacuum

centrifuge.

Normal phase-high performance liquid chromatographyFor normal phase-high performance liquid chromatography (NP-

HPLC), each dried sample was ressuspended in 20 mL water and


80 mL acetonitrile. NP-HPLC was performed using a TSKgel Amide-

80 3 mm (150 mm � 4.6 mm) column (Tosoh Bioscience) for 60-

min runs, at 308C, using 50 mM ammonium formate as Solvent A

and acetonitrile as Solvent B. The runs were performed on a 2695

Alliance separations module (Waters) with a 2475 multi-wave-

length fluorescence detector (Waters), with excitation and emis-

sion wavelengths at 330 and 420 nm, respectively. Conditions of

the 60 min method were a linear gradient of 35 to 47% Solvent A

over 48 min at a flow rate of 0.48 mL/min, followed by a minute at

47 to 100% Solvent A and 4 min at 100% Solvent A, returning to

35% solvent A over 1 min and then finishing with 35% solvent A

for 6 min.

The systems were calibrated by running an external standard of

2-AB dextran ladder (2-AB labeled glucose homopolymer) along-

side the sample runs.

Processing of samplesA fifth-order polynomial distribution curve was fitted to the dex-

tran ladder and used to allocate glucose unit (GU) values from

retention times, using Empower GPC software (Waters). Tentative

assignment of structures to the peaks was then made by matching

the GU values obtained with those in GlycoBase (http://glycoba-

se.nibrt.ie/), and using the known IgG1 profile as a guide. A

confirmation of the structures through exoglycosidase digestion

was not possible due to a limited amount of sample. Nevertheless,

for the purposes of comparison of this study, such final allocation

was not essential.

After structure assignment, the composition of the samples was

characterized in terms of: agalactosylated (G0), monogalactosy-

lated (G1), digalactosylated (G2), trigalactosylated (G3), galacto-

sylated (G1+G2+G3), core fucosylated (F), core fucosylated


www.elsevier.com/locate/nbt 3




Research

Pap

er

agalactosylated (FG0), core fucosylated monogalactosylated (FG1),

core fucosylated digalactosylated (FG2), core fucosylated trigalac-

tosylated (FG3), monosialylated (S1), disialylated (S2), trisialylated

(S3), sialylated (S1+S2+S3), monosialylated monogalactosylated

(S1G1), monosialylated digalactosylated (S1G2), disialylated diga-

lactosylated (S2G2), disialylated trigalactosylated (S2G3), high

mannose, and complex glycan structures. These characteristics

were determined by summing the areas of the peaks pertaining

to each of the abovementioned structures.

Results and discussionThe adaptation of the mAb-producing CHO-K1 cells to growth in

serum-free conditions has proven to be time-consuming (taking

around 280 days) and its success depended on the combination of

supplements used. Indeed, it was only possible to achieve a full

adaptation using the combination of supplements C3, whereas for

combinations C1 and C2 cell death occurred after reaching a

serum level of 0.625% in EXCELL. Of the supplements used in

the combinations, it is known that iron, copper and zinc play an

important role in several mechanisms that influence cell growth,

differentiation and development, and that selenite acts as an

antioxidant [58]; the remaining trace elements used have been

associated with many functions in vivo but their effects in vitro have


FIGURE 1

N-glycosylation profile of the monoclonal antibody produced by CHO-K1 cells cultu

free medium. D – DMEM, E – EXCELL, Cx – combination of supplements x (x = 1, 2medium.


not been fully disclosed [59]. It is interesting to note, therefore,

that in this study the combinations using the most known ele-

ments (C1 and C2) had the worst performance. This may have

resulted of antagonistic effects caused by the interaction of the

elements used in these combinations, or from the presence of the

element of iron in both C1 and C2, which is absent in C3. Iron is a

crucial component in cell culture but its uptake by cells is complex

and typically mediated by transferrin [60–62]. This protein is

present in the serum supplemented to the culture medium but

is absent in the EXCELL medium, so it is possible that cells growing

in EXCELL supplemented with iron-containing C1 and C2 at lower

percentages of serum had suffered from oxidative stresses caused

by uncontrolled uptake of iron by the cells [62,63].

As a consequence of the results obtained with C1 and C2, the

glycosylation analysis of the mAb produced in medium supple-

mented with these two combinations was only performed for two

stages of adaptation: D 2.5% and E 0.625%.

For combination C3, the stage of E 0.31% was the most demand-

ing, requiring the longest time (101 days) for cells to acquire a

healthy growth, and causing a shift on the cell growth mode from

adherent to suspension. These modifications could have an effect

on mAb glycosylation, so samples were taken at two distinct

periods at this stage (initial and final). Additionally, it should


red in different conditions during a process of gradual adaptation to serum-

, and 3), % referent to the percentage of serum supplemented in the culture




ResearchPap

er

be mentioned that cell density and viability were improved at C3 E

0.075% and C3 E 0%, reaching levels similar to those of the initial

culture conditions (D 10%).

As mentioned, the aim of this work was to assess how the

glycosylation profile of a mAb is affected by the process of adapta-

tion of CHO-K1 cells to serum-free conditions. In Figure 1 it is

possible to observe the profile obtained at the different stages

mentioned in Table 1.

The N-glycosylation profile obtained for the mAb produced in

the normal serum-supplemented culture conditions (D 10%) was

in accordance with the known human IgG1 profile [41,64–67],

with the three main structures (in the usual order of prevalence)

present: the core fucosylated agalactosylated (peak 4, FG0), mono-

galactosylated (peaks 8 and 9, two isoforms of FG1), and digalac-

tosylated (peak 12, FG2) glycans [28,29,38,68]. Moreover, the

relation between the two FG1 isoforms (peaks 8 and 9) matches

the ratio usually displayed in endogenous human IgG1 (approxi-

mately 3:1) [69,70].

However, it is clear from Figure 1 that this profile undergoes

several modifications during the process of adaptation to the

absence of serum, as a consequence of the changes in the culture

conditions. The proportions of the main structures vary, as well as

the presence/absence of other glycan structures. Such variations of

the main and less abundant glycoforms can affect product quality

and bioactivity [26,71].

