ana rita martins costa - core.ac.uk · trago uma caixinha mágica, à qual recorro com frequência,...

Ana Rita Martins Costa

October 2012UM

inho

|201

2

Expression and characterization ofa therapeutic monoclonal antibodyin mammalian cells

Universidade do Minho

Escola de Engenharia

Ana

Rita

Mar

tins

Cos

taE

xpre

ssio

n a

nd

ch

ara

cte

riza

tio

n o

f a

th

era

pe

uti

c

mo

no

clo

na

l a

nti

bo

dy

in m

am

ma

lia

n c

ells

Doctoral dissertation for PhD degree inBiomedical Engineering

Supervisor:Professor Joana AzeredoCo-supervisor:Professor Mariana Henriques

Ana Rita Martins Costa

October 2012

Expression and characterization ofa therapeutic monoclonal antibodyin mammalian cells

Universidade do Minho

Escola de Engenharia

Para os meus Pais

“Success is not final, failure is not fatal: it is the courage to continue that counts”

O sucesso não é final, o fracasso não é fatal: é a coragem para continuar que conta.

Winston Churchill

v

ACKNOWLEDGMENTS

Cheguei ao meu destino. A viagem foi longa e dura, mas valeu a pena. Olho para trás e traço

os caminhos percorridos, revejo os dias intensos, os desafios e as conquistas, as desilusões e as

alegrias. E vejo que não vim sozinha. Trago comigo companhia, e uma mala cheia de tesouros.

Trago um livro que consulto, quando estou numa encruzilhada, tem nas folhas palavras amigas,

tem ideias impensadas. Inclui mapas do caminho, incentiva à aventura, uma edição especial,

que faz crescer a cada leitura. Professora Joana

Trago um pequeno raio de sol, de luz intensa e generosa, que faz nascer da escuridão a

primavera harmoniosa. Tem o brilho da coragem, tem a cor da alegria, traz escondidos na sua

luz, uns pozinhos de magia. Professora Mariana

Trago um conjunto precioso e raro, de sabedoria e bondade, que mantenho bem guardado, num

cantinho recatado, para sempre admirar. Professora Rosário

Trago uma caixinha mágica, à qual recorro com frequência, transforma incertezas e

preocupações, em sorrisos, gargalhadas, histórias partilhadas e confidências.

Susana M., Paula, Dina, Sílvia P., Cláudia B., Melyssa, Sónia, Eva, Cristina, Isabel C.,

Isabel P., Joana C., Yun Lei, Pedro, Filipa, Patrícia e todos os colegas do LCCT

I bring chocolates of irish cream, with many flavors mixed within, a little wit, a touch of warmth,

joyfulness and even charm, but of them all my favorite is, the unexpected kindness they release.

Eoin, Li, Niaobh, Karina, Barbara, Jo, Tharmala, Jayne, Radka, Mark and Pauline. Larry and the

nice irish lady who once offered me pancakes on a bus stop, just because I was having a bad day

ACKNOWLEDGMENTS│

vi

Trago as cores da amizade, todas elas tão diferentes, pintam a tela da minha vida, com o pincel

da sua presença. Amélia, Elisa A., Roberto, Mariana, David, Inês, Cristina

Trago uma ave delicada, que ainda não sabe (receia) voar, tem forças escondidas, tem vontade

de cantar. Para já canta baixinho, alguns dos cantos sei de cor, partilhamos confidências,

trocamos o nosso melhor. A minha ave favorita, um dia vai arriscar, vai atirar-se ao vento, tenho

a certeza que irá voar. Elisa

Trago um postal antigo, que me acorda a memória, que me lembra a saudade, que me conta a

história, de olhos verdes que nunca julgaram, de gentileza e carinho, de tantos saberes que

resultaram da sua presença no meu caminho. Padrinho

Mas acima de tudo trago amor, um amor inigualável. Este guardado dentro de mim, no fundo do

meu ser, junto das minhas raízes, das minhas asas, e do amor que tenho para vos oferecer.

Os meus Pais, José e Rosa

É um facto curioso, este que noto ao chegar, os tesouros tornam a mala mais leve, em vez de a

carregar…

O trabalho apresentado nesta tese foi financiado pela Fundação para a Ciência e Tecnologia

(FCT), através da bolsa referência SFRH/BD/46660/2008.

O anticorpo monoclonal utilizado neste trabalho foi gentilmente cedido pela Biotecnol S.A.

(Lisboa, Portugal).

The authorization to evaluate the OSCARTM technology was gently conceded by the University of

Edinburgh.

vii

ABSTRACT

The advent of therapeutic recombinant proteins has revolutionized modern medicine. Since

the approval of recombinant insulin in 1982 to treat diabetes, many other recombinant proteins

have emerged for a diversity of previously incurable conditions. In these years, the manufacturing

processes have greatly evolved, but have also often disregarded product quality, an issue only

recently addressed and currently a major challenge in biopharmaceutical industry. Regulatory

agencies require now a tight control of product quality during production, which typically involves

characterization of the protein’s glycosylation, known to affect properties like serum half-life,

biological activity and immunogenicity. However, control of this property is not simple due to its

intrinsic high variability and its high sensitivity to small changes in the manufacturing process, in

ways that are far from being understood. Therefore, the development of products of consistent

high-quality requires a deeper understanding on the effect of the production parameters/

processes on glycosylation. In this context, the work described in this thesis intended to

contribute to this field of knowledge by evaluating changes on glycosylation of a monoclonal

antibody (mAb) produced by Chinese hamster ovary (CHO) cells during different stages of

process development, namely cell line transfection, serum-free media adaptation and culture.

To accomplish the main goal, a mAb-producing CHO-K1 cell line was initially established,

through a novel expression system (OSCARTM), potentially more expeditious than the traditional

methods. OSCARTM was indeed considerably faster, requiring about two months to obtain

producer cells compared to the minimum six months usually required. OSCARTM did not affect cell

growth and reached high levels of productivity (10 pg/cell/day) without requiring the typical

processes of amplification. The levels of productivity were initially difficult to stabilize, with a 10-

fold decay observed in the first weeks of culture, but remained stable thereafter for at least two

years. Furthermore, minigene selection was critical and seemed to be cell/product-specific. This

work also evaluated two methods for the initial selection of the highest producer clones obtained

after transfection, based on: absorbance values of supernatants; or mAb productivity calculated

using mAb yield determined by enzyme-linked immunosorbent assay and cell concentration

obtained from cell counting with a haemocytometer. It was shown that the methodology chosen

is highly influential on the product yield achievable in the final process of production, and that the

ABSTRACT│

viii

productivity-based method, albeit more laborious, is much more reliable than the commonly used

absorbance method. Additionally, the highest-producer clones were evaluated for mAb

glycosylation by high-performance liquid chromatography, with no differences detected.

Glycosylation was also assessed during the adaptation of mAb-producing cells to serum-free

medium using a gradual methodology and supplementation with trace elements. This process

was long and highlighted the importance of using proper media supplements, avoiding aggressive

procedures (centrifugation, enzymes), and giving cells enough time to adapt at each stage. More

importantly, it was shown that adaptation alters glycosylation: in the middle stages, fucosylation

decreased and galactosylation, sialylation and glycoform heterogeneity increased, with the first

two being positive and the last two undesirable for efficacy; in the final stages and after full

adaptation, fucosylation returned to the initial (serum-supplemented) levels, while galactosylation,

fucosylation and heterogeneity decreased, with the last two being positive. Divergences between

the stages were related to lower cell concentrations and viabilities in the middle stages of

adaptation, and to a shift of the growth mode of cells from adherent to suspended.

The impact of a microporous microcarrier culture on mAb glycosylation was evaluated and

compared to common T-flask cultures. The influence of several culture conditions was assessed,

including initial culture volume and cell concentration, rocking mechanism and speed, and

culture vessel. The microcarrier cultures led to a different mAb glycosylation compared to that

obtained in the T-flask culture, attributed to different mAb productivities, as well as to the use of

rocking and the generation of microenvironments (pH, accumulation of extracellular enzymes) in

the microcarrier cultures. Specifically, higher galactosylation and decreased fucosylation, both

beneficial changes, and a variable sialylation were found in the microcarrier cultures. Sialylation

was more sensitive to the culture parameters, particularly the type of culture vessel, being almost

absent in shake flask cultures. In addition to this advantageous modification, shake flasks also

led to a more homogeneous glycosylation, potentially due to improved cell densities.

In summary, the work described in this thesis contributes to a better understanding of how

glycosylation is affected by technologies used in process development, highlighting the need to

implement quality control at early stages. By increasing knowledge on this subject, it will be

possible to control and improve the quality and efficacy of therapeutic proteins, which will

ultimately lead to their administration at lower doses and frequency.

Keywords: Chinese hamster ovary cells; Monoclonal antibody; Glycosylation; Transfection;

Serum-free medium; Adaptation; Microcarriers; Quality control

ix

SUMÁRIO

EXPRESSÃO E CARACTERIZAÇÃO DE UM ANTICORPO MONOCLONAL TERAPÊUTICO EM CÉLULAS ANIMAIS

O desenvolvimento de proteínas terapêuticas recombinantes revolucionou a medicina atual,

trazendo novas formas de tratamento a doenças que se pensava incuráveis. A sua produção tem

sido alvo de grandes evoluções, que quase sempre negligenciaram a qualidade do produto.

Recentemente, este tornou-se um dos maiores desafios da indústria biofarmacêutica, com as

agências reguladoras a exigirem controlo da qualidade durante a produção. Este controlo envolve

normalmente a caracterização da glicosilação da proteína devido à sua influência em propriedades

como tempo de meia vida, imunogenicidade e atividade biológica, mas é dificultado pela sua

variabilidade intrínseca e pela sensibilidade a pequenas alterações nos processos de fabrico, de

forma ainda não compreendida. De facto, é necessário aprofundar o conhecimento atual sobre o

efeito dos parâmetros/processos de produção na glicosilação para a obtenção de produtos de

qualidade consistente. Neste contexto, o trabalho descrito na presente dissertação pretende

contribuir para o aumento do conhecimento na área, avaliando a glicosilação de um anticorpo

monoclonal (AcM) produzido por células animais durante diferentes estágios do desenvolvimento

de processos, nomeadamente transfecção, adaptação a meio sem soro e cultura.

Com este fim, foi estabelecida uma linha de células de ovário de hamster chinês (CHO-K1)

produtora de AcM, tendo-se avaliado um sistema de expressão (OSCARTM) recente e potencialmente

mais expedito do que os métodos tradicionais. De facto, este sistema foi mais célere, obtendo-se

células produtoras em apenas dois meses comparando com o mínimo de seis meses

normalmente necessário. Por outro lado, o sistema OSCARTM não afetou o crescimento celular e

permitiu obter elevadas produtividades (10 pg/cél/dia) sem recorrer a processos de amplificação.

Estes níveis de produtividade foram inicialmente difíceis de estabilizar, tendo-se observado um

decréscimo abrupto do seu valor nas primeiras semanas, mas mantendo-se posteriormente

estáveis por dois anos. Foram ainda avaliados dois métodos para seleção inicial dos clones mais

produtivos obtidos após transfeção, com base nos valores de absorvência dos sobrenadantes, ou

na produtividade calculada através das concentrações celular (contagem com hemocitómetro) e de

AcM (ensaio immunoenzimático) de cada clone. Demonstrou-se que a metodologia selecionada

tem uma grande influência nos níveis de produção atingidos no processo final de produção, sendo

SUMÁRIO│EXPRESSÃO E CARACTERIZAÇÃO DE UM ANTICORPO MONOCLONAL TERAPÊUTICO EM CÉLULAS ANIMAIS

x

o método da produtividade o mais aconselhado, pois embora mais trabalhoso é sem dúvida o de

maior precisão. Avaliou-se ainda a glicosilação do AcM produzido pelos clones selecionados, por

cromatografia líquida de alta eficiência, não se tendo detetado diferenças.

A glicosilação do AcM foi avaliada durante a adaptação celular gradual a meio sem soro, com

suplementação de oligoelementos. Este longo processo realçou a importância de usar suplementos

adequados, evitar procedimentos agressivos (centrifugação, enzimas) e permitir tempo suficiente

de adaptação a cada estágio. Demonstrou-se ainda que a glicosilação é alterada pelo processo:

nos passos intermédios houve decréscimo da fucosilação e aumento da galactosilação, sialilação e

heterogeneidade de glicoformas, sendo as duas primeiras positivas e as últimas indesejáveis; nos

últimos passos a fucosilação retomou os valores iniciais, e a galactosilação, fucosilação e

heterogeneidade diminuiram, sendo as duas últimas desejáveis. As diferenças entre passos

intermédios e finais estarão relacionadas com a concentração e viabilidade celular, inferiores nos

primeiros, e com a mudança no modo de crescimento celular de aderido para suspenso.

O impacto da cultura em suportes microporosos (Cytodex 3) na glicosilação do AcM foi

avaliada e comparada com a cultura aderida em frascos comuns. Analisou-se ainda a influência de

diferentes condições de cultura, como o volume e concentração celular iniciais, mecanismo e

velocidade de agitação, e frasco de cultura nos perfis de glicosilação. A cultura em suportes

microporosos originou diferentes padrões de glicosilação comparada com a cultura em frascos

comuns, atribuída a produtividades divergentes, assim como ao uso de agitação e à criação de

microambientes (pH, acumulação de enzimas extracelulares) nas culturas em suportes

microporosos. Em concreto, a galactosilação aumentou e a fucosilação diminuiu (mudanças

positivas), tendo a sialilação demonstrado ser mais sensível a parâmetros de cultura específicos,

particularmente ao tipo de frasco usado, sendo praticamente ausente em frascos agitados. Para

além desta modificação desejável, a cultura em frascos permitiu ainda uma glicosilação mais

homogénea, possivelmente devido a densidades celulares mais elevadas.

Resumindo, o trabalho efetuado contribui para uma melhor compreensão do efeito de

tecnologias envolvidas no desenvolvimento de processos na glicosilação do produto, salientando a

necessidade de implementar um controlo de qualidade já nos estágios iniciais. O aumento do

conhecimento sobre este tema irá, a longo prazo, permitir uma melhoria da eficácia das proteínas

terapêuticas, podendo resultar na redução da dose e frequência de administração necessárias.

Palavras-chave: Células de ovário de hamster chinês; Anticorpo monoclonal; Glicosilação;

Transfecção; Meio sem soro; Adaptação; Suportes microporosos; Controlo de qualidade

xi

TABLE OF CONTENTS

ACKNOWLEDGMENTS ................................................................................................................................................... v

ABSTRACT .................................................................................................................................................................... vii

SUMÁRIO ....................................................................................................................................................................... ix

LIST OF FIGURES .......................................................................................................................................................... xv

LIST OF TABLES ...........................................................................................................................................................xix

LIST OF ABBREVIATIONS, ACRONYMS AND SYMBOLS ........................................................................................... xxiii

THESIS PREAMBLE .................................................................................................................................................... xxxi

LIST OF PUBLICATIONS ............................................................................................................................................ xxxv

CHAPTER 1

GENERAL INTRODUCTION

1.1. INTRODUCTION ........................................................................................................................... 5

1.2. THERAPEUTIC PROTEINS AND THE SPECIFICITIES OF MONOCLONAL ANTIBODIES ..................... 6

1.3. EXPRESSION SYSTEM ................................................................................................................ 11

1.4. TRANSFECTION ......................................................................................................................... 13

1.4.1. EXPRESSION VECTOR ...................................................................................................................... 15

1.4.2. SELECTABLE MARKERS FOR STABLE TRANSFECTION ..................................................................... 17

1.4.3. DELIVERY METHODS ....................................................................................................................... 20

1.4.4. SELECTION...................................................................................................................................... 23

1.5. PROCESS OPTIMIZATION AT SMALL SCALE ................................................................................ 24

1.5.1. MODE OF GROWTH ......................................................................................................................... 24

1.5.2. MEDIUM .......................................................................................................................................... 26

1.5.3. PROCESS OPTIMIZATION ................................................................................................................. 27

1.6. GLYCOSYLATION........................................................................................................................ 28

1.6.1. BASIC CONCEPTS OF GLYCOSYLATION ........................................................................................... 29

1.6.2. BIOSYNTHESIS OF O-GLYCANS IN HUMAN CELLS ........................................................................... 30

1.6.3. BIOSYNTHESIS OF N-GLYCANS IN HUMAN CELLS ........................................................................... 32

1.6.4. GLYCOSYLATION AND THERAPEUTIC PROTEIN EFFICACY ............................................................... 38

1.6.5. EXTERNAL FACTORS INFLUENCING GLYCOSYLATION ..................................................................... 40

TABLE OF CONTENTS│

xii

1.6.6. GLYCOENGINEERING FOR IMPROVED THERAPEUTIC EFFICACY ...................................................... 53

1.6.7. STRATEGIES FOR GLYCOSYLATION ANALYSIS ................................................................................. 57

1.7. CONCLUDING REMARKS ............................................................................................................ 67

1.8. REFERENCES............................................................................................................................. 68

CHAPTER 2

THE OSCARTM SYSTEM FOR THE TRANSFECTION OF CHO CELLS FOR MONOCLONAL ANTIBODY PRODUCTION

2.1. INTRODUCTION ....................................................................................................................... 103

2.2. MATERIALS AND METHODS ..................................................................................................... 105

2.2.1. CELLS, PLASMIDS AND MINIGENES ............................................................................................... 105

2.2.2. SELECTION AND VALIDATION OF HPRT-DEFICIENT CLONES .......................................................... 105

2.2.3. TRANSFECTION ............................................................................................................................. 106

2.2.4. SELECTION OF THE HIGHEST MONOCLONAL ANTIBODY PRODUCERS ........................................... 107

2.2.5. ANTIBODY QUANTIFICATION BY ENZYME-LINKED IMMUNOSORBENT ASSAY ................................. 107

2.2.6. CELL GROWTH CHARACTERISTICS ................................................................................................ 109

2.2.7. ANALYSIS OF THE GLYCOSYLATION PROFILE ................................................................................. 109

2.3. RESULTS AND DISCUSSION ..................................................................................................... 113

2.3.1. ANALYSIS OF THE TWO MAIN PHASES OF THE OSCARTM TECHNOLOGY .......................................... 113

2.3.2. LEVEL AND STABILITY OF CLONE PRODUCTIVITY ........................................................................... 114