Table 2 presents more detailed data for the analysis of the

modifications observed in Figure 1, including the glucose unit

(GU) values, tentative assignments, and percentage of area of the

peaks obtained at the different stages of the process of cell adapta-

tion. The results shown confirm the variation of the proportions of

the main structures observed in Figure 1 for some samples. Indeed,

cells growing in C2 D 2.5% produce a mAb where FA2G1 is the

main glycan. Furthermore, in some cases the typically less abun-

dant structures become more prevalent than the usual three main

glycans, as is the case of A1G1 for C2 D 2.5%, C3 D 2.5%, C1 E

0.625%, and C3 E 0.625%, all at the expense of the FA2G2

structure. For C3 E 0.15%, FA3G3S2 and A2G2S2 become the main

glycans at the expense of FA2G1 and FA2G2.

It is curious to note that the reduction of the proportion of the

main glycan structures, which goes from the initial 82% down to

54%, occurs only at the middle stages of the process of adaptation.

Indeed, at the final stages of adaptation (C3 E 0.075% and C3 E

0%), the main structures reach percentages even superior (89%) to

those of the initial culture conditions. This may be related to the

behavior of the cells during adaptation. Indeed, at this final phase,

cells have reacquired proliferation rates and viabilities similar to

those observed initially, contrasting with the low cell densities and

viabilities obtained during the middle steps of adaptation. Low cell

viability may cause the accumulation of extracellular glycosidases

in the medium, and potentially alter the glycan profile of the mAb

produced [72,73]. Furthermore, at the stage of 0.31% serum, cells

shifted their mode of growth from adherent to suspended, which

may have led to modifications of the mAb glycosylation profile

thereafter. Another factor to consider is the uncommon long time

needed for cell adaptation (usually achieved in just a few months)

during which clonal selection may have occurred and thereby

caused some of the discrepancies observed in cell growth and in

the glycosylation profile of the mAb produced.


It should be noted that structures with higher GU values,

corresponding to glycans containing sialic acids, are absent in

some samples, which may impact the biological activity of the

mAb produced. Indeed, biological activity is closely related to the

glycosylation pattern, particularly to the following functional

elements: galactose (Gal), core fucose (F), sialic acid (S) and bisect-

ing N-acetylglucosamine (GlcNAc) [26,74]. Concerning the latter,

it is known that CHO cells lack the enzyme (N-acetyl-glucosami-

nyltransferase III) that mediates the transfer of bisecting GlcNAc to

complex glycans, so no particular attention will be given to this

element. By contrast, the degrees of galactosylation, fucosylation,

and sialylation of the mAb produced, and their possible impacts on

mAb biological activity, will be discussed in more detail. For this,

Table 3 shows the levels of the abovementioned main structures

composing the mAb obtained at the different stages of the process

of adaptation to serum-free conditions.

GalactosylationGalactosylation is one of the most studied glycosylation features of

any glycoprotein [8]. However, the effect of terminal galactose on

mAb effector functions is still not fully understood. Nevertheless,

it is suggested that higher levels of galactosylation are desirable

[27,28], because agalactosylated (G0) structures in antibodies are

correlated with several pathologies [75]. Antibodies with higher

galactose content could have a more effective action through

improvements of CDC [30,76].

In the present study, the levels of galactosylation of the mAb

obtained in the normal culture conditions (D 10%, 53% galactosy-

lation) are comparable to those obtained previously for normal

human IgG [28,77]. These levels are increased for serum levels

between 0.625 and 0.15% for all the combinations of supplements

assessed (C1-3). This is mainly due to the increase of monogalacto

(G1) structures, except for C3 E 0.31% and C3 E 0.15%, where

digalacto (G2) and trigalacto (G3) structures contribute majorly

to levels of galactosylation close to 80%. It has been suggested that

hypergalactosylation arises when cells are grown in stationary

culture [78] or low density static culture [79]; so the fact that at

these stages of the process cells were growing at low densities may

justify the higher levels of galactosylation observed. The antibody

produced in these conditions may be more effective with improved

CDC [15,30,76]. However, with further reduction of serum percen-

tage in the medium (C3 E 0.075% and C3 E 0%), galactosylation

decreases to levels below those observed at the initial culture con-

ditions. The lower proportions of glycans with terminal galactose

obtained in this study for cells growing in serum-free conditions in

comparison with those of serum-containing media are in accor-

dance with previous works concerning CHO [80] and hybridoma

cells [28], and may reduce the efficacy of the mAb.

FucosylationCore fucosylation of IgG has been intensively studied because of its

role in ADCC [8]. Fucose interferes with binding of the IgG to the

FcyRIIIa receptor and dampens its ability to destroy target cells

through ADCC [7,26,81,82]. Consequently, lack of core fucose

enhances the clinical efficacy of mAbs that exert their therapeutic

effect by ADCC mediated killing [24,32,33]. Indeed, it has been

shown that IgGs deficient in core fucose have ADCC activity

enhanced by up to 100-fold [24].




RESEARCH

PAPER

New

Biotech

nology�Volume

00,

Number

00�D

ecember

2012

NB

T-5

72

; N

o o

f P

ages

10

Please

cite th

is article

in p

ress as:

Costa,

A.R

. et

al., T

he

impact

of

cell ad

aptatio

n to

serum

-free co

nditio

ns

on

the

gly

cosy

lation

pro

file

of

a m

onoclo

nal

antib

ody

pro

duced

by

Chin

eseham

ster o

vary cells,

New

Bio

technol.

(2013),

http

://dx.d

oi.o

rg/1

0.1

016/j.n

bt.2

012.1

2.0

02

TABLE 2

Data of mean glucose unit (GU) value, tentative assignment, and percentage of area of the peaks detected on the monoclonal antibody produced by CHO-K1 cells duringadaptation to serum-free conditions. D – DMEM, E – EXCELL, GU – glucose units, Cx – combination of supplements x (x = 1, 2, 3), % referent to the percentage of serumsupplemented in the culture media. The structures of higher percentage are shown in bold, with those that are not the typical main IgG structures underlined.