2.3.3. EFFECT OF TRANSFECTION ON CELL GROWTH CHARACTERISTICS ............................................... 118

2.3.4. ANALYSIS OF THE MONOCLONAL ANTIBODY GLYCOSYLATION PROFILE ........................................ 119

2.4. CONCLUSION .......................................................................................................................... 121

2.5. REFERENCES........................................................................................................................... 122

CHAPTER 3

IMPACT OF THE PROCESS OF CELL ADAPTATION TO SERUM-FREE CONDITIONS ON PRODUCT QUALITY

3.1. INTRODUCTION ....................................................................................................................... 129


3.2.1. CELL LINE AND CULTURE CONDITIONS ......................................................................................... 131

3.2.2. CELL ADAPTATION TO SERUM-FREE CONDITIONS ......................................................................... 131



3.4. CONCLUSION .......................................................................................................................... 148

3.5. REFERENCES........................................................................................................................... 149

│TABLE OF CONTENTS

xiii

CHAPTER 4

IMPACT OF MICROCARRIER CULTURE ON PRODUCT QUALITY

4.1. INTRODUCTION ....................................................................................................................... 161


4.2.1. CELL LINE AND CELL CULTURE ..................................................................................................... 163

4.2.2. MICROCARRIER PREPARATION ...................................................................................................... 163

4.2.3. MICROCARRIER CULTURE ............................................................................................................. 163

4.2.4. CULTURE MONITORING ................................................................................................................. 164

4.2.5. ANTIBODY QUANTIFICATION BY ENZYME-LINKED IMMUNOSORBENT ASSAY ................................. 165



4.4. CONCLUSION .......................................................................................................................... 178

4.5. REFERENCES........................................................................................................................... 179

CHAPTER 5

GENERAL CONCLUSIONS AND CONSIDERATIONS FOR FUTURE WORK

5.1. GENERAL CONCLUSIONS ........................................................................................................ 187

5.2. CONSIDERATIONS FOR FUTURE WORK .................................................................................... 192

xv

LIST OF FIGURES

CHAPTER 1

FIGURE 1.1. Representation of the structure of a human IgG antibody, composed of two light chains (L,

represented by traced rectangles) and two heavy chains (H, represented by empty rectangles). Both

L and H chains contain variable (VL, VH) and constant (CL, CH1-3) regions. The light chains are

connected to the heavy chains by disulfide bonds forming three independent protein moieties, two

fragment antigen binding (Fab) and one fragment crystallisable (Fc), which are connected through

the hinge region. The CH2 domains in the Fc region contain a highly conserved site (represented by

a star) for the covalent attachment of a glycan. ............................................................................... 9

FIGURE 1.2. Illustration of the antibody mechanisms for the elimination of antigen by (a) apoptosis, (b)

antibody-dependent cellular cytotoxicity (ADCC) mediated by cross-linking of leucocyte IgG-Fc

receptors, (c) engagement and activation of the C1 component of complement ............................. 10

FIGURE 1.3. Schematic diagram of (a) transient transfection and (b) stable transfection of foreign genetic

material (DNA, grey wave) into mammalian cells. In transient transfection (a) the foreign DNA enters

the nucleus and is expressed only temporarily due to a rapid loss of the foreign genetic material by

environmental factors and cell division. In stable transfection (b) the foreign DNA enters the nucleus

and is integrated into the genome of the host cell, allowing its stable expression over extended

periods of time. ............................................................................................................................ 14

FIGURE 1.4. Representation of the eight known core structures of mucin-type O-glycans. Core 1 to 4 are

the most prevalent; Core 7 has not been observed in humans. Core 1 has a Gal attached in a β1-3

linkage to the core GalNAc. Core 2 forms by the addition of a GlcNAc β1-6 linked to the GalNAc of

Core 1. Core 3 has a GlcNAc attached in a β1-3 linkage to the core GalNAc. Core 4 forms by the

addition of another GlcNAc β1-6 linked to the GalNAc of Core 3. Core 5 has a GalNAc attached in a

α1-3 linkage to the core GalNAc, Core 6 has a GlcNAc in a β1-6 linkage, Core 7 has a GalNAc in a

β1-6 linkage, and Core 8 has a Gal α1-3 linked to the core GalNAc. .............................................. 31

FIGURE 1.5. Pathway for the biosynthesis of the dolichol glycan precursor for protein N-glycosylation and

transfer of the precursor to asparagine residues (Asn-X-Thr/Ser) on nascently translated proteins by

the oligosaccharyltransferase (OST) complex. ................................................................................ 33

LIST OF FIGURES│

xvi

FIGURE 1.6. Processing of the initial high mannose N-glycan in the endoplasmic reticulum and cis-Golgi to

generate the core N-glycan substrate used for further diversification in the Golgi. ........................... 34

FIGURE 1.7. Representation of the N-glycan diversification in the Golgi, which generates three subtypes:

high-mannose, hybrid, and complex glycans.................................................................................. 36

FIGURE 1.8. Structures exemplifying complex N-glycans with varying number of antennae: biantennary,

triantennary, and tetra-antennary. ................................................................................................. 36

FIGURE 1.9. Schematic representation of N-glycan structures found in human cells and in the different

expression systems available for recombinant protein production. ................................................. 42

FIGURE 1.10. Approaches for the glycosylation analysis of proteins. The glycoprotein can be analyzed intact

or fractioned into glycopeptides, glycans or monosaccharides for more detailed assessment. The

analysis by one approach can be followed by another, with the most and less common analytical

pathways shown as full and dashed lines, respectively. ................................................................. 58

CHAPTER 2

FIGURE 2.1. Initial clone selection after transfection based on two approaches: (a) absorvance values of

supernatant read at 450 nm (ABS450), and (b) productivity (qmAb). The interrupted line denotes the

cutoff limit established for clone selection: (a) ABS450 of 100, and (b) qmAb of 10 pg/cell/day; and the

dark color highlights the selected clones. ............................................................................................. 115

FIGURE 2.2. Values of productivity (qmAb) of the selected clones obtained initially and after three weeks in

culture. .................................................................................................................................................... 116

FIGURE 2.3. Evolution of the productivity (qmAb) levels of clones 18, 27, 32, and 95 during 6 weeks in

culture. .................................................................................................................................................... 117

CHAPTER 3

FIGURE 3.1. N-glycosylation profile of the monoclonal antibody produced by CHO-K1 cells cultured in

different conditions during a process of gradual adaptation to serum-free medium ......................... 140

│LIST OF FIGURES

xvii

CHAPTER 5

FIGURE 5.1. Summary of the modifications, and potential biological effects, detected in the glycosylation

profile of the monoclonal antibody produced by CHO-K1 cells during adaptation to growth in serum-

free conditions. The divergences found in the glycosylation profile were divided into two stages:

during adaptation, and at the final stages of adaptation and thereafter. Probable causes for the

differences found are mentioned on the right....................................................................................... 189

FIGURE 5.2. Summary of the modifications, and potential biological effects, detected in the glycosylation

profile of the monoclonal antibody produced by CHO-K1 cells during microcarrier culture. Probable

causes for the differences found in comparison to the normal adherent culture in T-flasks are

mentioned on the left. ............................................................................................................................ 191

xix

LIST OF TABLES

CHAPTER 1

TABLE 1.1. The top-ten-selling biopharmaceutical products of 2010 ................................................................ 8

TABLE 1.2. Advantages and limitations of different expression systems for therapeutic glycoprotein production

................................................................................................................................................................... 12

TABLE 1.3. Characteristics of selective markers commonly used in expression vectors for the stable

transfection of mammalian cells ............................................................................................................. 18

TABLE 1.4. Comparison of biological, chemical and physical methods for the delivery of genetic material

into cells ...................................................................................................................................... 21

TABLE 1.5. Monosaccharides commonly present in mammalian glycoproteins, with respective abbreviation

and symbolic notation .............................................................................................................................. 30

TABLE 1.6. Divergences found on protein glycosylation occurring in different expression systems in

comparison to the glycosylation obtained in human cells, and respective effects on protein activity ..... 42

TABLE 1.7. Structural differences in the N-glycans produced by the main mammalian cells used as hosts

for glycoprotein production, concerning human glycosylation ............................................................... 44

TABLE 1.8. List of the main results reported on the effects of different variables of cell culture processes on

product glycosylation ................................................................................................................................ 47

TABLE 1.9. Exoglycosidase enzymes used for glycan analysis and respective monosaccharide and linkage

specificity ................................................................................................................................................... 62

TABLE 1.10. Incremental glucose unit (GU) values of different monosaccharides and linkages for the

structural assignment of 2-AB labeled N-glycans ................................................................................... 63

TABLE 1.11. Analytical techniques to characterize glycosylation, their basic principle, approaches in which

they apply, information provided and main advantages and limitations ............................................... 65

LIST OF TABLES│

xx

CHAPTER 2

TABLE 2.1. Characteristics of the disabled HPRT minigenes used in the OSCARTM expression system, in

comparison with the fully functional HPRT gene pBT/PGK-HPRT[RI] ................................................. 105

TABLE 2.2. Comparison of cloning rings and cloning disks for the isolation of HPRT-deficient clones

obtained during the first phase of the OSCARTM expression system, considering simplicity, speed and

cell recovery ............................................................................................................................................ 113

TABLE 2.3. Comparison of the three HPRT minigenes with varying degrees of expression disability for the

co-transfection of the CAB051 plasmid into the CHO-K1 cells using the OSCARTM expression system

................................................................................................................................................................. 114

TABLE 2.4. Values of productivity (qmAb) obtained for Clone 18 and Clone 32 after two years in culture..... 118

TABLE 2.5. Doubling times (tD) of the original CHO-K1 cells and of Clones 18 and 32, after two years in

culture ..................................................................................................................................................... 118

TABLE 2.6. Data of mean glucose unit (GU) value, tentative structure assignment, and relative area (%) of

the peaks obtained by high performance liquid chromatography for the monoclonal antibody

produced by Clones 18 and 32 ............................................................................................................. 119

CHAPTER 3

TABLE 3.1. Composition of the five combinations of supplements assessed for the adaptation of

monoclonal antibody-producing CHO-K1 cells to growth in serum-free conditions ............................ 132

TABLE 3.2. Description of the samples taken at different stages of the process of adaptation of monoclonal

antibody-producing CHO-K1 cells to serum-free medium, for the evaluation of the glycosylation

profile of the monoclonal antibody produced ....................................................................................... 133

TABLE 3.3. Timeline of the process of adaptation of the monoclonal antibody-producing CHO-K1 cells to

serum-free culture conditions ................................................................................................................ 138

TABLE 3.4. Data of mean glucose unit (GU) value, tentative assignment, and percentage of area of the

different peaks detected on the monoclonal antibody produced by CHO-K1 cells during the process

of adaptation to serum-free conditions. ................................................................................................. 142

TABLE 3.5. 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-free conditions ....... 144

│LIST OF TABLES

xxi

CHAPTER 4

TABLE 4.1. Conditions tested for the culture of the monoclonal antibody-producing CHO-K1 cells in Cytodex

3 (CY) microcarriers ............................................................................................................................... 164

TABLE 4.2. Data of mean glucose unit (GU) value, tentative structure assignment, and relative (%) area of

the peaks of the monoclonal antibody produced by CHO-K1 cells under different culture conditions

................................................................................................................................................................. 171

TABLE 4.3. Data of average cell concentration and productivity of monoclonal antibody-producing CHO-K1

cells cultured in different conditions ...................................................................................................... 172

TABLE 4.4. Galactose, core fucose and sialic acid composition of the monoclonal antibody produced by

CHO-K1 cells during microcarrier culture in different conditions ........................................................ 173

xxiii

LIST OF ABBREVIATIONS, ACRONYMS AND SYMBOLS

$ Dollar

% Percentage

® Registered trademark

µm Micrometer

µM Micromolar

2-AA 2-aminoanthranilic acid

2-AB 2-aminobenzamide

2-AP 2-aminopyridine

2D-PAGE Two-dimensional polyacrylamide gel electrophoresis

a Estimated response at zero concentration

ABS Absorbance

Abs Arthrobacter ureafaciens sialidase

ABSx Absorbance read at x nanometers

ADCC Antibody-dependent cellular cytotoxicity

AMAC 2-aminoacridone

Amf Almond meal fucosidase

ANOVA One-way analysis of variance

ANTS 8-aminonaphtalene-1,3,6-trisulfonic acid

Aph Aminoglycoside phosphotransferase

APS Ammonium peroxisulphate

Asn Asparagine

Asp Aspartate

ATCC American Type Culture Collection

b Slope factor

BHK Baby hamster kidney

Bkf Bovine kidney focusidase

BSA Bovine serum albumin

BSA-PBS Bovine serum albumin in phosphate buffered saline

Btg Bovine testis galactosidase

c Mid-range concentration

LIST OF ABBREVIATIONS, ACRONYMS AND SYMBOLS│

xxiv

C1 Complement component 1

C1q Complement component 1, q subcomponent

CaCl2 Calcium chloride

Cbg Coffee bean α-galactosidase

Ccell Cell concentration, in cells per milliliter

Ccell/microcarrier Cell concentration, in cells per microcarrier

CDC Complement-dependent cytotoxicity

CE Capillary electrophoresis

CE-LIF Capillary electrophoresis with laser-induced fluorescence detection

CE-LIF-MS Capillary electrophoresis with laser-induced fluorescence detection and mass

spectrometry

CE-MS Capillary electrophoresis and mass spectrometry

CH1-3 Constant region of the heavy chains, domain 1-3

CHO Chinese hamster ovary

CmAb Concentration of monoclonal antibody, in micrograms per mL

Cmicrocarriers in g/mL Concentration of microcarriers in grams of dry weight per milliliter

Cmicrocarriers/mL Concentration of microcarriers in number of microcarriers per milliliter

CMV Cytomegalovirus

CO2 Carbon dioxide

CuSO4 Copper sulfate

Cx Cell concentration at time x, in cells per milliliter

d Estimated response at infinite concentration

Da Dalton

DHFR Dihydrofolate reductase

DMEM Dulbecco’s Modified Eagle’s Medium

DMSO Dimethylsulfoxide

DNA Deoxyribonucleic acid

DO Dissolved oxygen

Dol-P Dolichol phosphate

DTT Dithiothreitol

e.g. Exempli gratia (for example)

EASE Expression augmenting sequence elements

ELISA Enzyme-linked immunosorbent assay

ENGase Endo-β-N-acetylglucosaminidase

EPO Erythropoietin

│ LIST OF ABBREVIATIONS, ACRONYMS AND SYMBOLS

xxv

ER Endoplasmic reticulum

ESI-MS Electrospray ionization mass spectrometry

et al Et alii (and others)

F Dilution factor

Fab Fragment antigen binding

FBS Fetal bovine serum

Fc Fragment crystallisable

FcγR Fragment crystallisable gamma receptor

FcRn Neonatal fragment crystallisable receptor

FDA Food and Drug Administration

FeC6H5O7.NH4OH Ammonium iron citrate

FG0 Core fucosylated agalactosylated

FG1 Core fucosylated monogalactosylated

FG2 Core fucosylated digalactosylated

FG3 Core fucosylated trigalactosylated

Fuc, F Fucose

fut8 α1,6-fucosyltransferase gene

g Centrifugal force relative to the gravitational force of the earth

g Gram

G0 Agalactosylated

G1 Monogalactosylated

G2 Digalactosylated

G3 Trigalactosylated

G418 Geneticin

GAG Glycosaminoglycan

Gal, G Galactose

GalNAc N-acetylgalactosamine

GalT Galactosyltransferase

GC Gas chromatography

GC-MS Gas chromatography and mass spectrometry

GDP Guanosine diphosphate

Glc Glucose

GlcA Glucoronic acid

GlcNAc, A N-acetylglucosamine

Glu Glutamate


xxvi

GnT N-acetylglucosaminyltransferase

GS Glutamine synthetase

GU Glucose unit

h Hour

HAT Hypoxanthine, aminopterin, thymidine

HBS Hydroxyethyl piperazineethanesulfonic acid-buffered salt solution

HCO3- Bicarbonate

HEK Human embryonic kidney

HEPES Hydroxyethyl piperazineethanesulfonic acid

HILIC Hydrophilic interaction chromatography

hisD Histidinoldehydrogenase

HPAEC High performance anion exchange chromatography

HPAEC-PAD High performance anion exchange chromatography with pulse amperometric

detection

Hph Hygromycin-B phosphotransferase

HPLC High performance liquid chromatography

HPRT Hypoxanthine phosphoribosyltransferase

i.e. Id est (that is, in other words)

IAA Iodoacetamide

IBM International Business Machines

IEF-PAGE Isoelectric focusing polyacrylamide gel electrophoresis

IFN Interferon

Ig Imunoglobulin

Jbh Jack bean N-acetylhexosaminidase

Jbm Jack bean mannosidase

KCl Potassium chloride

KDa Kilodalton

KH2PO4 Potassium dihydrogen phosphate

L Liter

LC-MS Liquid chromatography and mass spectrometry

Leu Leucine

Log Logarithm

m Meter

M Molar

mAb Monoclonal antibody


xxvii

MALDI Matrix assisted laser desorption ionization

MALDI-MS Matrix assisted laser desorption ionization mass spectrometry

MALDI-MS-TOF Matrix assisted laser desorption ionization mass spectrometry with time of flight

analysis

Man, M Mannose

ManNAc N-acetyl mannosamine

MARS Matrix-attachment regions

MDCK Madin Darby canine kidney cells

mg Milligram

min Minute

mL Milliliter

mM Millimolar

mm2 Square millimeter

mRNA Messenger ribonucleic acid

MS Mass spectrometry

MSX Methionine sulfoxamide

MTX Methotrexate

MWCO Molecular weight cut-off

MΩ.cm Mega omega centimeter, measure of resistivity

Na2HPO4 Sodium phosphate dibasic

Na2HPO4.2H2O Sodium phosphate dibasic dihydrate

Na2SeO3 Sodium selenite

NaCl Sodium chloride

Nan1 Streptococcus pneumoniae sialidase

Ncells Number of cells

NeuGc N-glycolylneuraminic acid

NeuNAc N-acetylneuraminic acid

NH4Fe(SO4)2 Ferrous ammonium sulfate

NH4VO3 Ammonium metavanadate

NIBRT National Institute for Bioprocessing, Research and Training

NiCl2 Nickel chloride

NK cells Natural killer cells

nm Nanometer

Nmicrocarriers in sample Number of microcarriers in a sample

Nmicrocarriers/g dry weight Approximate number of microcarriers per gram of dry weight


xxviii

NMR Nuclear magnetic resonance

NP-HPLC Normal-phase high-performance liquid chromatography

NS0 Murine myeloma cells

º Degree

ºC Celsius degrees

OST Oligosaccharyltransferase

Pac Puromycin N-acetyltransferase

PAD Pulsed amperometric detection

PAS Periodic acid Schiff

PBS Phosphate buffered saline

pDWM128 Hypoxanthine phosphoribosyltransferase minigene



PER.C6 Human embryonic retinal cells

pg Picogram

Phe Phenylalanine

pI Isoelectric point

PNGase Peptide N-glycosidase

Pro Proline

qmAb Monoclonal antibody productivity, in picograms per cell per day

rhInsulin Recombinant human insulin

RNA Ribonucleic acid

RP-HPLC Reverse phase high-performance liquid chromatography

rpm Revolutions per minute

S Sialic acid

SA Anonymous society

S1 Monosialylated

S1G1 Monosialylated monogalactosylated

S1G2 Monosialylated digalactosylated

S2 Disialylated

S2G2 Disialylated digalactosylated

S2G3 Disialylated trigalactosylated

S3 Trisialylated

S3G3 Trisialylated trigalactosylated

SARS Scaffold-attachment regions


xxix

SDS Sodium dodecyl sulfate

SDS-PAGE Sodium dodecyl sulfate polyacrylamide gel electrophoresis

Ser Serine

sh ble Zeocin/bleomycin resistance gene

SnCl2 Stannous chloride

SOP Standard operating procedure

Sp2/0 Murine myeloma cells

Spg Streptococcus pneumoniae galactosidase

Sph Streptococcus pneumoniae N-acetylhexosaminidase

SPSS Statistical Package for the Social Sciences

SV40 Simian virus 40

t Time

tD Cell doubling time, in hours

TEMED Tetramethylethylenediamine

Thr Threonine

TIMP Tissue inhibitor of metalloproteinases

TK Thymidine kinase

TM Trademark

TMB 3,3’,5,5’-Tetramethylbenzidine

tPA Tissue plasminogen activator

Trp Tryptophan

UDP Uridine diphosphate

UK United Kingdom

USA United States of America

v/v Volume to volume

Vsample Volume of sample, mL

WAX-HPLC Weak anion exchange high-performance liquid chromatography

Xyl Xylose

ZnSO4 Zinc sulfate

xxxi

THESIS PREAMBLE

SCOPE OF THE THESIS

For many years, the biopharmaceutical industry has centered its efforts on increasing the

capacity of their processes to obtain higher quantities of the product at lower costs. Only recently

concerns about the quality of the therapeutic products obtained in these processes have risen,

resulting in a change of priorities to accommodate this crucial parameter. Indeed, the regulatory

agencies have now settled several requirements for the approval of new therapeutics that include

the characterization of properties related to their quality/efficacy, and which must be performed

throughout the lifecycle of the product. Of those properties, glycosylation has been considered as

one of the most critical, and has drawn the attention of the biopharmaceutical industry. This post-

translational modification is known to be affected by several process changes, ranging from the

choice of the expression system to the final culture conditions, which may change the therapeutic

efficacy of the product, the frequency and dosage required, and even its potential side effects. As

a result, biopharmaceutical industry is trying to understand the ways in which the incredible

diversity of parameters involved in process development for the production of therapeutics affect

glycosylation. As a more far-reaching objective, glycosylation is being regarded as a potential

means to improve the quality of the product, which may be accomplished with glycoengineering

methods or more simply through the manipulation of the culture conditions.

THESIS PREAMBLE│

xxxii

AIM OF THE THESIS

Taking into consideration the facts exposed, it is clear that a more in-depth understanding

about how glycosylation is affected by different processes of biopharmaceutical manufacturing is

crucial for the development of high quality therapeutics. The work described in the present thesis

intended to contribute to the progress in this field by evaluating the impact of different steps of

early process development on the glycosylation profile of a monoclonal antibody (mAb) produced

by mammalian cells.

The first step of process development is to establish a producer cell line by transfecting cells

with the product gene. This is commonly a time-consuming process that yields a high number of

cell clones, which must be screened for the selection of the highest-producers. The clones

obtained have different growth and production characteristics and could therefore also yield a

product with divergent glycosylation profiles. Therefore, the first aim of the study presented

herein was to evaluate a novel, potentially faster, methodology for the transfection of a Chinese

hamster ovary (CHO) cell line for mAb production, followed by clone selection and comparison of

the mAb glycosylation profile among the highest-producer clones.

Another important stage of process development is the adaptation of the producing cells to

serum-free culture, due to regulatory and safety concerns that require the elimination of any

animal components from the culture media. This process of adaptation has repercussions on cell

growth, viability and product yield, and may also affect product quality. Consequently, a second

aim of the work was to assess the impact of each of the steps of a gradual adaptation of mAb-

producting cells to serum-free culture on the resulting mAb glycosylation profile, including the test

of different combinations of media supplements.

Biopharmaceutical processes are typically performed with cells growing in suspension due

to the highest cell and product yields provided when compared to adherent culture. However,

alternative systems for high-yield anchorage-dependent culture are now being established and

used, with microcarriers currently regarded as the most successful technology in the field. The

establishment of microcarrier culture requires the optimization of several parameters that affect

cell growth, product yield and potentially quality. Therefore, the final task of this work was to

assess the impact of microcarrier culture on mAb glycosylation, in comparison to normal T-flask

cultures, and to evaluate the profile obtained under different culture conditions.

│THESIS PREAMBLE

xxxiii

OUTLINE OF THE THESIS

The work developed to accomplish the scientific goals previously outlined is reported in this

thesis in the form of five chapters, according to the following organization:

CHAPTER 1 overviews the most relevant steps of process development for biopharmaceutical

production of therapeutic proteins. Particular emphasis is given to glycosylation, covering its

influence on the therapeutic efficacy of the product, the current knowledge about the process

parameters that affect its profile, the methods employed for its monitoring, and the advances

made in the field of glycoengineering to improve product quality.

CHAPTER 2 evaluates the novel OSCARTM expression system for the transfection of CHO cells for

mAb expression. It covers the impact of this methodology on cell growth characteristics and

levels and stability of production, and compares the glycosylation profile obtained with different

cell clones.

CHAPTER 3 is centered on a process of gradual adaptation of mAb-producing CHO cells to

serum-free culture, and the analysis of the effects of its different steps on the glycosylation profile

of the product.

CHAPTER 4 assesses the impact of microcarrier culture on mAb glycosylation, evaluating the

influence of several culture parameters typically used for its optimization.

CHAPTER 5 closes the thesis by connecting the most significant conclusions of the former

chapters and suggesting paths for future research.

xxxv

LIST OF PUBLICATIONS

The work developed during the four years of the PhD program resulted in several

publications in peer-reviewed international journals and book chapters, as well as abstracts in

proceedings of international conferences. These publications are compiled below.

PAPERS IN PEER-REVIEWD INTERNATIONAL JOURNALS

ANA RITA COSTA, Maria Elisa Rodrigues, Mariana Henriques, Rosário Oliveira, Joana Azeredo

(2010). Guidelines to cell engineering for monoclonal antibody production. European Journal

of Pharmaceutics and Biopharmaceutics, 74 (2):127-138.

Maria Elisa Rodrigues, ANA RITA COSTA, Mariana Henriques, Joana Azeredo, Rosário Oliveira

(2010). Technological progresses in monoclonal antibody production systems. Biotechnology

Progress, 26 (2):332-351.

ANA RITA COSTA, Maria Elisa Rodrigues, Mariana Henriques, David W Melton, Philip Cunnah,

Rosário Oliveira, Joana Azeredo (2012). Evaluation of the OSCARTM system for the production

of monoclonal antibodies by CHO-K1 cells. International Journal of Pharmaceutics, 430 (1-

2):42-6.

ANA RITA COSTA, Joanne Withers, Maria Elisa Rodrigues, Niaobh McLoughlin, Mariana Henriques,

Rosário Oliveira, Pauline M Rudd, Joana Azeredo (2012). The impact of cell adaptation to

serum-free conditions on the glycosylation profile of a monoclonal antibody produced by CHO

cells. Submitted to New Biotechnology.

ANA RITA COSTA, Joanne Withers, Maria Elisa Rodrigues, Niaobh McLoughlin, Mariana Henriques,

Rosário Oliveria, Pauline M Rudd, Joana Azeredo (2012). The impact of microcarrier culture

optimization on the glycosylation profile of a monoclonal antibody. Submitted to Journal of

Industrial Microbiology and Biotechnology.

LIST OF PUBLICATIONS│

xxxvi

ANA RITA COSTA, Maria Elisa Rodrigues, Mariana Henriques, Rosário Oliveira, Joana Azeredo

(2012). Glycosylation: impact, control and improvement during therapeutic protein

production. Submitted to Critical Reviews in Biotechnology.

ANA RITA COSTA, Maria Elisa Rodrigues, Mariana Henriques, Joana Azeredo, Rosário Oliveira

(2012). Comparison of commercial serum-free media for CHO-K1 cell growth and monoclonal

antibody production. International Journal of Pharmaceutics, 437(1-2):303-305.


(2012). Wave characterization for mammalian cell culture: residence time distribution. New

Biotechnology, 29 (3):402-408.

Maria Elisa Rodrigues, ANA RITA COSTA, Mariana Henriques, Philip Cunnah, Joana Azeredo,

David W Melton, Rosário Oliveira (2012). Advances and drawbacks of the adaptation to

serum-free culture of CHO-K1 cells for monoclonal antibody production. Submitted to Applied

Biochemistry and Biotechnology.

Maria Elisa Rodrigues, ANA RITA COSTA, Pedro Fernandes, Mariana Henriques, Philip Cunnah,

David W Melton, Joana Azeredo, Rosário Oliveira (2012). Evaluation of macroporous and

microporous carriers for CHO-K1 cell growth and monoclonal antibody production. Submitted

to Biotechnology Progress.


(2012). Evaluation of microcarriers for the production of a monoclonal antibody by suspended

CHO-K1 cells in serum-free media. Submitted to Journal of Biotechnology.

BOOK CHAPTERS

ANA RITA COSTA, Maria Elisa Rodrigues, Mariana Henriques, Rosário Oliveira, Joana Azeredo.

Feed optimization in fed-batch culture. In: Portner, R (ed.). Animal Cell Biotechnology:

Methods and Protocols, third edition. Humana Press. Forthcoming 2013.

│LIST OF PUBLICATIONS

xxxvii

Maria Elisa Rodrigues, ANA RITA COSTA, Mariana Henriques, Joana Azeredo, Rosário Oliveira.

Evaluation of solid and porous microcarriers for cell growth and production. In: Portner, R

(ed.). Animal Cell Biotechnology: Methods and Protocols, third edition. Humana Press.

Forthcoming 2013.

ABSTRACTS AND PROCEEDINGS IN INTERNATIONAL CONFERENCES

ANA RITA COSTA, Maria Elisa Rodrigues, Mariana Henriques, Joana Azeredo, Rosário Oliveira.

Assessment of different approaches for the adaptation of CHO cells to serum-free medium.

BIT’s Life Sciences 2nd Annual International Congress of Antibodies, Beijing, China, 24-26

March 2010. Book of Abstracts, p. 493.


Wave bioreactor characterization: residence time distribution. BIT’s Life Sciences 2nd Annual

International Congress of Antibodies, Beijing, China, 24-26 March 2010. Book of Abstracts, p.

494.


Characterization of the Wave bioreactor: residence time distribution determination. HAH 2010

– Human Antibodies and Hybridomas, Porto, Portugal, 14-16 April 2010.

ANA RITA COSTA, Maria Elisa Rodrigues, Mariana Henriques, Rosário Oliveira, Joana Azeredo.

Strategies for adaptation of mAb-producing CHO cells to serum-free medium. 22nd European

Society for Animal Cell Technology (ESACT) Meeting on Cell Based Technologies, Vienna,

Austria, 15-18 May 2011. BMC Proceedings, 5(Suppl 8):112.

Maria Elisa Rodrigues, Pedro Fernandes, ANA RITA COSTA, Mariana Henriques, Joana Azeredo,

Rosário Oliveira. Preliminary evaluation of microcarrier culture for growth and monoclonal

antibody production of CHO-K1 cells. 22nd European Society for Animal Cell Technology

(ESACT) Meeting on Cell Based Technologies, Vienna, Austria, 15-18 May 2011. BMC

Proceedings, 5(Suppl 8):111.

hapter hapter hapter hapter 1111

General introduction

CC

THE PRESENT CHAPTER WAS ADAPTED FROM THE FOLLOWING REVIEW PAPERS

ANA RITA COSTA, Maria Elisa Rodrigues, Mariana Henriques, Joana Azeredo, Rosário

Oliveira (2012). Glycosylation: impact, control and improvement during therapeutic

protein production. Submitted to Critical Reviews in Biotechnology.

ANA RITA COSTA, Maria Elisa Rodrigues, Mariana Henriques, Joana Azeredo, Rosário

Oliveira (2010). Guidelines to cell engineering for monoclonal antibody production.

European Journal of Pharmaceutics and Biopharmaceutics, 74:127-138.

3

he emergence of biopharmaceutical industry represented a major revolution for

modern medicine, through the development of recombinant therapeutic proteins that brought a

new breath of hope for many patients with previously untreatable diseases. The ever-growing

demand for this type of therapeutics forces the industry to be in constant evolution to increase

product yields while simultaneously reducing associated costs. However, process changes may

also affect the quality of the product, for example by modifying its glycosylation profile, a factor

that was initially overlooked but whose monitoring and control is now a major concern in

biopharmaceutical manufacturing and a requirement of regulatory agencies. In this scope, the

present chapter provides a review of the current status of process development for the

manufacture of therapeutic proteins, with emphasis on monoclonal antibodies, and of the

present knowledge about the impact of its different stages on product quality, particularly on the

glycosylation profile. The current approaches used to monitor this parameter are addressed, as

well as the strategies that are being explored to use glycosylation control as a means to improve

the therapeutic efficacy of the product.

Keywords: Therapeutic proteins; Monoclonal antibody; Expression system; Mammalian cells;

Transfection; Process optimization; Small-scale; Glycosylation; Glycoengineering; Analytical

methodologies

TT

CHAPTER 1 │ GENERAL INTRODUCTION

5

1.1. INTRODUCTION

The biopharmaceutical industry has revolutionized the treatment of many previously unmet

medical needs (Li and d’Anjou 2009) and currently represents the most significant and fastest

growing segment of the overall pharmaceutical market (Walsh 2010). An annual growth of 7-15 %

is expected in the next several years (Hillor 2009), having therapeutic proteins and particularly

monoclonal antibodies (mAbs) as the main drivers of this growth (Walsh 2010; Li and d’Anjou

2009). The relevance of mAbs is highlighted by a consistent and significant increase on the

number of mAbs approved in the last years for a wide range of therapeutic applications (Walsh

2010, 2003; Nilsang et al. 2007) that include the treatment of cancer, autoimmune disorders

and allergic diseases (Ozturk and Hu 2006).

The development of processes for the production of mAbs and other biopharmaceutical

proteins still faces many technological challenges at different stages of process development,

encompassing the selection of an appropriate expression system, the establishment of a suitable

producing cell line (transfection), the process optimization, and the scale-up and bioreactor

production (Baker, Flatman, and Birch 2007).

The primary goal of process development has been the improvement of volumetric

productivity (Woolley and Al-Rubeai 2009), with the current values of 10-15 g/L demonstrating a

major progress from the 100 mg/L initially obtained (Huang et al. 2010; Wurm 2004). This

upgrade is mainly attributed to advances on expression vectors (Aldrich, Viaje, and Morris 2003),

process control (Gagnon et al. 2011; Rodriguez et al. 2005; Trummer et al. 2006; Yoon et al.

2006) and host cell engineering (Seth et al. 2006), as well as to optimizations of culture medium

and feeding strategies (Huang et al. 2010; Wurm 2004; Altamirano et al. 2004).

Recently, the awareness about the potential impact of process changes on the quality of the

product has risen concerns that proteins lacking structural fidelity may provoke an immune

response that ultimately impacts therapeutic efficacy and/or causes harmful reactions (side

effects) (Baker, Flatman, and Birch 2007; Smalling et al. 2004). As a consequence of this

knowledge, biopharmaceutical industry is now considering both product yield and product quality

during process development, and the regulatory agencies have added product control and

characterization to their requirements. This characterization must be performed throughout the

product lifecycle, and is mainly focused on post-translational modifications, in special


6

glycosylation, due to their known impact on a vast range of critical protein properties that

ultimately affect therapeutic efficacy (Baker, Flatman, and Birch 2007; Walsh and Jefferis 2006;

Spearman et al. 2005).

This chapter presents an up-to-date review about the early stages of process development

for the manufacture of therapeutic proteins, in particular mAbs, including the construction and

selection of an adequate cell line and the optimization of the process at small scale. The potential

impact of these steps on the quality of the product, more precisely on the glycosylation profile, is

discussed, as well as the current strategies employed to control and improve this property.

1.2. THERAPEUTIC PROTEINS AND THE SPECIFICITIES OF MONOCLONAL ANTIBODIES

Therapeutic proteins have been around for more than a century, since the pioneer work of

Behring and Kitasato, which resulted in the market launch of the first therapeutic protein in

1894, an anti-diphtheria serum to combat a serious diphtheria epidemic in Europe (Strohl 2009).

After this, other therapeutic proteins derived from human or animal sources emerged, as

exemplified by the remarkable introduction of heterologous insulin purified from pigs and cows

(Iletin®) by Eli Lilly in 1923, which started a life-saving treatment for patients with type I diabetes

mellitus (Strohl 2009).