Peak Mean GUa Tentativeassignment

% Area

D10%

C1D 2.5%

C1E 0.625%

C2D 2.5%

C2E 0.625%

C3D 2.5%

C3D 0.625%

C3E 0.625%

C3E 0.31% i

C3E 0.31% f

C3E 0.15%

C3E 0.075%

C3E 0%

1 4.91 � 0.02 A1 2.09 0.84 8.08 4.67 5.42 8.12 2.60 9.89 5.67 0.53 2.31 3.60 2.97

2 5.34 � 0.01 FA1/A2 2.01 1.92 – – 3.82 3.62 5.63 2.37 3.55 1.11 2.51 2.28 2.61

3 5.68 � 0.07 A1G1 – – 25.72 20.32 9.27 11.06 – 14.85 – – 3.38 – –4 5.82 � 0.02 FA2 40.99 41.51 15.34 18.14 19.24 29.31 37.53 19.36 22.61 25.12 14.19 37.69 50.215 6.18 � 0.01 M5 1.50 1.98 – – 3.42 4.40 4.80 1.90 3.76 1.84 1.83 1.79 1.62

6 6.30 � 0.01 A2G1[6] – – – – – – 2.02 – 1.88 0.80 1.90 – –

7 6.43 � 0.00 A2G1[3] – – – – 5.43 5.52 – 7.47 – – – – –8 6.64 � 0.02 FA2G1[6] 23.52 24.21 17.86 14.96 15.31 18.50 23.42 13.75 20.31 21.67 10.10 26.27 23.279 6.75 � 0.03 FA2G1[3] 9.11 6.41 23.89 27.84 8.66 6.21 10.38 6.29 9.20 9.31 8.42 8.30 7.89

10 7.12 M6 – – – 1.03 – – –– – – – –

11 7.16 � 0.03 A2G2/M6 1.65 1.50 2.08 1.03 3.77 3.60 2.17 4.43 2.51 1.23 1.34 2.44 1.9412 7.61 � 0.02 FA2G2 8.40 8.55 7.03 11.15 10.45 7.26 9.24 8.04 12.79 9.27 8.50 10.00 7.3413 7.98 � 0.01 FA2G1S1 2.73 3.25 – 0.85 2.91 2.39 1.52 3.47 2.33 2.07 1.49 3.32 2.16

14 8.46 � 0.02 A2G2S1 0.91 1.62 – – 1.33 – 0.67 0.82 1.72 2.76 3.19 1.20 –15 8.75 � 0.00 FA2G2S(3)1 – – – – 1.28 – – 1.78 – – – – –

16 8.89 � 0.01 FA2G2S(6)1 0.98 1.23 – – 0.95 – – 0.77 1.70 2.37 2.64 2.03 –

17 9.59 � 0.09 A2G2S2 1.99 1.13 – – 1.46 – – 1.56 1.76 2.71 3.83 1.08 –

18 9.99 � 0.17 FA2G2S2 0.95 1.64 – – 2.44 – – 1.24 3.86 6.81 11.94 – –19 10.19 � 0.10 A3G3S2 – 1.12 – – – – – – 0.69 1.13 2.25 – –

20 10.45 � 0.02 FA3G3S2[6] 1.14 1.00 – – 3.63 – –– 2.00 4.22 7.58 14.33 – –

21 10.67 � 0.01 FA3G3S2[3,6] 0.92 0.86 – – – – – – – – – – –

22 10.97 � 0.05 A3G3S3 0.84 0.89 – – 1.21 – – – 1.45 3.68 5.85 – –23 11.41 � 0.02 FA3G3S3 0.27 0.34 – – – – – – – – – – –

Main structures 82.02 80.68 64.12 72.09 53.66 61.28 80.57 47.44 64.91 65.37 41.21 82.26 88.71

Other structures 17.98 19.32 35.88 27.90 46.34 38.71 19.41 52.55 35.10 34.62 58.79 17.74 11.30a Average of the GU value of all samples containing the peak.

6

www.else

vier.co

m/lo

cate/nbt

ResearchPaper


New

Biotech

nology�Volume

00,

Number

00�D

ecember

2012

RESEARCH

PAPER

NB

T-5

72

; N

o o

f P

ages

10

Please

cite th

is article

in p

ress as:

Costa,

A.R

. et

al., T

he

impact

of

cell ad

aptatio

n to

serum

-free co

nditio

ns

on

the

gly

cosy

lation

pro

file

of

a m

onoclo

nal

antib

ody

pro

duced

by

Chin

eseham

ster o

vary cells,

New

Bio

technol.

(2013),

http

://dx.d

oi.o

rg/1

0.1

016/j.n

bt.2

012.1

2.0

02

TABLE 3

Galactose, core fucose and sialic acid composition of the monoclonal antibody produced by CHO-K1 cells during the different steps of the process of adaptation to serum-freeconditions. D – DMEM, E – EXCELL, Cx – combination of supplements x (x = 1, 2, 3), % referent to the percentage of serum supplemented in the culture media.

Glycan structure % Area

D10%

C1D 2.5%

C1E 0.625%

C2D 2.5%

C2E 0.625%

C3D 2.5%

C3D 0.625%

C3E 0.625%

C3E 0.31% i

C3E 0.31% f

C3E 0.15%

C3E 0.075%

C3E 0%

Agalactosylated (G0) 47.03 47.00 24.46 24.36 33.79 47.25 51.65 35.74 36.85 29.22 21.51 46.58 58.38

Galactosylated(G1+G2+G3)

52.99 53.00 75.54 75.64 66.22 52.74 48.34 64.26 63.17 70.78 78.49 53.42 41.63

Monogalactosylated (G1) 33.85 33.87 67.47 63.97 41.58 43.68 37.34 45.83 33.72 33.85 25.29 37.89 33.32

Digalactosylated (G2) 15.33 14.92 8.07 11.67 19.80 9.06 11.00 16.43 23.09 24.54 30.77 15.53 8.31Trigalactosylated (G3) 3.81 4.21 0.00 0.00 4.84 0.00 0.00 2.00 6.36 12.39 22.43 0.00 0.00

Core fucosylated (F) 86.85 89.00 64.12 72.94 64.87 63.67 82.09 56.70 77.02 84.20 71.61 87.61 90.87

Core fucosylatedagalactosylated (FG0)

40.59 41.51 15.34 18.14 19.24 29.31 37.53 19.36 22.61 25.12 14.19 37.69 50.21

Core fucosylatedmonogalactosylated(FG1)

33.85 33.87 41.75 43.65 26.88 27.10 35.32 23.51 31.84 33.05 20.01 37.89 33.32

Core fucosylateddigalactosylated (FG2)

9.65 11.42 7.03 11.15 15.12 7.26 9.24 11.83 18.35 18.45 23.08 12.03 7.34

Core fucosylatedtrigalactosylated (FG3)

2.76 2.20 0.00 0.00 3.63 0.00 0.00 2.00 4.22 7.58 14.33 0.00 0.00

Sialylated (S1+S2+S3) 13.18 13.08 0.00 0.85 15.21 2.39 2.19 11.64 17.73 29.11 45.52 7.63 2.16Monosialylated (S1) 5.57 6.10 0.00 0.85 6.47 2.39 2.19 6.84 5.75 7.20 7.32 6.55 2.16

Disialylated (S2) 5.99 5.75 0.00 0.00 7.53 0.00 0.00 4.80 10.53 18.23 32.35 1.08 0.00

Trisialylated (S3) 1.62 1.23 0.00 0.00 1.21 0.00 0.00 0.00 1.45 3.68 5.85 0.00 0.00

Monosialylatedmonogalactosylated(S1G1)

3.01 3.25 0.00 0.85 2.91 2.39 1.52 3.47 2.33 2.07 1.49 3.32 2.16

Monosialylateddigalactosylated(S1G2)

2.56 2.85 0.00 0.00 3.56 0.00 0.67 3.37 3.42 5.13 5.83 3.23 0.00

Disialylateddigalactosylated(S2G2)

3.80 2.77 0.00 0.00 3.90 0.00 0.00 2.80 5.62 9.52 15.77 1.08 0.00

Disialylatedtrigalactosylated(S2G3)

2.19 2.98 0.00 0.00 3.63 0.00 0.00 2.00 4.91 8.71 16.58 0.00 0.00

Complex glycans 97.68 97.49 97.27 98.96 98.45 94.70 93.79 94.10 95.88 95.00 97.54 97.50 96.99High mannose

structures2.32 2.52 2.73 1.04 1.55 5.31 6.21 5.91 4.13 5.01 2.46 2.50 3.01

www.else

vier.co

m/lo

cate/nbt

7

Research Paper




Research

Pap

er

In the present study, the lowest and therefore most favorable

levels of core fucose were generally obtained in the middle stages

of the process of adaptation. However, these positive changes on

mAb structure are lost at the final stages of adaptation, with core

fucose levels at C3 E 0% of approximately 90%, which are similar

to those of D 10%, and in accordance with the normal levels

observed in human and CHO cell derived antibodies [20].

In IgGs, a ratio of fucosylated structures of 1:2:1 for agalacto:-

monogalacto:digalacto (FG0/FG1/FG2) structures has been

observed [83,84]. Interestingly, this ratio is not observed in the

mAb produced in the initial culture conditions, with FG0 as the

predominant fucosylated structure. This initial ratio changes

through the process of adaptation, with values closer to those

expected in C3 E 0.625% and C3 E 0.31%, FG2 predominant at C3

E 0.15%, and the initial FG0 predominance returned at C3 E

0.075% and C3 E 0%.

SialylationThe effects of sialylation of IgG have recently attracted much

attention, because it was shown that the addition of sialic acids

dramatically changes the physiological role of IgGs by converting

them from pro-inflammatory into anti-inflammatory agents [35–

37]. Reports have shown that higher sialylation of IgG associates

with low ADCC, due to reduced binding to specific Fc receptors

such as FcgRIII [34].

In general, the degree of sialylation reported for IgG has been

below 20% [8,28,85]. These values are in accordance with those

obtained in the present study, except for C3 E 0.31% and C3 E

0.15%, which show higher sialic acid contents (29.11 and 45.52%,

respectively). These increased levels are mainly related to a higher

prevalence of disialylated structures, both digalacto (S2G2) and

trigalacto (S2G3). By contrast, sialylated structures are less present

at C1 E 0.065%, C2 D 2.5%, and at the lowest serum stages of the

process of adaptation, C3 E 0.075% and C3 E 0%, with monosialic

acids as the most prevalent sialylated structure. These results of

lower mAb sialylation in serum-free conditions when compared to

serum-containing media are in accordance to previous studies

with hibridoma cells [28], and suggest an improvement of mAb

efficacy through ADCC.

Other elementsIn addition to the functional elements already discussed (galactose,

fucose and sialic acid), other less abundant glycoforms may also

have an important effect on the biological activity of therapeutic

IgGs. The presence of high mannose structures, for example, has

been associated with reduced ADCC and CDC [26], as well as with a

rapid IgG clearance from serum [27]. In this work, mAb produced at

all stages of the process of adaptation had low levels of high

mannose structures, with 94 to 99% of the glycoforms pertaining

to complex glycans. The high mannose structures were slightly

more present at the middle steps of adaptation (with combination

C3), returning to the initial lower levels at the end of the process.