However, concerns about the purity, consistency and, in particular, safety (potential viral

contamination) of these human/animal-derived products, resulted in a slow increment of their

therapeutic use (Sethuraman and Stadheim 2006). This changed in the last 30 years with the

emergence of recombinant protein technology (Strohl and Knight 2009; Sethuraman and

Stadheim 2006), which basically consists on the expression of cloned genetic material (the

protein) within living cells (the in vitro biological production system) (Jayapal et al. 2007). Once

again, insulin led the path and was the first recombinant human protein to be approved and

marketed, as Humulin®, in 1982 by Genentech and Eli Lilly, eliminating the issues of immune

response directed against the former heterologous insulins obtained from livestock (Strohl 2009;

Strohl and Knight 2009).


7

Since the advent of Humulin®, more than 200 therapeutic proteins, majorly of recombinant

sources, have been brought to the market for a broad array of therapeutic diseases, having a

huge impact on modern medicine by bringing novel and life-saving therapies to patients

(Durocher and Butler 2009; Walsh 2010). According to their application/function, these

biopharmaceutical proteins can be categorized into four major groups: protein therapeutics with

enzymatic and regulatory activity (e.g. insulin, human growth hormone), protein therapeutics with

special targeting activity (e.g. mAbs), protein vaccines (e.g. human papilloma virus vaccine), and

protein diagnostics (e.g. biomarkers and imaging agents) (Leader, Baca, and Golan 2008).

All categories together totaled sales of $99 billion by 2009, representing the fastest growing

segment of the pharmaceutical industry. In particular, mAb-based products constitute the most

preeminent group of the biopharmaceutical market, as demonstrated by the presence of five

mAb products in the top-ten-selling biopharmaceuticals of 2010 (and four in the top five) (Table

1.1) (Walsh 2010; Huggett, Hodgson, and Lahteenmaki 2011). Indeed, the recombinant mAbs

currently licensed represent a substantial success both in terms of revenue generated and

clinical benefit delivered, with therapeutic applications ranging from the dominant areas of

oncology, rheumatoid and autoimmune diseases, to other indications such as organ

transplantation, cardiology, viral infection, allergy, tissue growth and repair, and hemoglobinuria

(Beck et al. 2008; Jefferis 2009). Furthermore, with 240 mAb products currently in clinical trials,

along with an additional 120 recombinant proteins, it is expected that mAb-based approvals will

continue to dominate the market in the coming years (Sheridan 2010).

The efficacy of mAbs is a result of their specificity for target antigens and of their biological

activities, also called ‘effector functions’, which might be activated by the immune complexes

formed (Jefferis 2006; Jefferis 2009; Natsume, Niwa, and Satoh 2009). These effector functions

of mAbs diverge among the five distinct classes defined within human antibodies

(imunoglobulins, Ig): IgM, IgG, IgA, IgD and IgE (Burton and Woof 1992). Of these, the IgG class

has been the most studied due to its predominance in serum, and as a result of the accumulated

knowledge about its structure and function it became the class of choice in recombinant

technology for therapeutic application, currently comprising all the approved mAbs (Jefferis

2009; Wagner-Rousset et al. 2008; Natsume, Niwa, and Satoh 2009; Beck et al. 2008).


8

TABLE 1.1. The top-ten-selling biopharmaceutical products of 2010 (adapted from Huggett, Hodgson, and

Lahteenmaki 2011)

NAME CLASS INDICATION SALES $ billions

COMPANY

ENBREL

(etanercept)

Fusion protein Rheumatoid and psoriatic arthritis, ankylosing

spondylitis, psoriasis

6.808 Amgen

HUMIRA

(adalimumab)

Monoclonal

antibody

Rheumatoid and psoriatic arthritis, psoriasis, ulcerative

colitis, ankylosing spondylitis, Crohn’s disease

6.548 Abbott

REMICADE

(infliximab)

Monoclonal

antibody

Psoriasis, ulcerative colitis, ankylosing spondylitis,

Crohn’s disease, psoriatic and rheumatoid arthritis

6.478 Johnson &

Johnson

AVASTIN

(bevicuzumab)

Monoclonal

antibody

Colorectal, breast, brain, renal cell, and non-small cell

lung cancers

6.193 Roche

RITUXAN

(rituximab)

Monoclonal

antibody

Non-Hodgkin’s lymphoma, rheumatoid arthritis, chronic

lymphocytic leukemia

6.088 Biogen-Idec

HERCEPTIN

(trastuzumab)

Monoclonal

antibody

Breast cancer 5.210 Roche

LANTUS

(insulin glargine)

Hormone Diabetes mellitus, type 1 and 2 4.653 Sanofi

LOVENOX

(enoxaparin)

Low molecular

weight heparin

Venous thromboembolism, acute coronary syndrome 3.722 Sanofi

NEULASTA

(pegfilgrastim)

Growth factor Neutropenia/leucopenia 3.558 Amgen

COPAXONE

(glatiramer)

Immunomodulator Multiple sclerosis 3.315 Teva

MAbs of the IgG class have a basic structure of two light and heavy chains in covalent and

non-covalent association, with variable and constant regions, that result in the formation of three

independent protein moieties – two fragment antigen binding (Fab) regions and one fragment

crystallisable (Fc) region – which are connected through a flexible linker named the hinge region

(Figure 1.1) (Natsume, Niwa, and Satoh 2009; Beck et al. 2008).


9

FIGURE 1.1. Representation of the structure of a human IgG antibody, composed of two light chains (L, represented

by traced rectangles) and two heavy chains (H, represented by empty rectangles). Both L and H chains contain

variable (VL, VH) and constant (CL, CH1-3) regions. The light chains are connected to the heavy chains by disulfide

bonds forming three independent protein moieties, two fragment antigen binding (Fab) and one fragment

crystallisable (Fc), which are connected through the hinge region. The CH2 domains in the Fc region contain a highly

conserved site (represented by a star) for the covalent attachment of a glycan (adapted from Natsume, Niwa, and

Satoh 2009).

Both Fab regions of an individual antibody are identical in structure and express a specific

antigen-binding site. For its turn, the Fc region expresses interaction sites for ligands that activate

the effector mechanisms (Jefferis, Lund, and Pound 1998; Jefferis and Lund 2002). These

effector ligands include three homologous cellular Fc receptor types (FcγRI, FcγRIIa-c,

FcγRIIIa,b), the C1q component of the complement, and the neonatal Fc receptor (FcRn)

(Nimmerjahn and Ravetch 2008; Burton and Woof 1992). The interaction sites of the Fc region

to the effector ligands are mostly located in the hinge and in the second domain of the constant

region of the heavy chains (CH2) (Figure 1.1). These CH2 domains bear a highly conserved site

(asparagine 297, Asn-297) for the covalent attachment of a glycan, which strongly influences the

conformation of the Fc region and its effector functions (Wuhrer et al. 2007; Clark 1997; Morgan

et al. 1995; Burton and Woof 1992).

As a consequence of the IgG structure, the physiological activities may be mediated by three

mechanisms (Figure 1.2): the direct neutralization of the antigen and/or induction of apoptosis

through the specific and bivalent binding to the target antigen in the Fab region; the activation of

the complement cascade through the initial binding and activation of the C1q molecule in the Fc


10

region, which culminates in complement

inflammatory and immunoregulatory molecules;

macrophage, natural killer (NK) cells) and mediation of antibody

(ADCC) through the binding of the Fc receptor in the IgG to the IgG

leucocytes. Additionally, the FcRn receptor controls the catabolic half

has a role in placental transpor

Woof 1992; Nimmerjahn and Ravetch 2008

FIGURE 1.2. Illustration of the antibody mechanisms for the elimination of antigen by (a) apoptosis, (b) antibody

dependent cellular cytotoxicity (ADCC)

activation of the C1 component of complement

The IgG class of mAbs is further divided into four subclasses, defined by different heavy

chains and numbered according to their relative concentration in serum: IgG1, IgG2, IgG3, and

IgG4 (Beck et al. 2008). Each subclass

choice when developing recombinant mAb therapeutics is being

critical decision, and may vary with the specific therapeutic application

Braslawsky 2002). For example, in oncology it seems beneficial to maximize the potential to

region, which culminates in complement-dependent cytotoxicity (CDC) and generation of

inflammatory and immunoregulatory molecules; and the stimulation of effector cells (

macrophage, natural killer (NK) cells) and mediation of antibody-dependent cellular cytotoxicity

(ADCC) through the binding of the Fc receptor in the IgG to the IgG-Fc receptors present in

leucocytes. Additionally, the FcRn receptor controls the catabolic half-life of the IgG molecule and

has a role in placental transport from mother to fetus (Wagner-Rousset et al. 2008

Nimmerjahn and Ravetch 2008; Basta 2008; Kibe et al. 1996).

Illustration of the antibody mechanisms for the elimination of antigen by (a) apoptosis, (b) antibody

(ADCC) mediated by cross-linking of leucocyte IgG-Fc receptors, (c) engagement and

activation of the C1 component of complement (adapted from Jefferis 2009).



. Each subclass expresses a unique profile of effector functions, and

choice when developing recombinant mAb therapeutics is being increasingly considered as a

critical decision, and may vary with the specific therapeutic application (Reff, Hariharan, and

mple, in oncology it seems beneficial to maximize the potential to

dependent cytotoxicity (CDC) and generation of

ation of effector cells (e.g.

ependent cellular cytotoxicity

Fc receptors present in

life of the IgG molecule and

Rousset et al. 2008; Burton and

Illustration of the antibody mechanisms for the elimination of antigen by (a) apoptosis, (b) antibody-

Fc receptors, (c) engagement and



expresses a unique profile of effector functions, and its

increasingly considered as a

Reff, Hariharan, and

mple, in oncology it seems beneficial to maximize the potential to


11

induce ADCC and CDC to eliminate the target cancer cells; however, in chronic diseases the

central objective might be the neutralization of a soluble target, without excessive effector activity,

which could be detrimental if the target is also a membrane protein on host cells (Jefferis 2009;

Jefferis 2006). Therefore, although the predominant choice for therapeutic applications is

currently IgG1, as a consequence of its long-life in serum (about 21 days) and its enhanced

effector functions, the other IgG subclasses are now also being evaluated (Burton and Woof

1992; Wuhrer et al. 2007; Clark 1997; Jefferis 2009, 2007; Beck et al. 2008). In particular, the

IgG2 appears as a good alternative when minimum effector functions are desired, also

demonstrating a long serum half-life and stability (Jefferis 2006; Jefferis 2007).

1.3. EXPRESSION SYSTEM

The choice of an expression system (host cell) for the production of therapeutic proteins has a

profound impact on the product characteristics and the maximum expression yields that are

possible to obtain (Jayapal et al. 2007). Therefore, a number of requirements should be met by

the cells to be apt for biopharmaceutical application. First, cells should be amenable to genetic

modifications, allowing the easy introduction of foreign deoxyribonucleic acid (DNA), and be able

to support high-level product expression over long periods of time while maintaining high viable

cell density and genetic stability. Secondly, from an industrial perspective it is highly desirable

that cells are able to adapt to growth in suspension and serum-free conditions to allow for

volumetric and simple scalability, and to reduce safety concerns about the presence of animal-

derived components. Thirdly, for the production of complex proteins with correct pharmacokinetic

and pharmacodynamic properties, cells should be capable of performing adequate (human-like)

post-translational processing. Lastly, cells should be safe for human use, not allowing the

propagation of any adventitious pathogenic agents that may eventually find their way into humans

(Ozturk and Hu 2006; Jayapal et al. 2007).

Considering these characteristics, a diversity of expression systems have been considered for

the production of therapeutic proteins, including bacteria (Waegeman and Mey 2012), yeast

(Cereghino and Cregg 2000; Schumann and Ferreira 2004), insect (Kost, Condreay, and Jarvis


12

2005; Kirkpatrick et al. 1995), plant (Hellwig et al. 2004; Soderquist and Lee 2005) and

mammalian (Dingermann 2008; Durocher and Butler 2009) cells, each system possessing

different advantages and limitations, as resumed in Table 1.2.

TABLE 1.2. Advantages and limitations of different expression systems for therapeutic glycoprotein production (Dasgupta

et al. 2007; Durocher and Butler 2009; Zhu 2011; Altmann et al. 1999; Harcum 2006; Hellwig et al. 2004)

EXPRESSION SYSTEM MARKETED PRODUCTS MAIN ADVANTAGES (+) AND LIMITATIONS (-)

BACTERIA

Insulin, human growth

hormone

+ Simple and fast production, low costs, high productivity

- Unable to perform glycosylation

YEAST

Insulin, human growth

hormone, hirudin, albumin

+ Robust expression, scalable fermentation, high titers

- Glycosylation potentially immunogenic to humans

INSECT CELLS None + Low costs, simple and fast production, safety


PLANT CELLS Glucocerebrosidase (2012

approval)

+ Inexpensive and simple culture, intrinsically safe

(neither harbor pathogens nor produce endotoxins)


MAMMALIAN CELLS Many, including interferon,

mAbs, erythropoietin, and

tissue plasminogen activator

+ Human-like glycosylation

- High cost of goods, potential to propagate infectious

agents, long development time

At the moment, biopharmaceutical manufacturing is mostly relying on mammalian cell

expression systems, as indicated by the preeminence of therapeutics approved by the Food and

Drug Administration (FDA) produced in these cells (Zhu 2011; Walsh 2010; Wurm 2004;

Andersen and Reilly 2004). The main motivator of this preference is the ability of mammalian

cells to synthesize proteins that are similar to those naturally occurring in humans with respect to

structure and biochemical properties (Zhu 2011).

Among mammalian systems, the Chinese hamster ovary (CHO) cells are undoubtedly the

dominant choice, accounting for nearly 70 % of all recombinant protein therapeutics currently

produced (Jayapal et al. 2007; Walsh 2010), which include four of the top-five selling

biopharmaceuticals in 2010 (Enbrel, Humira, Avastin and Rituxan) (Huggett, Hodgson, and

Lahteenmaki 2011). The extensive use of CHO cells is related to the compliance with most of the


13

requirements mentioned above, which include ease of transfection, presence of a powerful gene

amplification system, ease of adaptation to growth in suspension and serum-free medium, ability

to grow at high densities, ability to perform human-like post-translational modifications, and

safety of the product (Wurm 2004; Butler 2004; Trill, Shatzman, and Ganquly 1995; Mohan et

al. 2008). Additionally, the knowledge and expertise accumulated during two decades of

commercial production with CHO cells will ease the process of FDA approval for production of

new CHO-based therapeutics (Jayapal et al. 2007; Durocher and Butler 2009).

Other mammalian cells have also been used, although to a lower extent, including murine

myeloma (NS0 and Sp2/0) and baby hamster kidney (BHK-21) cells (Birch and Racher 2006;

Zhang 2010; Walsh 2010). Similarly to CHO, these cells are able to grow at high densities, in

suspension and serum-free culture, and are readily transfected (Trill, Shatzman, and Ganquly

1995; Butler 2004).

More recently, human cells have become quite attractive to the biopharmaceutical industry,

mainly due to their ability to produce genuine human post-translational modifications (Swiech,

Picanço-Castro, and Covas 2012; Niklas et al. 2011; Schräder et al. 2011). Some cell lines (e.g.

human embryonic kidney (HEK) and human embryonic retinal (PER.C6) cells) have shown

efficient production at laboratory scale, but their application in commercial biopharmaceutical

production is still on its very beginning and is therefore still limited (Swiech, Picanço-Castro, and

Covas 2012; Schiedner et al. 2008; Zhu 2011). In particular, the human retina-derived PER.C6

cell line has shown interesting characteristics for therapeutic protein expression, particularly high

production titers and rapid development of high-producing stable clones (in a few months) due to

the absence of requirements for gene amplification or selective marker (Wurm et al. 1992; Jones

et al. 2003; Zhu 2011).

1.4. TRANSFECTION

Transfection is the process that introduces foreign nucleic acids into cells to genetically

modify them (Kim and Eberwine 2010). In biopharmaceutical manufacturing, this process is

applied to introduce the gene of the desired therapeutic protein into the selected mammalian


14

host cell to obtain producer cells. Two types of transfections, transient and stable (Figure 1.3),

are routinely performed in mammalian systems, depending on the nature of the genetic materials

(Kim and Eberwine 2010; Kingston 2003).

FIGURE 1.3. Schematic diagram of (a) transient transfection and (b) stable transfection of foreign genetic material

(DNA, grey wave) into mammalian cells. In transient transfection (a) the foreign DNA enters the nucleus and is

expressed only temporarily due to a rapid loss of the foreign genetic material by environmental factors and cell

division. In stable transfection (b) the foreign DNA enters the nucleus and is integrated into the genome of the host

cell, allowing its stable expression over extended periods of time.

Transient transfection relies on the introduction of the foreign genetic material into the

nucleus of the host cell where it is replicated as an extrachromosomal unit (Ozturk and Hu 2006;

Recillas-Targa 2006; Ma and Chen 2005). This results in a fast process of transfection, but with

temporary levels of production due to the rapid loss of the genetic material by environmental

factors and cell division (Kim and Eberwine 2010; Ozturk and Hu 2006; Liao and Sunstrom

2006). In stable transfection, the foreign genetic material is introduced into the nucleus and

integrated into the genome of the host cell, which permits sustaining the gene expression over

extended periods of time (Glover, Lipps, and Jans 2005; Ozturk and Hu 2006; Ma and Chen

2005). Approximately one in ten thousand cells in a transfection will stably integrate the nucleic

acid (Mortensen and Kingston 2009); therefore, selectable markers can be used to enable the

identification and selection of these stably transfected cells by the characteristics provided by the

(a)

Nucleus

Expression

Transfection

DNA

(b)

Nucleus

Expression

Transfection

Integration

DNA


15

marker. This form of transfection ultimately delivers larger quantities of protein from cells (Rosser

et al. 2005), but it is labor-intensive, time-consuming, and requires considerable investment on

resources/equipment (Ozturk and Hu 2006; Rosser et al. 2005).

The choice between stable or transient transfection depends on the objective of the

experiment. For clinical or commercial (large-scale) purposes stable transfection is usually the

preferred method (Wurm 2004), while transient transfection is mostly valuable for high-

throughput screening in drug discovery processes, in vivo evaluation and early product analysis

(Liao and Sunstrom 2006; Ozturk and Hu 2006; Rosser et al. 2005).

1.4.1. EXPRESSION VECTOR

The expression of heterologous proteins in mammalian cells requires the use of specialized

vectors to transfer the gene of the product into the cells (Özdemir 1998). These vectors should

display three main features: expression levels independent from the site of integration in the

genome, expression levels correlated to the number of integrated transgene copies, and

maintenance of expression efficiency over time (Blaas et al. 2009). As a consequence of these

requirements, mammalian vectors are usually plasmids, circular DNA molecules that exist in

bacterial cells apart from their main chromosome, and are able to replicate within cells

independently from bacterial or eukaryotic chromosomes (Özdemir 1998; Nehlsen et al. 2009).

For the expression of recombinant proteins in mammalian cells, plasmids should include the

following elements: (i) a promoter/enhancer that drives the expression of the coding regions of

the gene of interest, a critical element for efficient expression which typically consists of a strong

viral promoter/enhancer such as cytomegalovirus (CMV) and simian virus 40 (SV40) (Mortensen

and Kingston 2009; Carswell and Alwine 1989; Ozturk and Hu 2006; Liao and Sunstrom 2006);

(ii) sequences that stabilize or enhance translation, such as polyadenylation signals (to extend the

half-life of messenger ribonucleic acid (mRNA) in cytoplasm), Kozak sequence (to improve the

translational initiation of mRNA of the gene of interest), and intervening sequences (to improve

mRNA stability and export from the nucleus to the cytoplasm) (Jeffs 2007; Kozak 1986; Kozak

1995; Huang and Gorman 1990; Wurm 2004). Additionally, the generation of a stable producing

cell line requires the use of a sequence encoding a selectable marker gene, either on a separate


16

or on the same expression vector as the recombinant gene (Wurm 2004; Mortensen and

Kingston 2009).

The integration of the expression vector into the genome of the host mammalian cell occurs

randomly, being simple and straight forward but lacking reproducibility (Blaas et al. 2009).

Indeed, it has been shown that the site of integration influences the transcription rate of the

recombinant gene, a phenomenon known as the position effect (Wurm 2004; Eszterhas et al.

2002). Depending on the surrounding chromatin by the integration site, expression of the

product can be high, low or even null (Wurm 2004). Furthermore, the expression of the

recombinant gene tends to be inactivated (silenced) over time (Wurm 2004; Blaas et al. 2009).

Consequently, the selection of suitable high-producing clones becomes tedious and time

consuming (Wurm 2004). To overcome the position effects different strategies have been

developed (Kwaks et al. 2003; Huang et al. 2007; Wang, Yang, et al. 2008) and include the use

of anti-repressor elements flanking the vectors (e.g. expression augmenting sequence elements,

EASE) (Kwaks et al. 2003), the integration of vectors specifically into chromosomal loci with open

chromatin (highly transcribed) (Huang et al. 2007), and the use of nuclear regulatory factors

(matrix-attachment regions (MARS) or scaffold-attachment regions (SARS)) (Girod, Zahn-Zabal,

and Mermod 2005; Wang, Yang, et al. 2008).

The design of expression vectors presents additional issues when mAbs are concerned.

Indeed, the structure of heavy and light chain subunits, whose interaction influences the kinetics

of mAb assembly, (Li et al. 2007) results in the need to express these subunits at optimal

stoichiometric ratios (of equal extent) (Li et al. 2007; Schlatter et al. 2005). This has led to the

development of different strategies for vector design (Ozturk and Hu 2006), with the most

common consisting of the use of two vectors, one for each chain, for co-transfection into the cells

(Schlatter et al. 2005; Leitzgen, Knittler, and Haas 1997). Their relative expression is controlled

by the proportion of each gene used in the co-transfection cocktail. Nevertheless, this control is

not accurate (Fussenegger et al. 1999; Kaufmann and Fussenegger 2003) and this approach is

considered to be the least efficient to obtain a balanced expression of both chains (Wurm 2004).

Alternatively, a single vector using two identical promoters, one for each chain, can be used (Li et

al. 2007). Although a single vector system has the advantage of assuring equal introduction of

genes into cells, the randomness of their integration into the cell genome remains a problem that

results in inaccurate control of the relative expression of these genes (Yuansheng et al. 2009). A

more accurate control of the relative expression of multiple genes across a transfected cell pool


17

can be achieved with a single vector containing multiple compatible inducible promoters, with

each gene under the control of one independent promoter (Yahata et al. 2005; Kunes et al.

2009). This allows the introduction of an equal amount of different genes into each cell (Belshaw

et al. 1996), and also the simultaneous variation of the expression level of different genes by

changing the concentration of the corresponding chemical inducers (Belshaw et al. 1996;

Hartenbach et al. 2007; Mijakovic, Petranovic, and Jensen 2005).

1.4.2. SELECTABLE MARKERS FOR STABLE TRANSFECTION

As mentioned above, the difference between a transient and a stable transfection is the

incorporation of the nucleic acid into the genome of the host cell in the latter. To isolate the

stably transfected cells, a selection strategy must be implemented to provide the transfected cell

with a growth advantage in culture. For this, a selectable marker is transfected into the cells

along with the gene of interest (Mortensen and Kingston 2009; Wurm 2004). There are several

types of markers available for mammalian expression systems, which can be divided into

recessive and dominant classes (Shen et al. 2006). The markers providing recessive resistance

have specific requirements for the host cells, which must be deficient in the activity by which they

will be selected; whereas the genes conferring dominant resistance can be used independently of

the phenotype of the cell (Castillo 2008; Keown, Campbell, and Kucherlapati 1990; Shen et al.

2006). Some common selection markers are shown in Table 1.3.

The selection systems using a dominant marker are based on resistance to antibiotics or

detoxification of the effects of a protein synthesis inhibitor on the cell. A large number of drugs

are available for this purpose, including the most routinely used aminoglycoside

phosphotransferase (Aph), hydromycin-B phosphotransferase (Hph) and puromycin N-

acetyltransferase (Pac) (Castillo 2008; Shen et al. 2006). The antibiotic-resistance marker allows

the transfected cells to survive in a culture containing the corresponding antibiotic (G418,

hygromycin B and puromycin, respectively), while cells that do not have the marker stably

integrated into their genome perish after several weeks in culture (Mortensen and Kingston

2009).


18

TA

BLE

1.3

. Characteristics of selective markers commonly used in expression vectors for the stable transfection of mammalian cells (Li et al. 2010; Shen et al. 2006; Mortensen and

Kingston 2009; Castillo 2008)

SELECTABLE MARKER

REQUIREMENTS

SELECTIVE AGENT

INHIBITOR FOR

AMPLIFICATION

BASIS FOR SELECTION

DOMINANT

Aminoglycoside

phosphotransferase (Aph)

Geneticin (G418)/Neomycin

G418 blocks protein synthesis in mammalian cells. Gene expression results in

detoxification of G418.

Hygromycin-B phosphotransferase

(Hph)

Hygromycin-B

Hygromycin-B inhibits protein synthesis. Gene expression detoxifies hygromycin-B

by phosphorilation.

Puromycin N-acetyltransferase

(Pac)

Puromycin

Puromycin inhibits protein synthesis. The gene encodes an enzyme for the

inactivation of puromycin by acetylation.

Zeocin/Bleomycin resistance gene

(sh ble)

Zeocin/Bleomycin

The gene encodes a protein that stoichiometrically binds the zeocin/bleomycin and

inactivates it.

Histidinol dehydrogenase

(hisD)

Histidinol

Medium deficient in histidine and containing histidinol will not support cell growth.

The gene encodes an enzyme that detoxifies histidinol.

RECESSIVE

Thymidine kinase

(TK)

Cells deficient in TK activity

Media containing hypoxanthine,

aminopterin, thymidine (HAT)

TK is essential for cell growth in the presence of aminopterin, blocker of thymidine

synthesis. The gene allows cell survival in HAT media (contains aminopterin).

Dihydrofolate reductase

(DHFR)

Cells lacking DHFR activity (it can

be used in cells expressing DHFR

through the application of MTX)

Media lacking glycine,

hypoxanthine and thymidine

Methotrexate

(MTX)

DHFR is required for purine biosynthesis. The gene allows cell growth without

exogenous purines. Amplification is obtained by increasing MTX concentrations, an

inhibitor of DHFR.

Glutamine synthase

(GS)

Cells not expressing glutamine (it

can be used with cells expressing

GS though the application of MSX)

Glutamine-free media

Methionine

sulfoximine

(MSX)

GS is essential for glutamine formation. MSX inhibits endogenous GS activity, so

without glutamine supplemented in the medium, only cells transfected with GS are

able to survive.

Hypoxanthine phosphoribosyl

transferase (HPRT)

Cells lacking HPRT activity

Media containing HAT

HPRT is essential for purine synthesis via the normal salvage pathway. HAT blocks

the de novo purine synthesis. Cells expressing HPRT are able to survive through

the salvage pathway.


19

In selection systems relying on recessive markers, selection is based on restoring an

enzyme or protein that is critical for cell survival, and therefore requires host cells that are

deficient in that activity. In the non-transfected host cell, the requirement for the activity is

replaced by supplementing the media with the metabolic byproduct of the enzyme’s action. But

upon withdrawal of these media components, only cells that acquired the needed activity through

transfection are able to survive (Shen et al. 2006). Some of the recessive markers promote a

phenomenon called gene amplification, which consists of applying increasing concentrations of

the selection agent to stimulate cells to increase the number of copies of the marker gene

(Mortensen and Kingston 2009; Castillo 2008). The significant feature of gene amplification is

that regions of the chromosome which are adjacent to the marker gene are also amplified, hence

the gene of the product of interest that is co-transfected with the marker gene will also be

amplified and overexpressed (Mortensen and Kingston 2009). The two most commonly used

amplifiable systems are based on dihydrofolate reductase (DHFR) with methotrexate (MTX)

amplification in DHFR- CHO cells, and glutamine synthetase (GS) with methionine sulfoxamide

(MSX) amplification in both CHO-K1 and murine myeloma NS0 cells (Bebbington et al. 1992;

Jeffs 2007; Wurm 2004; Page and Sydenham 1991; Castillo 2008).

The DHFR expression system is based on the dhfr gene coding for the DHFR enzyme, which

is involved in nucleotide metabolism (Wurm 2004) by catalyzing the conversion of dihydrofolate

to tetrahydrofolate (Chusainow et al. 2009). In the absence of tetrahydrofolate, the primary

pathway for synthesis of purines and pyrimidines is inhibited, and the salvage pathway that

involves the conversion of hypoxanthine and thymidine can be used. Therefore, if cells deficient

in HPRT are supplied with glycine, hypoxanthine, and thimidine from an exogenous source, they

are able to survive. However, if the media is not supplemented with these precursors, the cells

will die unless they have acquired the ability to express DHFR through transfection (Shen et al.

2006; Wurm 2004). The process of amplification in this system is based on the use of MTX, a

potent inhibitor of the DHFR enzyme. Using increasing concentrations of MTX, transfected cells

containing the dhfr gene will have to increase their capacity for DHFR synthesis in order to

survive, by amplifying the dhfr gene (Chusainow et al. 2009; Andersen and Reilly 2004). The

gene of interest that is located in the same expression vector as the dhfr gene or adjacently

resides in the host chromosomal DNA will also be co-amplified (Kaufman 1990).

In the GS systemTM (Lonza), the glutamine metabolism in mammalian cells is explored.

Glutamine formation follows an enzymatic pathway of biosynthesis from glutamate and


20

ammonium which is dependent of the GS enzyme. Without glutamine in the growth medium, this

GS enzyme is essential for the survival of mammalian cells in culture (Wurm 2004; Shen et al.

2006). Since some cell lines do not express sufficient GS to survive, it is possible to use a

transfected GS gene as a selectable marker by permitting growth in a glutamine-free medium

(Barnes, Bentley, and Dickson 2000; Shen et al. 2006). Cell lines that express sufficient GS to

survive can also be used with the GS system through the application of MSX, an inhibitor of the

endogenous GS activity (Castillo 2008; Shen et al. 2006). Thus, only the transfectants with

additional GS activity are able to survive. It is also possible to perform gene amplification using

increasing levels of MSX, in a similar manner to MTX in the DHRF system. It has been shown

that the GS system has a time advantage over the DHFR system during development, and

requires fewer copies of the recombinant gene per cell (Brown et al. 1992), allowing a faster

selection of high-producing cell lines.

A more recently explored selectable marker consists of the HPRT gene. A system developed

at University of Edinburgh, known as OSCARTM system, relies on a series of partially disabled

minigene vectors that encode for hypoxanthine phosphoribosyltransferase (HPRT), essential for

purine synthesis via the normal cellular salvage pathway (Barnes, Bentley, and Dickson 2001).

HPRT-deficient mammalian cells transfected with one of these minigenes along with the gene of

interest are placed in a selective hypoxanthine aminopterin thymidine (HAT) medium that blocks

the de novo purine synthesis. Consequently, cell survival becomes dependent on the salvage

pathway using a disabled HPRT enzyme, and only the transfected cells are able to grow.

Furthermore, since large amounts of the disabled HPRT enzyme are required for cell survival,

gene amplification occurs. This implies that selection and amplification occur in a single step,

providing a time advantage of the OSCARTM system over both DHFR and GS systems which

require multiple rounds of amplification after selection (Melton et al. 2001).

1.4.3. DELIVERY METHODS

For both transient and stable transfection, the introduction of the gene of interest into the

mammalian cells can be accomplished by several delivery methods, whose ideal characteristics

include high transfection efficiency, low cell toxicity, minimal effects on normal cell physiology,

ease to use, and reproducibility (Kim and Eberwine 2010). The methods which are commonly


21

used today can be generally divided into three categories: biological, chemical and physical, as

shown in Table 1.4, and their choice is dependent on cell type, purpose intended, as well as

economic and technical factors (Ma and Chen 2005).

TABLE 1.4. Comparison of biological, chemical and physical methods for the delivery of genetic material into cells

BIOLOGICAL CHEMICAL PHYSICAL

CHARACTERISTIC Vírus-

mediated

Cationic

polymer

Cationic

lipid

Calcium

phosphate

Electroporation Microinjection Biolistic

Immunogenicity + - +/- - - - -

Cytotoxicity + + + + - - +/-

Limited insert size + - - - - - -

Safety risk + - - - - - -

Simple handling - + + + + - +

Expeditious - + + + + + +

Efficiency in adherent cells + +/- + +/- + + +

Efficiency in suspended

cells + - - - - - +

Viability High High High High Low High Low

Costs High Low Low Low Moderate High High

The biological method of virus-mediated transfection, also known as transduction, is highly

efficient and easy to achieve due to the viral nature of integration into the host genome,

characteristics that make it the most commonly used method in clinical research (Kim and

Eberwine 2010; Pfeifer and Verma 2001). However, the potential dangers of immunogenicity and

cytotoxicity associated with this form of transfection, particularly for applications in gene therapy

and manufacturing of therapeutics, has led to the development of safer and more adequate non-

viral approaches (Glover, Lipps, and Jans 2005; Wurm 2004; Cavazzana-Calvo, Thrasher, and

Mavilio 2004).

In contemporary research, chemical methods are the most widely used, relying on carrier

molecules to overcome the cell-membrane barrier (Kim and Eberwine 2010), and include

cationic polymer, calcium phosphate, and cationic lipid transfection (Holmen et al. 1995;

Schenborn and Goiffon 2000; Washbourne and McAllister 2002; Jordan, Schallhorn, and Wurm

1996; Derouazi et al. 2004; Tait et al. 2004). The principle of these methods consists on the

interaction between the negatively charged nucleic acids with the positively charged chemicals


22

(i.e. polymers and lipids), forming a positively-charged complex that is attracted to the negatively

charged cell membrane and is then incorporated into the cell apparently by endocytosis (Kim and

Eberwine 2010; Graham and van der Eb 1973; Wilson and Smith 1997; Ozturk and Hu 2006).

They have the advantages of simple performance, absence of mutagenesis, and no limitation on

the size of the packaged nucleic acid. However, the transfection efficiency is highly variable,

depending on factors such as the cell type, the relative concentrations of reagents and nucleic

acid, as well as environmental (pH, temperature) and cell membrane conditions (Ozturk and Hu

2006; Kim and Eberwine 2010; Gao and Huang 1995).

More recently, physical transfection methods have been developed to enable the direct

transfer of nucleic acids into the cell cytoplasm or nucleus by physical or mechanical means.

These methods include electroporation, microinjection, and biolistic particle delivery. Of these,

electroporation is the most widely used for being very simple and fast (Ozturk and Hu 2006;

Chusainow et al. 2009; Kim and Eberwine 2010). In this technique, a pulsed electric field

disrupts the voltage gradient across the plasma membrane, presumably creating reversible pores

that allow the entry of the nucleic acid into the cell (Canatella et al. 2001; Kingston 2003). This

method is very efficient and can be used with virtually any type of cell, but requires the

optimization of parameters like amplitude and length of pulse for each case (Ozturk and Hu

2006; Kingston 2003). Furthermore, due to a low cell viability after transfection, electroporation

requires more cells and nucleic acid than other transfection methods (Ozturk and Hu 2006;

Potter and Heller 2010).

In microinjection, the nucleic acid is directly injected into the cytoplasm or nucleus of the

cell, creating a highly efficient transfection, but with the limitations of a very time-consuming and

expensive procedure (Kim and Eberwine 2010) that requires specialized operator skills, and of

the inadequacy to the transfection of a large number of cells (Rose 2007).

In biolistic particle delivery, the nucleic acid adheres to biological inert particles (i.e. gold or

tungsten) forming conjugates that are shot into recipient cells at a high velocity (Kim and

Eberwine 2010; Ma and Chen 2005). This method is reliable but requires expensive instruments

and causes physical damage to the cells, therefore requiring the use of high cell numbers (Kim

and Eberwine 2010).


23

1.4.4. SELECTION

To generate stable producer cell lines, the transfected cells are subjected to a process of

screening to select those with the most desirable characteristics.