The differences found for all the functional elements between

the intermediate and final stages of the process of adaptation may

be related to the composition of the medium. At these stages cells

were growing in EXCELL, whose composition is much more

complex than that of DMEM to compensate for the absence of

serum and to promote mAb production. Because this medium is



not formulated to be supplemented with serum, it is possible that

its combination with serum at the percentages used in the inter-

mediate stages of adaptation have influenced cell growth and mAb

glycosylation differently than at the final stages of adaptation

(absence of serum). Indeed, the interaction between EXCELL

and serum may have led to antagonist and/or synergistic effects

of their compounds. Furthermore, the proteolytic activities of

extracellular enzymes (e.g. sialydase, galactosydase) have been

shown to vary in the presence or absence of serum [86–88], and

may have also caused modifications in the glycan composition

of the mAb produced at the different stages of the process of

adaptation.

Although the presence of IgG from fetal bovine serum in the

samples could have influenced the glycosylation profiles obtained,

in this work the methodology used for IgG purification had a high

specificity for human IgG while providing weak purification for

bovine IgG, reducing the likelihood of such contamination.

Furthermore, other studies in similar conditions have shown a

low contribution of bovine IgG to the peaks of the glycosylation

profile (between 0.1 and 2.0% for most peaks, with 4.7% for the

A2G2F peak) [28]. Therefore, the conclusions of the present study

are most probably not affected by the presence of bovine IgG.

ConclusionsThe establishment of a process for the large-scale production of

therapeutics comprises several steps that aim to maximize product

yield. However, it is essential to understand and monitor the effect

of these processes on product quality (e.g. glycosylation) to assure

that its therapeutic action is not hindered. In this study, it was

clear that one of the most time-consuming steps of process devel-

opment – the adaptation of producer cells to growth in serum-free

conditions – has a strong impact on product glycosylation/quality.

The initial profile (DMEM, 10% serum) was in accordance with the

typical IgG1 profile, but the prevalence of the common main

structures (core fucosylated agalacto, monogalacto and digalacto

glycans) decreased from 82 to 54% during the middle stages of the

process, and increased to 89% in the final steps. Indeed, it was

interesting to note the different behaviors of mAb glycosylation

for the middle and final steps of adaptation. At the middle stages,

galactosylation was increased and fucosylation reduced, both

having a positive impact on the clinical efficacy of the mAb

through improvements of the ADCC and CDC activity. Conver-

sely, sialylation was increased and exceeded the levels usually

reported (below 20%), particularly for combination C3 with

EXCELL at 0.15% serum (45.52% sialic acid content), which results

in reduced ADCC activity and, consequently, low mAb efficiency.

At the final steps of adaptation, in serum-free medium, the levels of

fucosylation were similar to those of the initial serum-supplemen-

ted culture conditions, but galactosylation and sialylation were

decreased, with divergent potential effects (negative and positive,

respectively) on the biological activity of the mAb. The diver-

gences found may be associated with a low cell density and

viability observed at the middle steps, as well as to a shift of the

cell growth mode from adherent to suspension.

The final result of the interplay of these different mAb char-

acteristics is not predictable and would need to be assessed in vivo.

However, it is clear that the process of cell adaptation to serum-free

conditions impacts the quality of the product, suggesting that





product quality control and monitoring should be implemented

even at these initial stages of process development.

AcknowledgementsThe authors acknowledge funding and support from the

Portuguese Foundation for Science and Technology (FCT),


namely grant reference SFRH/BD/46660/2008 for Ana Rita

Costa and SFRH/BD/46661/2008 for Maria Elisa Rodrigues.

Also, a special acknowledgement to Biotecnol S.A. (Lisbon,

Portugal) for providing the monoclonal antibody used in this

study.

References

ResearchPap

er

[1] Primack J, Flynn GC, Pan H. A high-throughput microchip-based glycan screen-ing assay for antibody cell culture samples. Electrophoresis 2011;32:1129–32.

[2] Reichert JM, Rosensweig CJ, Faden LB, Dewitz MC. Monoclonal antibodysuccesses in the clinic. Nature Biotechnology 2005;23:1073–8.

[3] Carter PJ. Potent antibody therapeutics by design. Nature Reviews Immunology2006;6:343–57.

[4] Reichert JM, Valge-Archer VE. Development trends for monoclonal antibodycancer therapeutics. Nature Reviews Drug Discovery 2007;6:349–56.

[5] Rodrigues ME, Costa AR, Henriques M, Azeredo J, Oliveira R. Technologicalprogresses in monoclonal antibody production systems. Biotechnology Progress2009;26:332–51.

[6] LeFloch F, Tessier B, Chenuet S, Guillaume J-M, Cans P, Goergen J-L, et al.Related effects of cell adaptation to serum-free conditions on murine EPOproduction and glycosylation by CHO cells. Cytotechnology 2006;52:39–53.

[7] Hossler P, Khattak SF, Li ZJ. Optimal and consistent protein glycosylation inmammalian cell culture. Glycobiology 2009;19:936–49.

[8] Pucic M, Knezevic A, Vidic J, Adamczyk B, Novokmet M, Polasek O, et al. Highthroughput isolation and glycosylation analysis of IgG – variability and herit-ability of the IgG glycome in three isolated human populations. Molecular andCellular Proteomics 2011;10.

[9] Sutton B, Phillips D. The three-dimensional structure of the carbohydrate withinthe Fc fragment of immunoglobulin G. Biochemical Society Transactions1983;11:130–2.

[10] Clark M. IgG effector mechanisms. Chemical Immunology 1997;65:88–110.[11] Fernandes D. Demonstrating comparability of antibody glycosylation during

biomanufacturing. European Biopharmaceutical Reviews 2005;106–10.[12] Lowe JB, Marth JD. A genetic approach to mammalian glycan function. Annual

Review of Biochemistry 2003;72:643–91.[13] Lund J, Takahashi N, Popplewell A, Goodall M, Pound JD, Tyler R, et al.

Expression and characterization of truncated forms of humanized L243 IgG1.European Journal of Biochemistry 2000;267:7246–57.

[14] Mimura Y, Church S, Ghirlando R, Ashton PR, Dong S, Goodall M, et al. Theinfluence of glycosylation on the thermal stability and effector function expres-sion of human IgG1-Fc: properties of a series of truncated glycoforms. MolecularImmunology 2000;37:697–706.

[15] Davies J, Jiang LY, Pan L-Z, LaBarre MU, Andersen D, Reff M. Expression ofGnTIII in recombinant anti-CD20 CHO production cell line: expression ofantibodies with altered glycoforms leads to an increase in ADCC through higheraffinity for FcyRIII. Biotechnology and Bioengineering 2001;74:288–94.

[16] Krapp S, Mimura Y, Jefferis R, Huber R, Sondermann P. Structural analysis ofhuman IgG-Fc glycoforms reveals a correlation between glycosylation andstructural integrity. Journal of Molecular Biology 2003;325:979–89.

[17] Yamane-Ohnuki N, Kinoshita S, Miho I-U, Kusunoki M, Iida S, Nakano R, et al.Establishment of FUT8 knockout Chinese hamster ovary cells: an ideal host cellline for producing completely defucosylated antibodies with enhanced anti-body-dependent cellular cytotoxicity. Biotechnology and Bioengineering2004;87:614–22.

[18] Magdelaine-Beuzelin C, Kaas Q, Wehbi V, Ohresser M, Jefferis R, Lefranc M-P,et al. Structure–function relationships of the variable domains of monoclonalantibodies approved for cancer treatment. Critical Review in Oncology/Hema-tology 2007;64:210–25.

[19] Jefferis R. Glycosylation of recombinant antibody therapeutics. BiotechnologyProgress 2005;21:11–6.

[20] Raju TS. Glycosylation variations with expression systems and their impact onbiological activity of therapeutic immunoglobulins. Bioprocess International2003;1:44–53.

[21] Elliott S, Lorenzini T, Asher S, Aoki K, Brankow D, Buck L, et al. Enhancement oftherapeutic protein in vivo activities through glycoengineering. Nature Bio-technology 2003;21:414–21.

[22] Hashimoto K, Goto S, Kawano S, Aoki-Kinoshita KF, Ueda N, Hamajima M, et al.KEGG as a glycome informatics resource. Glycobiology 2006;16:63R–70R.

[23] Tsuchiya N, Endo T, Matsuta K, Yoshinoya S, Aikawa T, Kosuge E, et al. Effects ofgalactose depletion from oligosaccharide chains on immunological activities ofhuman IgG. Journal of Rheumatology 1989;16:285–90.

[24] Shields RL, Lai J, Keck R, O’Connell LY, Hong K, Meng YG, et al. Lack of fucoseon human IgG1 N-linked oligosaccharide improves binding to human FcgRIIIand antibody-dependent cellular toxicity. Journal of Biological Chemistry2002;277:26733–40.

[25] Goetze AM, Liu YD, Arroll T, Chu L, Flynn GC. Rates and impact of humanantibody glycation in vivo. Glycobiology 2012;22:221–34.

[26] Kanda Y, Yamada T, Mori K, Okazaki A, Inoue M, Kitajima-Miyama K, et al.Comparison of biological activity among nonfucosylated therapeutic IgG1antibodies with three different N-linked Fc oligosaccharides: the high-mannose,hybrid, and complex types. Glycobiology 2007;17:104–18.

[27] Abes R, Teillaud J-L. Impact of glycosylation on effector functions of therapeuticIgG. Pharmaceuticals 2010;3:146–57.

[28] Serrato JA, Hernandez V, Estrada-Mondaca S, Palomares LA, Ramırez OT.Differences in the glycosylation profile of a monoclonal antibody producedby hybridomas cultured in serum-supplemented, serum-free or chemicallydefined media. Biotechnology and Applied Biochemistry 2007;47:113–24.

[29] Jefferis R, Lund J. Glycosylation of antibody molecules: structural and func-tional significance. Chemical Immunology 1997;65.

[30] Hodoniczky J, Zheng YZ, James DC. Control of recombinant monoclonalantibody effector functions by Fc N-glycan remodeling in vitro. BiotechnologyProgress 2005;21:1644–52.

[31] Ferrara C, Brunker P, Suter T, Moser S, Puntener U, Umana P. Modulation oftherapeutic antibody effector functions by glycosylation engineering: influenceof Golgi enzyme localization domain and co-expression of heterologous beta1,4-N-acetylglucosaminyltransferase III and Golgi alpha-mannosidase II. Biotech-nology and Bioengineering 2006;93:851–61.

[32] Siberil S, de Romeuf C, Bihoreau N, Fernandez N, Meterreau J-L, Regenman A,et al. Selection of a human anti-RhD monoclonal antibody for therapeutic use:impact of IgG glycosylation on activating and inhibitory FcgR functions.Clinical Immunology 2006;118:170–9.

[33] Shinkawa T, Nakamura K, Yamane N, Shoji-Hosaka E, Kanda Y, Sakurada M,et al. The absence of fucose but not the presence of galactose or bisecting N-acetylglucosamine of human IgG1 complex-type oligosaccharides shows thecritical role of enhancing antibody-dependent cellular cytotoxicity. Journal ofBiological Chemistry 2003;278:3466–73.

[34] Scallon BJ, Tam SH, McCarthy SG, Cai AN, Raju TS. Higher levels of sialylated Fcglycans in immunoglobulin G molecules can adversely impact functionality.Molecular Immunology 2007;44:1524–34.

[35] Kaneko Y, Nimmerjahn F, Ravetch JV. Anti-inflammatory activity of immuno-globulin G resulting from Fc sialylation. Science 2006;313:670–3.

[36] Anthony R, Ravetch J. A novel role for the IgG Fc glycan: the anti-inflammatoryactivity of sialylated IgG Fcs. Journal of Clinical Immunology 2010;30:9–14.

[37] Anthony RM, Nimmerjahn F, Ashline DJ, Reinhold VN, Paulson JC, Ravetch JV.Recapitulation of IVIG anti-inflammatory activity with a recombinant IgG Fc.Science 2008;320:373–6.

[38] Jenkins N, Parekh RB, James DC. Getting the glycosylation right: implicationsfor the biotechnology industry. Nature Biotechnology 1996;14:975–81.