In a first stage of screening, the cell pool obtained after transfection is cultured under

conditions that allow the selective survival of the cells successfully transfected with the marker

gene (e.g. using medium lacking glycine, hypoxanthine and thymidine, and containing low levels

of MTX, for DHFR) (Chisholm 1995; Jayapal et al. 2007). The surviving cells, presumably

expressing the product of interest, are recovered and can be subjected to gene amplification, by

repetitive rounds of exposure to higher concentrations of inhibitors of selective markers (i.e. MTX

for DHFR) (Ozturk and Hu 2006; Li et al. 2005), which in most cases leads to the generation of a

pool of cells containing clones with enhanced expression levels (Ozturk and Hu 2006; Li et al.

2005). However, this pool is heterogeneous, composed of cells with different integration sites,

copy numbers, and varying specific productivities; therefore, a series of limiting dilutions are

performed in multi-well plates, to generate single colonies with uniform cell populations that can

be isolated (Jayapal et al. 2007; Li et al. 2010), and expanded to produce clonal populations

(Wurm 2004). Each population is then evaluated, usually in terms of growth and product titer,

and the highest producers are selected for additional round(s) of cultivation and analysis (Wurm

2004; Li et al. 2005). This analysis typically includes cell growth, cell specific productivity and

volumetric productivity (titer), but may or should also incorporate product quality attributes such

as the glycosylation profile (Li et al. 2010) and other important characteristics specific to the type

of process to be used, as for example the ability to grow in serum-free medium or the resistance

to the stress and shear conditions found in bioreactors (Lloyd et al. 2000; Böhm et al. 2004).

The cell line stability is another critical factor that should be considered during selection,

since volumetric and specific productivities may decline with cell age (Li et al. 2010), and the

scale-up of the culture process to industrial levels takes significant time. Usually, cells maintained

in the presence of selective agents exhibit stable production during many months in culture

(Shen et al. 2006). However, due to the toxicity, high costs and complicated downstream

purification associated with these agents, it would be important to guaranty stability of production

without depending on their use (Barnes, Bentley, and Dickson 2001; Böhm et al. 2004; Li et al.

2005; Wurm 2004). Studies have shown that in the absence of selective pressure there are signs

of instability, with decreasing levels of production, which is often attributed to a loss of


24

recombinant gene copies during long-term culture (Page and Sydenham 1991; Kim, Kim, and

Lee 1998; Brown et al. 1992). This is particularly critical for cells in continuous production

cultures that run over long periods in the absence of the selective agent, so care should be taken

to screen the cell line for stability or to devise expression strategies that minimize the need for

amplification (Shen et al. 2006).

The screening and selection of a highly productive and stable clone in a short time frame is

still currently a major challenge because the desired characteristics are often dependent on cell

culture conditions (Li et al. 2005; Li et al. 2010). In this sense, the use of miniaturized high

throughput bioreactors that allow full control of process parameters could help identify the best

production clone at a very early stage, by mimicking the intended final large-scale bioreactor

environment (e.g. feeding regime, final production medium) (Li et al. 2005; Li et al. 2010;

Jayapal et al. 2007).

1.5. PROCESS OPTIMIZATION AT SMALL SCALE

After transfection of the host cells with the gene of interest and selection of the clones with

the most desirable characteristics, the next step on biopharmaceutical process development is

the adaptation of cells to grow in the culture environment most suitable for production at large-

scale. This transition on cultivation conditions may impart drastic changes in cellular physiology

and consequently alter cell growth, product yield and product quality. Therefore, it is also an

established practice to monitor and optimize different culture parameters to improve these

characteristics, while simultaneously searching for ways to reduce the costs associated.

1.5.1. MODE OF GROWTH

The production of therapeutic proteins usually requires large-scale processes to satisfy the

market demands. In these processes, cells are commonly grown in suspension due to several

advantages provided over the adherent mode of growth. Of these, perhaps the major advantage


25

is that suspension cultures are not limited by the surface available to adhesion, therefore offering

a higher surface-to-volume ratio that results in increased cell densities and productivities (Ozturk

and Hu 2006). Additional benefits related to simpler scale-up and process control support the

current preference for suspension culture (Chu and Robinson 2001).

However, since the majority of mammalian cells are attachment-dependent, a period of

adaptation is required to obtain cells growing is suspension (Kwaks et al. 2003). This is usually

accomplished at low scale, by culturing the cells in adequate vessels (e.g. spinner flasks, shake

flasks or roller bottlers) and under rocking conditions (Ozturk and Hu 2006; Zahn-Zabal et al.

2001; Berson and Friederichs 2008). Some critical parameters must be considered to guarantee

a successful adaptation, such as initiating the culture with a suitable high cell density and

optimizing the rocking conditions according to the capacities of each cell type to avoid death

caused by shear (Shen et al. 2006). In fact, the sensitivity of mammalian cells to shear stresses

is one of the main problems related to suspension cultures. It limits the rocking velocities that

can be applied in the processes of production and, consequently, has strong implications on the

oxygen and nutrient distribution in large bioreactors, ultimately affecting the final product yield

(Zhu, Mollet, and Hubert 2007). Mostly for this reason, the use of adherent culture for large-scale

production has not been completely put aside, with several developments made in the field.

These have mainly focused on maximizing the surface-to-volume ratio of adherent culture, with

the most successful system currently available being microcarrier culture (Ziao et al. 2002;

Tharmalingam et al. 2011; Knibbs et al. 2003). Microcarriers are small beads that may be solid

(microporous microcarriers) or porous (macroporous microcarriers), which allow growth of cells

in their external and internal (only for macroporous) surface and are maintained in suspension in

the culture medium (Wurm 2004; Blüml 2007; Ozturk and Hu 2006). Therefore, this system

provides a pseudo-suspension culture for anchorage-dependent cells with increased surface area

available for cell adhesion and growth, combining the advantages of both adherent and

suspension cultures (Blüml 2007; Hirtenstein et al. 1980; Rudolph et al. 2008). Among them are

improved cell densities and productivities, protection against shear stress, easy scale-up, and

simplified downstream processing due to cell retention (Blüml 2007; Ozturk and Hu 2006;

Tharmalingam et al. 2011).


26

1.5.2. MEDIUM

The supplementation of basal culture media with animal serum is essential for cell growth

and proliferation, due to its rich composition of hormones, growth factors, lipids, and trace

elements (Brunner et al. 2010). It also offers protection against potentially adverse conditions

such as pH fluctuations or shear forces, consequence of its high albumin content (Butler 2005).

However, the use of animal serum in cell culture bears a number of disadvantages for the

production of therapeutics. These include the high lot-to-lot component variability that jeopardizes

process consistency, the burden put on downstream processing due to increased protein

content, the high costs, and the risk of transmission of diseases from animals to humans

(Brunner et al. 2010; LeFloch et al. 2006; Zhang and Robinson 2005; Liu and Chang 2006).

Consequently, the removal of serum and all animal-derived components from the culture

medium is now a required step in process development for therapeutic protein production, for

both economical and safety reasons (Butler 2005; Froud 1999; Ozturk and Hu 2006; LeFloch et

al. 2006). This poses a considerable challenge for media development since it has been found

that the producer cell lines are quite fastidious in their growth requirements and these

requirements vary considerably among cell lines. Indeed, it has not been possible to develop a

single serum-free formulation to act as a serum substitute suitable for the growth of all cell lines,

like serum does. Therefore, serum-free and animal-component-free formulations are developed

according to the needs of each specific producer cell line (Butler 2005).

The adaptation of the producer cell line to these serum-free media is technologically more

challenging than the adaptation of cells to suspension growth in serum-containing media (Shen et

al. 2006). Basically two approaches are commonly used: direct adaptation, with the complete

removal of serum in a single step; or the more common gradual adaptation, by the sequential

reduction of serum concentration in the medium (Doyle and Griffiths 1998; Sinacore, Drapeau,

and Adamson 2000). Both approaches represent a major change in the cell environment,

consequently affecting cell growth, viability, and potentially product yield and quality (Shen et al.

2006; LeFloch et al. 2006; Ozturk et al. 2003). Extensive investigations to optimize medium

composition for a specific cell line are required to recover equivalent performances (Castro et al.

1992; Liu, Chu, and Hwang 2001; Liu and Chang 2006).


27

1.5.3. PROCESS OPTIMIZATION

After cell line generation, clone selection, and cell adaptation, a stage of process

optimization is required before establishing the final process of production at large scale. This

typically involves the control and evaluation of culture parameters that influence the cell

environment, aiming to achieve a maximum product yield while maintaining acceptable product

quality profiles (Li et al. 2010; Woolley and Al-Rubeai 2009; Jain and Kumar 2008). Common

parameters include physical factors such as temperature, gas flow rate and agitation speed, and

chemical factors like dissolved oxygen (DO), carbon dioxide (CO2), pH, osmolality, and nutrient

and metabolite levels (Li et al. 2010; Jain and Kumar 2008). For example, it has been shown

that a reduction of the typical cell culture temperature (37 ºC) causes a decrease in cell growth

but may improve the specific productivity and product quality (Stiens et al. 2000; Bollati-Fogolín

et al. 2005; Yoon et al. 2005; Mohan et al. 2008; Fox et al. 2004; Oguchi et al. 2006). DO has

also been shown to influence product quality although cell growth and production levels are kept

stable within a large range (Kunkel et al. 1998; Li et al. 2005; Griffiths 2000). The nutrient and

metabolite concentrations are another limiting factor in cell culture, particularly glucose,

glutamine, lactate and ammonia levels, and have led to the development of different feeding

strategies to reduce their negative effects on the course of the culture (Kontoravdi et al. 2007;

Lao and Toth 1997; Europa et al. 2000; Altamirano et al. 2004).

Another important factor in process development is the selection of an adequate mode of

operation of the bioreactor. The most commonly used are batch, fed-batch and perfusion, each

having their advantages and limitations. Batch culture is the simplest process, consisting of an

initial inoculation of the bioreactor without any further additions or withdrawals performed during

the culture (Pierce and Shabram 2004; Noe et al. 1992; Fenge and Lüllau 2006). Consequently,

the nutrient concentration decreases over time while the concentration of waste products

increases, leading to a gradual deterioration of the cell environment that leads to reduced cell

concentrations and negative impacts on both product yield and quality (Fenge and Lüllau 2006;

Wong et al. 2010). To overcome the limitations of batch, a controlled nutrient feeding may be

performed during the course of the culture, resulting in a fed-batch mode of operation, with

increased culture longevity, cell yield and overall productivity (Fenge and Lüllau 2006; Sandadi,

Ensari, and Kearns 2005; Dorka et al. 2009; Huang et al. 2010). However, the accumulation of

waste products still occurs and may impact the product quality (Wong et al. 2010). Therefore, to


28

provide a continuous renewal of medium with the introduction of fresh nutrients and removal of

metabolites, a perfusion process can be used (Fenge and Lüllau 2006; Voisard et al. 2003;

Adams et al. 2011). This mode of culture provides the highest cell concentrations and product

yield, and due to the reduced residence time of the product in the culture environment, the

product quality is also generally improved (Yang et al. 2000; Voisard et al. 2003; Ryll et al. 2000;

Lipscomb et al. 2005).

1.6. GLYCOSYLATION

Glycosylation is the most prevalent and structurally complex post-translational modification

naturally occurring in proteins, with more than half of the human proteins and more than one-

third of approved biopharmaceuticals being glycosylated (Wong 2005; Walsh and Jefferis 2006;

Solá and Griebenow 2010). It is a complex process that occurs in the endoplasmic reticulum

(ER) and in the Golgi apparatus, involving more than a hundred different enzymes that define the

glycan composition of the protein (Dingermann 2008; Li and d’Anjou 2009). This glycan moiety

is often critical in determining the pharmacological properties of therapeutic glycoproteins,

including conformation, stability, solubility, pharmacokinetics, in vivo activity, and immunogenicity

(Li and d’Anjou 2009; Helenius and Aebi 2001).

However, glycosylation is naturally highly variable and is also dependent of cell type- and

species-specific factors (i.e. levels of specific enzymes and availability of appropriate

monosaccharides) as well as of local environmental factors (Serrato et al. 2004; Yoon et al.

2005; Trummer et al. 2006; Crowell et al. 2007). Therefore, the choice of the expression system

(host cell) and of the culture conditions for glycoprotein production have a strong impact on the

glycosylation profile and consequently on the functionality of the protein (Royle et al. 2007). As a

result, this modification has become a major concern in biopharmaceutical production, which

now regards the quantitative analysis and monitoring of glycosylation as a means to control the

efficacy and safety of therapeutics (Hossler 2012; Hossler, Khattak, and Li 2009; Li and d’Anjou

2009).


29

1.6.1. BASIC CONCEPTS OF GLYCOSYLATION

The tremendous diversity observed in the process of glycosylation begins with the natural

diversity of its most simple component, the monosaccharides, both in chemical structure and

connectivity. Concerning chemical structure, monosaccharides can be divided into two general

classes according to their carbonyl group: the aldoses (aldehyde group, -CH=O) and the ketoses

(ketone group, C=O). The monosaccharides of both classes (with the exception of triketose) have

at least one asymmetric carbon atom (stereocentre), which allows them to have different possible

isomeric forms (stereoisomers, same molecular formula but different three-dimensional

orientation), generating variability. An absolute configuration, D or L, is attributed to each

monosaccharide considering the configuration of the stereocentre which is further from the

anomeric center (the carbon in the carbonyl group) (D on the right and L on the left, in the Fisher

projection) (Bertozzi and Rabuka 2009). Furthermore, the monosaccharides can exist in an open

form or undergo a cyclization reaction that yields a cyclic form and produces two possible

isomers of the anomeric carbon (α and β) (Bertozzi and Rabuka 2009; Marino et al. 2010).

Regarding the chemical connectivity, the cyclic monosaccharides establish connections

between the anomeric carbon of one monosaccharide and the hydroxyl group of another to form

higher-order structures (glycans). The linkages established have a high variability due to the α

and β isomers of the anomeric carbon and the diverse hydroxyl groups that are available for

connection in the monosaccharides. Furthermore, the presence of different hydroxyl groups also

enables the involvement of the same monosaccharide in more than two linkages, which results in

the formation of branches. This branching phenomenon is unique to glycans and creates very

different three-dimensional structures that contribute to their wide diversity, and also influences

its compactness, flexibility and diverse physical and biochemical properties (Bertozzi and Rabuka

2009; Marino et al. 2010; Marth 1999; Brooks 2009).

The glycans formed possess a reducing end that bears a free anomeric center that may

establish a bond to other structures, such as a polypeptide, by the process of glycosylation

(Bertozzi and Rabuka 2009; Marth 1999). The glycans associated with glycoproteins are typically

composed of up to a dozen monosaccharides, and may be divided into two general types, as

defined by the linkage established with the protein: N-glycans, which attach to the nitrogen atom

of asparagine side chains; and O-glycans, which bind to the oxygen atom of serine or threonine

side chains (Bertozzi and Rabuka 2009; Marino et al. 2010; Wopereis et al. 2006). Of these, N-


30

glycans are the most prevalent and widely studied, and are the main focus of this chapter (Bañó-

Polo et al. 2011; Lin et al. 2012).

In animals, the glycans attached to proteins mostly contain ten types of monosaccharides,

with seven being more prevalent in humans (Brooks 2009; Bertozzi and Rabuka 2009). These

are represented in Table 1.5 with the abbreviation and symbolic notation adopted in this work.

TABLE 1.5. Monosaccharides commonly present in mammalian glycoproteins, with respective abbreviation and

symbolic notation (Harcum 2006; Merry et al. 2004; Brooks 2009; Bertozzi and Rabuka 2009)

MONOSACCHARIDE ABBREVIATION SYMBOLIC NOTATION

Neutral Glucose Glc

Galactose Gal

Fucose Fuc

Mannose Man

Xylose Xyl

Amino N-acetylglucosamine GlcNAc

N-acetylgalactosamine GalNAc

Acidic N-acetylneuraminic acid NeuNAc

N-glycolylneuraminic acid NeuGc

Glucoronic acid GlcA

1.6.2. BIOSYNTHESIS OF O-GLYCANS IN HUMAN CELLS

O-linked protein glycosylation is much less well understood than N-linked glycosylation due

to the existence of few inhibitors of O-linked glycosylation compared to the range of inhibitors of

specific steps of N-glycosylation that has facilitated its study in detail (Brooks, Dwek, and

Schumacher 2002; Elbein 1987, 1991). Nevertheless, there are seven types of O-linked glycans

found in humans, classified on the basis of the first sugar attached to a serine (Ser) or threonine

(Thr) residue of a protein: O-linked N-acetylgalactosamine (GalNAc) or mucin-type; O-linked

glycosaminoglycan (GAG); O-linked N-acetylglucosamine (GlcNAc); O-linked galactose (Gal); O-

linked mannose (Man); O-linked glucose (Glc); and O-linked fucose (Fuc) (Wopereis et al. 2006;


31

Haltiwanger 2004). Of these, the mucin-type O-glycan, with GalNAc at the reducing end, is the

most common form found in humans, occurring in over 10 % of human proteins (Roth, Yehezkel,

and Khalaila 2012; Wopereis et al. 2006).

This post-translational modification is normally initiated in the Golgi apparatus by the

transfer of a GalNAc monosaccharide from the nucleotide uridine diphosphate (UDP)-GalNAc into

a Ser or Thr residue of a fully folded and assembled protein, through the action of a N-acetyl

galactosaminyltransferase (GalNAc transferase) (Butler 2004; Van den Steen et al. 1998). No

general consensus sequence has yet been identified for O-glycosylation, but the glycosylated Ser

or Thr residues are often flanked by proline (Pro) (Brooks 2004; Spiro 2002). Subsequently, a

sialic acid residue may be added, terminating the chain, or a stepwise enzymatic elongation by

specific glycosyltransferases can occur yielding different long linear or branching chains that

compose the core structures (Van den Steen et al. 1998; Brooks 2004; Spiro 2002). There are

eight core structures currently identified, as shown in Figure 1.4, which can be distinguished by

the second monosaccharide and respective linkage (Wopereis et al. 2006).

FIGURE 1.4. Representation of the eight known core structures of mucin-type O-glycans. Core 1 to 4 are the most

prevalent; Core 7 has not been observed in humans. Core 1 has a Gal attached in a β1-3 linkage to the core

GalNAc. Core 2 forms by the addition of a GlcNAc β1-6 linked to the GalNAc of Core 1. Core 3 has a GlcNAc

attached in a β1-3 linkage to the core GalNAc. Core 4 forms by the addition of another GlcNAc β1-6 linked to the

GalNAc of Core 3. Core 5 has a GalNAc attached in a α1-3 linkage to the core GalNAc, Core 6 has a GlcNAc in a β1-

6 linkage, Core 7 has a GalNAc in a β1-6 linkage, and Core 8 has a Gal α1-3 linked to the core GalNAc.