[39] Sethuraman N, Stadheim TA. Challenges in therapeutic glycoprotein produc-tion. Current Opinion in Biotechnology 2006;17:341–6.

[40] Chotigeat W, Watanapokasin Y, Mahler S, Gray PP. Role of environmentalconditions on the expression levels, glycoform pattern and levels of sialyltrans-ferase for hFSH produced by recombinant CHO cells. Cytotechnology1994;15:217–21.

[41] Serrato JA, Palomares LA, Meneses-Acosta A, Ramırez OT. Heterogeneous con-ditions in dissolved oxygen affect N-glycosylation but not productivity of amonoclonal antibody in hybridoma cultures. Biotechnology and Bioengineer-ing 2004;88:176–88.

[42] Muthing J, Kemminer SE, Conradt HS, Sagi D, Nimtz M, Karst U, et al. Effects ofbuffering conditions and culture pH on production rates and glycosylation ofclinical phase I anti-melanoma mouse IgG3 monoclonal antibody R24. Bio-technology and Bioengineering 2003;83:321–34.

[43] Yoon SK, Choi SL, Song JY, Lee GM. Effect of culture pH on erythropoietinproduction by Chinese hamster ovary cells grown in suspension at 32.5 and37.08C. Biotechnology and Bioengineering 2005;89:345–56.

[44] Yoon SK, Kim SH, Lee GM. Effect of low culture temperature on specificproductivity and transcription level of anti-4-1BB antibody in recombinantChinese hamster ovary cells. Biotechnology Progress 2003;19:1383–6.

[45] Trummer E, Fauland K, Seidinger S, Schriebl K, Lattenmayer C, Kunert R, et al.Biphasic cultivation-A tool for enhancing the volumetric productivity of batchprocesses using Epo-Fc expressing CHO cells. Biotechnology and Bioengineering2006;94:1045–52.

[46] Gu X, Xie L, Harmon BJ, Wang DIC. Influence of Primatone RL supplementationon sialylation of recombinant human interferon-g produced by Chinese ham-ster ovary cell culture using serum-free media. Biotechnology and Bioengineer-ing 1997;56:353–60.

[47] Patel T, Parekh R, Moellering B, Prior C. Different culture methods lead todifferences in glycosylation of a murine IgG monoclonal antibody. BiochemicalJournal 1992;285:839–45.

[48] Wong CFD, Wong TKK, Goh TL, Heng KC, Yap GSM. Impact of dynamic onlinefed-batch strategies on metabolism, productivity and N-glycosylation quality inCHO cell cultures. Biotechnology and Bioengineering 2005;89:164–77.

[49] Kunkel JP, Jan DC, Butler M, Jamieson JC. Comparisons of the glycosylation of amonoclonal antibody produced under nominally identical cell culture condi-tions in two different bioreactors. Biotechnology Progress 2000;16:462–70.






Research

Pap

er

[50] Andersen DC, Goochee CF. The effect of cell-culture conditions on the oligo-saccharide structures of secreted glycoproteins. Current Opinion in Biotechnol-ogy 1994;5:546–9.

[51] Ozturk SS, Hu W-S. Cell culture technology for pharmaceutical and cell-basedtherapies. New York: CRC Press; 2006.

[52] Geigert J. The challenge of CMC regulatory compliance for biopharmaceuticals.New York: Springer; 2004.

[53] Doyle A, Griffiths JB. Cell and tissue culture: laboratory procedures in biotech-nology. Chichester, England: John Wiley & Sons; 1998.

[54] Crowell CK, Grampp GE, Rogers GN, Miller J, Scheinman RI. Amino acid andmanganese supplementation modulates the glycosylation state of erythropoietinin a CHO culture system. Biotechnology and Bioengineering 2007;96:538–49.

[55] Ozturk S, Kasebo G, Mahaworasilpa T, Coster HG. Adaptation of cell lines toserum-free culture medium. Hybridoma and Hybridomics 2003;22:267–72.

[56] Griffiths JB. Serum and growth factors in cell culture media – an introductoryreview. Developments in Biological Standardization 1987;66:155–60.

[57] Costa AR, Rodrigues ME, Henriques M, Melton D, Cunnah P, Oliveira R, et al.Evaluation of the OSCARTM system for the production of monoclonal antibodiesby CHO-K1 cells. International Journal of Pharmaceutics 2012;430:42–6.

[58] Jacobia S, Kenerson R, Tescione L, Gruber D, Jayme D, Munroe D, et al. Traceelement optimization enhances performance and reproducibility of serum-freemedium. In: Iijima S, Nishijima K-I, editors. Animal cell technology: basic &applied aspects. Netherlands: Springer; 2006. p. 193–9.

[59] Kenerson RW. Cell culture medium comprising transition metals or traceelements. In: Office UPT, editor. United States Patent Application Publication.US: Invitrogen Corporation; 2005.

[60] Conrad M, Umbreit J, Moore E. Iron absorption and transport. American Journalof the Medical Sciences 1999;318:213–29.

[61] Elliott RL, Elliott MC, Wang FEN, Head JF. Breast carcinoma and the role of ironmetabolism. Annals of the New York Academy of Sciences 1993;698:159–66.

[62] Kovar J, Franek F. Growth-stimulating effect of transferrin on a hybridoma cellline: relation to transferrin iron-transporting function. Experimental CellResearch 1989;182:358–69.

[63] Garcıa-Alfonso C, Lopez-Barea J, Sanz P, Repetto G, Repetto M. Changes inantioxidative activities induced by Fe(II) and Fe(III) in cultured Vero cells.Archives of Environment Contamination and Toxicology 1996;30:431–6.

[64] Takahashi N, Tomiya N. Analysis of N-linked oligosaccharides: application ofglycoamidase A. In: Takahashi N, Muramatsu T, editors. Handbook of endo-glycosidases and glycoamidases. Boca Raton, FL: CRC Press; 1992. p. 199–332.