CORE 1 CORE 2 CORE 3 CORE 4

CORE 5 CORE 6 CORE 7 CORE 8


32

The Core 1 structure is formed by the addition of Gal in a β1-3 linkage to GalNAc by the

core 1 β1-3 galactosyltransferase. Core 2 O-glycans are branched Core 1 structures, synthesized

by the addition of a GlcNAc in a β1-6 linkage to the GalNAc of the Core 1 structure through the

action of one of three core 2 β1-6 N-acetylglucosaminyltransferases (C2GnTs) (Brockhausen,

Schachter, and Stanley 2009; Tarp and Clausen 2008).

Alternatively to Core 1 O-glycan formation, Core 3 may be formed through the addition of

GlcNAc in a β1-3 linkage to the GalNAc attached to Ser/Thr by the enzyme Core 3 β1-3 GnT

(C3GnT). The synthesis of Core 3 O-glycans is required for the synthesis of Core 4 by the M-type

β1-6 GnT (C2GnT-2), which adds another GlcNAc in a β1-6 linkage to the GalNAc attached to

Ser/Thr. In addition, Cores 5 to 8 O-glycans exist (Core 7 has not been found in humans) but are

rare and the enzymes responsible for their synthesis remain to be characterized (Brockhausen,

Schachter, and Stanley 2009; Tarp and Clausen 2008).

The core structures may be further elongated or modified by sialylation, fucosylation,

sulfatation, methylation or acetylation, given rise to a multitude of O-glycan structures (Van den

Steen et al. 1998; Butler 2004). The final O-linked glycans are smaller and less complex than

typical N-linked glycans, usually containing 1 to 20 monosaccharide residues (Harcum 2006).

1.6.3. BIOSYNTHESIS OF N-GLYCANS IN HUMAN CELLS

The biosynthesis of glycans occurs in the ER and Golgi apparatus and involves the intricate

and concerted activity of a multitude of specific transmembrane enzymes, namely

glycosyltransferases and glycosidases (van den Eijnden and Joziasse 1993; Sears and Wong

1998; Burda and Aebi 1999).

SYNTHESIS OF THE DOLICHOL GLYCAN PRECURSOR AND TRANSFER TO NASCENT PROTEINS

N-linked glycosylation begins with the synthesis of a dolichol glycan precursor which consists

of the lipid dolichol phosphate (Dol-P) linked to a glycan composed of 14 specific

monosaccharide linkages (Glc3Man9GlcNAc2), a structure that has been conserved among all

eukaryotes (Brooks 2004; Marth 1999; Taylor and Drickamer 2006). The assembly of this

precursor is outlined in Figure 1.5. The process begins on the cytosolic side of the ER


33

membrane, with the linkage of two core GlcNAc and five Man monosaccharides donated by UDP-

GlcNAc and guanosine diphosphate (GDP)-Man, respectively (Marth 1999; Butler 2004). Then,

by a mechanism not fully understood (Helenius and Aebi 2002), the precursor formed

(Man5GlcNAc2-Dol) flips across the membrane bilayer into the lumen of the ER, where four Man

and three Glc residues are added using the Dol-P-Man and Dol-P-Glc donors, completing the

assembly of the dolichol glycan precursor Glc3Man9GlcNAc2 (Marth 1999; Butler 2004).

The dolichol glycan precursor is transferred to an asparagine (Asn) residue on a nascently

translated protein, through the action of a multi subunit transmembrane protein complex termed

oligosaccharyltransferase (OST), present in the ER membrane, which cleaves the high-energy

GlcNAc-P bond, causing the release of Dol-P-P in the process (Figure 1.5) (Butters 2002;

Dempski Jr and Imperiali 2002; Marth 1999).

FIGURE 1.5. Pathway for the biosynthesis of the dolichol glycan precursor for protein N-glycosylation and transfer of

the precursor to asparagine residues (Asn-X-Thr/Ser) on nascently translated proteins by the

oligosaccharyltransferase (OST) complex. Dol-P: dolichol phosphate; GDP: guanosine diphosphate; UDP: uridine

diphosphate; Man:mannose; Glc: glucose; ER: endoplasmic reticulum.

Dol-PUDP-

GlcNAcGDP-Man

Dol-P-Man

Dol-P-Glc

Dol-P

GDP-Man

Dol-P-Man

Dol-P-Glc

OST OST OST

Ribosome RibosomeRibosome

ER LUMEN

CYTOSOL

CYTOSOL

UDP-Glc

Asn-X-Thr/Ser


34

A consensus sequence is required for the attachment of the glycan to the acceptor Asn. This

Asn must have a minimal surrounding sequence of Asn-X-Thr/Ser, where X can be any amino

acid except Pro (Brooks 2004). However, among the polypeptides bearing this consensus

sequence (potential N-glycosylation sites) only about 30 % appear to be used (Marth 1999).

EARLY PROCESSING STEPS IN THE ENDOPLASMIC RETICULUM

Following the attachment of the precursor to the Asn residues, a series of processing

reactions occurs (Figure 1.6).

FIGURE 1.6. Processing of the initial high mannose N-glycan in the endoplasmic reticulum and cis-Golgi to generate

the core N-glycan substrate used for further diversification in the Golgi.

The first several steps occur in the lumen of the ER and appear to be conserved among all

eukaryotic cells. These steps are associated with quality control of correct protein folding and are

essential for subsequent transport of the newly synthesized glycoprotein to the Golgi apparatus

(Parodi 2000; Ellgaard and Helenius 2001; Marth 1999). The initial steps of trimming consist of

the sequential removal of the three Glc residues by the action of glucosidases I and II, to yield

Dol-P

Glucosidase I

Glucosidase II

Endo-mannosidase

Mannosidase IA

Mannosidase IB

Oligosaccharyl-

transferase

ENDOPLASMIC RETICULUM

CIS-GOLGI


35

Man9GlcNAc2. This process of Glc removal is associated with protein folding mechanisms and

contributes to the ER retention time of the glycoprotein. Proteins that become improperly folded

are re-glucosylated by an α-glucosyltransferase and retained in the ER, where they are either

refolded into a proper conformation or deglucosylated and degraded (Marth 1999).

The correctly folded precursors are then acted on by an endo-mannosidase that generates a

Man8GlcNAc2 glycan structure than can proceed in the N-glycan processing and secretory

pathways (Marth 1999).

LATE PROCESSING STEPS IN THE GOLGI APPARATUS

The Man8GlcNAc2 structure produced in the ER is released to the cis-Golgi where it is

sequentially processed by distinct class I α-mannosidases (1A and 1B) that act specifically on

α1-2-linked Man residues (Figure 1.6). This processing most often results in a Man5GlcNAc2

glycan that becomes the core N-glycan substrate for diversification in the Golgi (Marth 1999;

Brooks 2004).

DIVERSIFICATION/MATURATION OF N-GLYCANS

The Man5GlcNAc2 structure can be trimmed and extended by the addition of other

monosaccharides or extended without further trimming, given rise to a diversification of N-glycans

in the Golgi (Brooks 2004). The structures of N-glycans created fall into three main categories:

high-mannose, hybrid, and complex types, which share a common trimannosyl core

(Man3GlcNAc2) but differ in their outer branches (Figure 1.7) (Marth 1999; Brooks 2004; Butler

2004).

The high-mannose type N-glycans may have between two and six additional Man residues

linked to the core; the complex type has no further Man residues and has two or more outer

branches containing GlcNAc (attached to the core), Gal and/or sialic acids; the hybrid type has

one outer branch of the high-mannose type and the other of the complex type (Marth 1999;

Brooks 2004; Butler 2004).

The enzyme GnT-I is the first involved in N-glycan diversification, adding a GlcNAc in β1-2

linkage to produce the hybrid N-glycan structure GlcNAc1Man5GlcNAc2. In the medial Golgi, α-

mannosidase II acts to remove the α1-3- and α1-6-linked Man residues as shown in Figure 1.7.


36

The resulting GlcNAc1Man3GlcNAc2 hybrid N-glycan is the specific substrate for the enzyme GnT-II

that catalyzes the conversion from hybrid to complex N-glycans (Marth 1999).

FIGURE 1.7. Representation of the N-glycan diversification in the Golgi, which generates three subtypes: high-

mannose, hybrid, and complex glycans.

The hybrid and complex N-glycans may have two or more GlcNAc-bearing branches which

are termed antennae. Branching reactions are very complex and may result in biantennary,

triantennary, and tetra-antennary glycans, as shown in Figure 1.8 (Harcum 2006).

FIGURE 1.8. Structures exemplifying complex N-glycans with varying number of antennae: biantennary, triantennary,

and tetra-antennary.

GlcNAc

transferase I

Mannosidase II GlcNAc

transferase II

Glactosyl-

transferase

Sialyl-

transferases

MEDIAL GOLGI

TRANS GOLGI

HYBRIDHIGH MANNOSE COMPLEX

BIANTENNARY TRIANTENNARY TETRA-ANTENNARY


37

The development of multi-antennary N-glycan structures requires the action, sometimes

competing, of four additional GnTs (GnT III to VI) to the already mentioned GnT-I and GnT-II

(Marth 1999). All GlcNAc residues except that added by GnT-III to the Man linked to the core

GlcNAc (termed bisecting GlcNAc) can be extended with additional monosaccharide linkages,

including Gal, sialic acid, and Fuc (Marth 1999). Galactosylation (the addition of Gal) is catalyzed

by a number of enzymes such as the β1-4 and β1-3 galactosyltransferases (GalTs). Sialylation

(the addition of sialic acids) may be catalyzed by different sialyltransferases specific for different

sialic acids and linkages, with the most common in humans being the linkage of N-

acetylneuraminic acid (NeuNAc) to the terminal Gal by either α2,3 sialyltransferase or α-2,6

sialyltransferase. Fuc is added by a fucosyltransferase in a α1-6 linkage to the GlcNAc residue

attached to Asn (core fucosylation). Fuc can be added at any time after the formation of the

Man5GlcNAc2 structure, but not after the action of GnT-III (bisecting GlcNAc) (Butler 2004; Marth

1999; Harcum 2006).

MACROHETEROGENEITY AND MICROHETEROGENEITY

Unlike the biosynthesis of a protein from mRNA, where the nucleic acid codes for only one

sequence, the attachment of a glycan to a protein is not template driven, and the presence of the

consensus sequence Asn-X-Ser/Thr does not guarantee glycosylation (Harcum 2006; Butler

2004). Additionally, a protein contains several potential glycosylation sites at different positions

that may remain unoccupied, adding variability and complexity to the process. This glycosylation

or non-glycosylation variability at a specific site in a protein is termed macroheterogeneity

(Harcum 2006). Influencing this macroheterogeneity may be factors such as the spatial

arrangement of the peptide during translation, which can expose or hide the consensus

sequence; the amino acids (X) surrounding the consensus sequence Asn-X-Ser/Thr, with high

occupancy level for Ser and Phenylalanine (Phe), intermediate for Leucine (Leu) and Glutamate

(Glu), and very low for Aspartate (Asp) and Tryptophan (Trp); and the availability of UDP and GDP

monosaccharide precursors (Butler 2004).

The generation of different N-glycan structures and their variable addition to the glycosylation

sites on a protein results in considerable glycosylation diversity, and is known has

microheterogeneity (Bill, Revers, and Wilson 1998; Royle et al. 2007; Harcum 2006).

Microheterogeneity is influenced by factors that include nucleotide metabolism, transport rates in


38

the ER and Golgi, and the localization and level of expression of the glycosyltransferases in the

Golgi apparatus (Walsh and Jefferis 2006).

Both macro- and microheterogeneity affect the biological activity of the glycoproteins

produced, which makes the prediction and/or control of their impact a concern for the

biopharmaceutical industry, for safety and regulatory reasons (Hossler, Khattak, and Li 2009).

1.6.4. GLYCOSYLATION AND THERAPEUTIC PROTEIN EFFICACY

Glycosylation often plays a critical role in establishing or maintaining the integrity of

glycoproteins. Indeed, the addition of glycans to a protein can result in a diverse set of changes

in the physicochemical and biological properties, which include stability, circulatory half-life,

efficacy, solubility and immunogenicity (Harcum 2006; Lowe and Marth 2003; Li and d’Anjou

2009; Van den Steen et al. 1998; Baker, Flatman, and Birch 2007).

The maintenance and stabilization of the physical properties of therapeutic proteins are

central for the storage and in vivo efficacy of the product, but may be hindered by many

physicochemical instabilities caused by inappropriate glycosylation (Yamane-Ohnuki and Satoh

2009). For example, the size and mass of the protein are affected by the number and type of

glycans attached, and the charge and solubility properties are influenced by the presence of

charged monosaccharides such as sialic acid residues (Lis and Sharon 1993; Narhi et al. 1991;

Reuter and Gabius 1999). Furthermore, the number, length, branching and charge of the glycans

have also been related to proteolytic stabilization, with the negative charge of sialic acids

improving the resistance to proteolysis and, consequently, enhancing the in vivo stability of the

protein (Solá and Griebenow 2009; Sareneva et al. 1993; Raju and Scallon 2007; Runkel et al.

1998). The attachment of glycans to some proteins also aids in folding (e.g. erythropoietin (EPO))

(Helenius and Aebi 2001; Narhi et al. 1991).

In addition to the physicochemical effects, glycosylation can also affect the efficacy of

therapeutic proteins by influencing their biological activity and clearance rate, especially through

the modulation of the pharmacokinetic properties. The pharmacokinetic profile of the protein

informs about the dosing size and frequency, which are key attributes for patient compliance and

acceptance of a drug (Li and d’Anjou 2009). Of the pharmacokinetic properties, the protein half-

life in circulation is one of the most important, and it has been shown to be improved by


39

glycosylation in different therapeutics (e.g. interferon (IFN), EPO) (Erbayraktar et al. 2003; Runkel

et al. 1998; Walsh and Jefferis 2006; Sareneva et al. 1996). In particular, the presence of sialic

acid residues can have a major influence in this property (Baker, Flatman, and Birch 2007; Zhu

2011; Egrie and Browne 2001), with increased levels of terminal sialylation improving protein

half-life apparently by ‘hiding’ the Gal residues that when exposed prompt a fast removal of the

protein from the circulation due to endocytosis-mediated uptake by Gal-specific receptors in the

hepatocytes (Kompella and Lee 1991; Baker, Flatman, and Birch 2007; Morell et al. 1971;

Wileman, Harding, and Stahl 1985; Raju 2003). Similarly, glycoproteins exposing glycans that

terminate in Man, GlcNAc or Fuc residues may also be more rapidly removed from the circulation

due to specific interactions with other lectin-like receptors expressed at different cell types

(Townsend and Stahl 1981; Weigel and Yik 2002; Raju 2003; Jones et al. 2007). This type of

clearance does not apply to all glycosylated therapeutics, as is the case of mAbs whose

clearance is mediated by other type of receptors (FcRn) that are not influenced by glycosylation

(Jefferis 2009). In fact, in the specific case of mAb-based therapeutics, glycosylation of the Fc

portion exerts a profound influence on the effector functions, and therefore on in vivo efficacy,

through the modulation of the binding affinities for other type of receptors, specifically the FcγR

receptors (Li and d’Anjou 2009; Jefferis 2009; Raju 2008; Crispin et al. 2009; Mimura et al.

2009). Of the glycan structures, Fuc residues are currently considered as the major influents on

mAb effector functions. Typically, the glycans at the Fc portion of mAbs contain a core Fuc

residue (Mizuochi et al. 1982); however, studies have shown that non-fucosylated variants exhibit

improved binding to FcγRIII receptors of up to 50-fold (Shields et al. 2002; Iida et al. 2006),

which results in dramatically enhanced ADCC (up to 100-fold) (Iida et al. 2006; Shields et al.

2002; Yamane-Ohnuki and Satoh 2009; Kanda et al. 2007). Furthermore, this improved affinity

for FcγRIII provides an advantage to the non-fucosylated therapeutic mAbs in the competition

with the human serum IgGs for binding to these receptors on NK cells (Vugmeyster and Howell

2004; Jefferis 2009; Preithner et al. 2006; Shinkawa et al. 2003).

The same enhanced ADCC can be accomplished by the attachment of bisecting GlcNAc, an

effect that has been attributed to the low Fuc content that is always associated with the presence

of this residue (since the bisecting GlcNAc inhibits further addition of Fuc during the N-glycan

biosynthesis) (Umana et al. 1999; Ferrara, Stuart, et al. 2006; Abès and Teillaud 2010).

Other variations of the Fc glycan structure have also been proven to alter the functional

effects of therapeutic mAbs, such as the levels of sialic acid, Gal and Man residues. In contrast to


40

other glycoproteins, where higher contents of sialic acid result in extended circulatory half-life and

improved efficacy, increased levels of sialylation have an adverse effect on mAb efficacy by

reducing ADCC activity (Scallon et al. 2007). It was proposed that these effects could be related

to a reduction of the flexibility of the mAb at the hinge region, which could have the following

effects: reduced binding to FcγRIIIa due to a greater difficulty in assuming the required sharp 90º

bend of the hinge region; or reduced bivalent (and increased monovalent) binding to the antigen

as a result of the mAb molecules not being able to orient themselves to enable both Fab arms to

simultaneously bind the antigen (Sondermann, Kaiser, and Jacob 2001; Scallon et al. 2007;

Kaneko, Nimmerjahn, and Ravetch 2006).

Contrarily to Fuc and sialic acid residues, the terminal content of Gal has no effects on

ADCC, but has shown to improve CDC as a result of increased antibody binding to C1q in

hypergalactosylated variants of the IgG (Hodoniczky, Zheng, and James 2005; Boyd, Lines, and

Patel 1995).

Additionally, the presence of high-mannose structures in mAbs, similarly to glycoproteins,

has been related to a rapid clearance from serum due to the binding of the exposed Man

residues to Man receptors expressed in phagocytic cells (Abès and Teillaud 2010; Wright et al.