[65] Masuda K, Yamaguchi Y, Kato K, Takahashi N, Shimada I, Arata Y. Pairing ofoligosaccharides in the Fc region of immunoglobulin G. FEBS Letters 2000;473:349–57.

[66] Raju TS, Briggs JB, Borge SM, Jones AJS. Species-specific variation in glycosyla-tion of IgG: evidence for the species-specific sialylation and branch-specificgalactosylation and importance for engineering recombinant glycoproteintherapeutics. Glycobiology 2000;10:477–86.

[67] Saba JA, Kunkel JP, Jan DCH, Ens WE, Standing KG, Butler M, et al. A study ofimmunoglobulin G glycosylation in monoclonal and polyclonal species byelectrospray and matrix-assisted laser desorption/ionization mass spectrometry.Analytical Biochemistry 2002;305:16–31.

[68] Routier F, Davies M, Bergemann K, Hounsell E. The glycosylation pattern of ahumanized IgGl antibody (D1.3) expressed in CHO cells. GlycoconjugateJournal 1997;14:201–7.

[69] Flynn GC, Chen X, Liu YD, Shah B, Zhang Z. Naturally occurring glycan forms ofhuman immunoglobulins G1 and G2. Molecular Immunology 2010;47:2074–82.

[70] Jefferis R, Lund J, Mizutani H, Nakagawa H, Kawazoe Y, Arata Y, et al. Acomparative study of the N-linked oligosaccharide structures of human IgGsubclass proteins. Biochemical Journal 1990;268:529–37.



[71] Mimura Y, Ghirlando R, Sondermann P, Lund J, Jefferis R. The molecularspecificity of IgG-Fc interactions with Fc gamma receptors. Advances in Experi-mental Medicine and Biology 2001;495:49–53.

[72] Gramer MJ, Goochee CF. Glycosidase activities in Chinese hamster ovary celllysate and cell culture supernatant. Biotechnology Progress 1993;9:366–73.

[73] Gramer MJ, Goochee CF. Glycosidase activities of the 293 and NS0 cell lines, andof an antibody-producing hybridoma cell line. Biotechnology and Bioengineer-ing 1994;43:423–8.

[74] Cabrera G, Cremata JA, Valdes R, Garcıa R, Gonzalez Y, Montesino R, et al.Influence of culture conditions on the N-glycosylation of a monoclonal anti-body specific for recombinant hepatitis B surface antigen. Biotechnology andApplied Biochemistry 2005;41:67–76.

[75] Parekh RB, Dwek RA, Sutton BJ, Fernandes DL, Leung A, Stanworth D, et al.Association of rheumatoid arthritis and primary osteoarthritis with changes inthe glycosylation pattern of total serum IgG. Nature 1985;316:452–7.

[76] Boyd P, Lines A, Patel A. The effect of the removal of sialic acid, galactose andtotal carbohydrate on the functional activity of Campath-1H. Molecular Immu-nology 1995;32:1311–8.

[77] Routier FH, Hounsell EF, Rudd PM, Takahashi N, Bond A, Hay FC, et al.Quantitation of the oligosaccharides of human serum IgG from patients withrheumatoid arthritis: a critical evaluation of different methods. Journal ofImmunological Methods 1998;213:113–30.

[78] Lund J, Takahashi N, Nakagawa H, Goodall M, Bentley T, Hindley S, et al.Control of IgG/Fc glycosylation: a comparison of oligosaccharides from chi-meric human/mouse and mouse subclass immunoglobulin Gs. MolecularImmunology 1993;30:741–8.

[79] Kumpel B, Rademacher T, Rook G, Williams P, Wilson I. Galactosylation ofhuman IgG monoclonal anti-D produced by EBV-transformed B-lymphoblas-toid cell lines is dependent on culture method and affects Fc receptor-mediatedfunctional activity. Human Antibodies and Hybridomas 1994;5:143–51.

[80] Lifely MR, Hale C, Boyce S, Keen MJ, Phillips J. Glycosylation and biologicalactivity of CAMPATH-1H expressed in different cell lines and grown underdifferent culture conditions. Glycobiology 1995;5:813–22.

[81] Nimmerjahn F, Ravetch JV. Fc[gamma] receptors as regulators of immuneresponses. Nature Reviews Immunology 2008;8:34–47.

[82] Kanda Y, Yamane-Ohnuki N, Sakai N, Yamano K, Nakano R, Inoue M, et al.Comparison of cell lines for stable production of fucose-negative antibodieswith enhanced ADCC. Biotechnology and Bioengineering 2006;94:680–8.

[83] Jones D, Kroos N, Anema R, Van Montfort B, Vooys A, van der Kraats S, et al.High-level expression of recombinant IgG in the human cell line PER.C6.Biotechnology Progress 2003;19:163–8.

[84] Beck A, Bussat M-C, Zorn N, Robillard V, Klinguer-Hamour C, Chenu S, et al.Characterization by liquid chromatography combined with mass spectrometryof monoclonal anti-IGF-1 receptor antibodies produced in CHO and NS0 cells.Journal of Chromatography B 2005;819:203–18.

[85] Arnold JN, Wormald MR, Sim RB, Rudd PM, Dwek RA. The impact of glycosyla-tion on the biological function and structure of human immunoglobulins.Annual Review of Immunology 2007;25:21–50.

[86] Castro P, Ison A, Hayter P, Bull A. The macroheterogeneity of recombinanthuman interferon-gamma produced by Chinese-hamster ovary cells is affectedby the protein and lipid content of the culture medium. Biotechnology andApplied Biochemistry 1995;21:87–100.

[87] Gawlitzek M, Valley U, Nimtz M, Wagner R, Conradt HS. Characterizationof changes in the glycosylation pattern of recombinant proteins from BHK-21cells due to different culture conditions. Journal of Biotechnology 1995;42:117–31.

[88] Yang M, Butler M. Effects of ammonia and glucosamine on the heterogeneity oferythropoietin glycoforms. Biotechnology Progress 2002;18:129–38.



the impact of cell adaptation to serum-free conditions on the glycosylation profile of a monoclonal...

Documents