2000).

1.6.5. EXTERNAL FACTORS INFLUENCING GLYCOSYLATION

The control and maintenance of a consistent glycosylation/glycoform profile during

glycoprotein manufacturing is a major concern and currently still a considerable challenge to the

biopharmaceutical industry (Walsh and Jefferis 2006). This is a consequence of the combination

of several factors. First, the high variability already inherent to the process of glycosylation

(macro- and microheterogeneity); second, the impact of glycosylation on protein functionality; and

third, the variability on the glycan profile introduced by several culture and process parameters,

whose mechanisms are still under investigation and therefore are hard to control. These

parameters include cell type- and species-specific factors (e.g. expression system) and

environmental factors and culture conditions (e.g. culture media, mode of culture) (Wong et al.

2005; Kunkel et al. 2000; Crowell et al. 2007; Gawlitzek et al. 1995; Hossler, Khattak, and Li

2009; Serrato et al. 2004; Trummer et al. 2006).


41

Therefore, to maintain product consistency and avoid unexpected changes during the

process of production of therapeutic glycoproteins it is crucial to understand how and at what

extent cell culture parameters affect glycosylation. As a more far-reaching objective, it may be

reasonable to control culture conditions in favor of reducing heterogeneity or even producing a

specific and more favorable glycoform (Butler 2004).

CELL TYPE- AND SPECIES-SPECIFIC FACTORS

Bacteria, yeast, insect, plant, and mammalian cells have evolved as the major recombinant

protein expression hosts. However, these expression systems vary widely in their ability to

perform human-like post-translational modifications such as glycosylation (Butler 2004;

Sethuraman and Stadheim 2006), so the choice of the most adequate host will depend on the

product and of the intended application. Some therapeutic proteins, such as insulin or human

growth hormone, do not need to be glycosylated for efficiency, so the ability to perform

glycosylation is not a critical parameter to be considered in host selection. However, most of the

therapeutics currently approved or in development are glycoproteins and, therefore, require

proper glycosylation for biological activity (e.g. tissue plasminogen activator (tPA), EPO, mAbs). In

this case, the choice of the expression system becomes crucial, since non-human glycoforms can

adversely impact the pharmacokinetic properties of the protein and raise immunogenicity and

safety concerns (James and Baker 2002; Sasaki et al. 1987; Beck et al. 2008; Jefferis 2005).

Indeed, the potential for significant changes in the biological activity of therapeutic proteins

depending on the expression system is a real concern from a pharmacological point-of-view,

particularly for mAbs because of their obvious complexity and their intended interaction with the

immune system of the patient (Beck et al. 2008; Sheeley, Merrill, and Taylor 1997).

The early events in the N-glycan biosynthesis pathway are largely conserved among

organisms. However, the latter steps of glycan diversification in the Golgi apparatus seem to have

been evolutionarily diversified (James and Baker 2002; Brooks 2004; Goochee et al. 1991).

These divergences can be related to differences found among species in the relative

expression/activity of several glycosyltransferases, which are a key factor in the synthesis of N-

glycans (Butler 2004; Butler 2005). In addition to the enzyme repertoire, other factors

contributing to the cell- and species-specific variability of the glycan profile are the competition

between different enzymes for one substrate, the transit time of the glycoproteins in the ER and


42

Golgi apparatus, the levels of sugar and nucleotide donors, and the competition between

glycosylation sites on the protein for the same pool of enzymes (Butler 2004).

An example of the N-glycan structures commonly obtained in humans and the different

expression systems available for therapeutic glycoprotein production can be seen in Figure 1.9.

FIGURE 1.9. Schematic representation of N-glycan structures found in human cells and in the different expression

systems available for recombinant protein production.

A more detailed list of the divergences found between glycan structures of different species

compared to human cells, and their respective effects on protein activity, is shown in Table 1.6.

TABLE 1.6. Divergences found on protein glycosylation occurring in different expression systems in comparison to the

glycosylation obtained in human cells, and respective effects on protein activity

EXPRESSION SYSTEM DIFFERENCES TO HUMAN EFECT

Bacteria No glycosylation Reduced or absent biological activity

Yeast High-mannose glycans Fast clearance, immunogenic

Insect High-mannose glycans

Paucimannose glycans

α1,3-fucose

Absence of sialylation

Fast clearance, immunogenic

Fast clearance, immunogenic

Immunogenic

Poor half-life

Plant α1,3-fucose

β1,2-xylose

Absence of sialylation

Absence of terminal Gal

Immunogenic

Immunogenic

Poor half-life

Mammalian Typically human-like

YEAST PLANT INSECT MAMMALIAN HUMAN


43

BACTERIA

Bacteria are generally regarded as unable to glycosylate proteins due to the absence of the

N-linked glycosylation machinery (Sethuraman and Stadheim 2006; Harcum 2006; Dingermann

2008). However, Eubacteria and Archaebacteria have shown ability to glycosylate surface

proteins (of the bacteria cell wall), with the biosynthetic pathways for this glycosylation still not

well characterized (Upreti, Kumar, and Shankar 2003; Schäffer, Graninger, and Messner 2001;

Benz and Schmidt 2002). Since the prokaryotic bacterial cells lack internal membrane-bound

organelles (including the ER and Golgi apparatus), glycan processing takes place on the outside

of the cell membrane and the glycans synthesized are mostly chemically very different from those

found on human cells (e.g. rhamnose, galactofuranose, N-acetylmannosamine). Nevertheless,

although possessing some glycosylation mechanisms, glycoproteins appear to be uncommonly

synthesized by bacteria, and the recombinant proteins currently produced in these organisms are

not glycosylated (e.g. insulin and human growth hormone produced in Escherichia coli) (Harcum

2006; Butler 2004; Brooks 2004).

YEAST

Yeast-based expression systems have been used for the production of some approved

therapeutic proteins (e.g. insulin, human growth hormone, hirudin, and albumin) (Sethuraman

and Stadheim 2006). However, the presence of yeast-specific glycans of the high-mannose type,

with up to 15 Man residues (hiper-mannosylation, Figure 1.9) (Marth 1999), in the proteins

produced is a concern since these groups are not common in humans and can be immunogenic

(Lam, Huang, and Levitz 2007; Dasgupta et al. 2007; Luong et al. 2007).

INSECT CELLS

Insect cells infected by baculovirus have been reported as efficient expression hosts for the

production of many glycoproteins, including mAbs (Altmann et al. 1999; Kost, Condreay, and

Jarvis 2005; Sethuraman and Stadheim 2006). They are capable of performing N-glycosylation

with similar core structures to other eukaryotic organisms, including humans (Harcum 2006;

Liang et al. 1997; Li et al. 2006). However, proteins expressed in insect cells also contain

glycans that are not of the complex type, such as hybrid, high-mannose and paucimannose

glycans (Figure 1.9) (Kim et al. 2005; Durocher and Butler 2009). Furthermore, in some insect

cell lines, non-human α1,3-Fuc may be added to the glycans, which raises immunogenicity and


44

safety concerns (Harcum 2006; Garcia-Casado et al. 1996; Altmann et al. 1999), and the

absence of sialic acid residues has also been detected (Altmann et al. 1999; Marchal et al.

2001). As a consequence of these modifications, which could compromise in vivo bioactivity and

potentially induce allergenic reactions, no approved therapeutic proteins are currently produced

in this system (Durocher and Butler 2009).

PLANT

Plant-based expression systems produce non-human glycan structures where Gal and sialic

acid residues are absent (Figure 1.9). Also, the presence of the potentially immunogenic α1,3-

Fuc and β1,2-xylose (β1,2-Xyl) residues are major limitations to the use of these systems for the

expression of therapeutic glycoproteins (Jin et al. 2008; Gomord and Faye 2004; Bardor et al.

2003).

MAMMALIAN CELLS

The N-glycosylation machinery is highly conserved across different mammalian species,

constituting the primary factor for the current preference for mammalian cells as the expression

host for the production of human-like therapeutic glycoproteins. Glycans produced are majorly of

the complex type, but some marginal differences between human and non-human mammalian

glycosylation exist, varying with the cell type, as summarized in Table 1.7 (Chung et al. 2008;

Dingermann 2008; Butler 2004).

TABLE 1.7. Structural differences in the N-glycans produced by the main mammalian cells used as hosts for

glycoprotein production, concerning human glycosylation

CELL LINE DIFFERENCE TO HUMAN EFFECT

Chinese hamster ovary (CHO) Absence of α2,6-linked sialic acids Undersialylation, reduced serum half-life

Absence of bisecting GlcNAc Reduced biological activity of mAbs

NeuGc produced at minor fractions Not immunogenic at the levels produced

Baby hamster kidney (BHK) Absence of α2,6-linked sialic acids Undersialylation, reduced serum half-life

NeuGc produced at minor fractions Not immunogenic at the levels produced

Murine myeloma (NS0, Sp2/0) Presence of α1,3-Gal Immunogenic

Predominance of NeuGc sialic acids Immunogenic


45

Among mammalian cell lines, CHO are the most widely used for the production of

therapeutics. These cells have not been implicated in any obvious adverse effects, but some

differences have been found in their potential for glycosylation compared to human cells

(Durocher and Butler 2009; Sethuraman and Stadheim 2006). One of the differences is on the

type of linkage established between sialic acid and Gal residues in the glycans. In humans, the

sialic acids are attached to Gal in a mixture of α2,6- and α2,3-linkages, with preference to the

former (Tarelli 2007). However, CHO cells lack the enzyme α2,6-sialyltransferase and thus can

only produce α2,3-linked terminal sialic acid residues, which can result in undersialylation and

consequently affect circulatory lifetime and other protein properties (Butler 2004; Tarelli 2007).

CHO cells also lack a functional GnT-III enzyme present in human cells, which prevents the

addition of bisecting GlcNAc residues and may impact the activity of certain glycoproteins, such

as mAbs (Jefferis 2005; Butler 2004; Jenkins and Curling 1994; Umana et al. 1999).

BHK cells are other common expression system in recombinant protein production, and like

CHO cells they only produce α2,3-linked terminal sialic acids due to the absence of a functional

α2,6-sialyltransferase (Butler 2004).

The differences found in the glycosylation potential of CHO and BHK cells, compared to

human cells, do not appear to result in glycoproteins that are immunogenic (Butler 2005).

However, immunogenic concerns appear for murine myeloma cells (e.g. NS0, SP2/0), another

expression system that is considered for glycoprotein production. These cells express the enzyme

α1,3-galactosyltransferase, which generates α1,3-Gal residues that are not present in humans

and are found to be highly immunogenic (Butler 2004; Harcum 2006; Jenkins, Parekh, and

James 1996). Indeed, about 1-2 % of circulating antibodies in humans are directed against these

α1,3-Gal residues (also known as the Galili antigen) (Galili et al. 1984), and therefore may

recognize therapeutic glycoproteins produced in these non-human expression systems (Durocher

and Butler 2009). Indeed, some reports about hypersensitivity reactions in patients receiving

treatment with a mAb produced in these cells have been reported due to the presence of the

α1,3-Gal epitope, which only highlight the concerns, especially if the therapeutic is to be

administered chronically in large doses, where severe immune responses may arise (Chung et al.

2008; Harcum 2006).

Murine myeloma cells also express glycoforms with terminal sialic acid that differ from those

present in humans. NeuNAc is the sialic acid produced by human cells, and although murine

cells also produce NeuNAc, the N-glycolylneuramic acid (NeuGc) is the sialic acid predominantly


46

present in their glycoproteins (Beck et al. 2008; Jenkins, Parekh, and James 1996; Tarelli 2007).

These NeuGc residues are potentially immunogenic in humans and can reduce the efficacy of the

therapeutic glycoprotein due to a rapid clearance by the immune system (Padler-Karavani et al.

2008; Jenkins, Parekh, and James 1996; Durocher and Butler 2009; Sheeley, Merrill, and Taylor

1997). CHO and BHK cells also produce these NeuGc residues but only as a minor fraction,

which does not seem to elicit any anti-NeuGc response (Dingermann 2008; Padler-Karavani et al.

2008; Beck et al. 2008).

Human cell-based expression systems, such as HEK and PER.C6 cells, are expected to

produce recombinant proteins with post-translational modifications more similar to their natural

counterpart. Therefore, the expression of therapeutic proteins in human cells could reduce

potential immunogenic reactions against the non-human epitopes found in other mammalian cell

lines as described above (Durocher and Butler 2009).

ENVIRONMENTAL FACTORS AND CULTURE CONDITIONS

Usually, several parameters of the cell culture environment are optimized during therapeutic

production with the intent to optimize cell growth and product yield. These parameters can also

change the protein glycosylation and consequently affect the biological activity of the product

(Gawlitzek et al. 1995; Hossler, Khattak, and Li 2009; Butler 2004; Yuen et al. 2003). The

parameters may be divided into the following categories: medium and nutrients, culture

conditions, technology platform, and cell-related factors. It is essential to acquire a better

understanding of the impact of these cell culture parameters on the resulting glycosylation to

improve the control of protein quality and maintain a consistent glycoform profile within the strict

limits required by the regulatory authorities (Walsh and Jefferis 2006). An extensive list of studies

on the effects of different culture conditions on glycosylation is provided in Table 1.8.


47

TA

BLE

1.8

. List of the main results reported on the effects of different variables of cell culture processes on product glycosylation

FACTORS

EFFECT

REFERENCE

MEDIUM AND NUTRIENTS

GLUCOSE

Glucose limitation reduced site occupancy of an IgG produced in mouse myeloma cells

(Stark and Heath, 1979)

Glucose limitation produced less glycosylated IFN-γ in CHO cells

(Hayter et al., 1992)

Glucose addition to the culture medium slightly improved galactosylation of an anti-Rhesus D antibody

(Nahrgang et al., 1999)

Glucose limitation led to decreased sialylation and increased hybrid and high-mannose type glycans of IFN-γ

produced in fed-batch cultures of CHO cells

(Wong et al., 2005)

GALACTOSE

Galactose feeding facilitated a more fully galactosylated N-glycan profile

(Andersen, 2004)

GLUTAMINE/AMMONIA

Increased ammonia led to a decrease in terminal galactosylation and sialylation of a recombinant tumor

necrosis factor receptor-IgG in CHO cells

(Gawlitzek et al., 2000)

Ammonia decreased sialylation of O-glycans even at low levels

(Andersen and Goochee, 1994)

High ammonia concentrations increased heterogeneity, decreased the terminal sialylation of O- and N- glycans

and decreased the content of O-glycans of EPO

(Yang and Butler, 2000a, b, 2002)

SERUM

Serum-free improved N- and O-glycosylation of IL-2, with enhanced terminal sialylation and proximal

fucosylation, in suspension and microcarrier cultures of BHK-12 cells

(Gawlitzek et al., 1995)

Serum-free culture increased terminal sialylation and galactosylation in a IgG1 monoclonal antibody produced

by murine hybridoma

(Patel et al., 1992)

Serum provided slightly higher terminal galactosylation of a mAb produced by CHO cells

(Lifely et al., 1995)

Serum caused proteolysis of IFN-γ produced by CHO cells

(Castro et al., 1995)

AMINO ACIDS

Supplementation with amino acids that have been depleted from the culture (cysteine, isoleucine, tryptophan,

valine, asparagine, aspartic acid, and glutamate) increased sialylation of EPO in CHO cell cultures

(Crowell et al., 2007)

MANGANESE

Manganese added to the culture increased galactosylation which in turn facilitated O- and N-glycosylation of

EPO in CHO cell cultures

(Crowell et al., 2007)

SODIUM-BUTYRATE

Sodium-butyrate added to a culture of CHO cells increased the sialylation of IFN-γ

(Lamotte et al., 1999)

Sodium-butyrate added to a culture of CHO cells increased the microheterogeneity and decrease the sialylation

and in vivo biological activity of human thrombopoietin

(Sung et al., 2004)


48

FACTORS

EFFECT

REFERENCE

GLYCEROL

Glycerol stabilized and enhanced sialylation of IFN-β in CHO cell cultures

(Rodriguez et al., 2005)

DIMETHYLSULFOXIDE

Dimethylsulfoxide (DMSO) decreased sialylation of IFN-β in CHO cell cultures

(Rodriguez et al., 2005)

LIPIDS

Lipids improved the N-glycan site occupancy of IFN- γ produced by CHO cells

(Castro et al., 1995)

Lipoprotein supplementation increased the amount of fully glycosylated IFN-γ produced by CHO cells

(Jenkins et al., 1994)

NUCLEOTIDE PRECURSORS

Cystidine and uridine increased the availability of nucleotide precursors and modified protein glycosylation

(Kornfeld and Kornfeld, 1985)

N-acetyl mannosamine (ManNAc), a direct precursor of CMP-NeuAc, added to CHO cell cultures increased

significantly the sialylation of IFN-γ

(Gu and Wang, 1998)

ManNAc added to NS0 cultures improved the NeuAc to NeuGc ratio of tissue inhibitor of metalloproteinase 1

(TIMP-1) to more favorable (human) levels

(Baker et al., 2001)

Glucosamine/uridine supplementation increased the antennarity of N-glycans of tissue inhibitor of

metalloproteinase-1 in CHO cells, but not in NS0 cells, and caused a decrease in sialylation in both CHO and

NS0 cultures

(Baker et al., 2001)

Glucosamine decreased sialylation and the proportion of tetraantennary glycans of EPO in CHO-K1 cells

(Yang and Butler, 2002)

CULTURE CONDITIONS

DISSOLVED OXYGEN

A decrease of the dissolved oxygen (DO) level led to a gradual decrease in the galactosylation (mainly digalacto

glycans) of an IgG in murine myeloma cell cultures

(Kunkel et al., 1998)

High levels of DO enhanced sialylation of a recombinant follicle-stimulating hormone in CHO cell cultures

(Chotigeat et al., 1994)

DO concentrations affected the final glycosylation state of a alkaline phosphatase protein in insect cell culture,

with optimal glycosylation occurring at intermediate DO levels

(Zhang et al., 2002)

pH

Culture pH shifted the glycosylation pattern of a mouse placental lactogen expressed in CHO cells

(Borys et al., 1993, 1994)

Culture pH affected the galactosylation and sialylation levels of a mAb produced by hybridoma cells

(Müthing et al., 2003)

Decreases in culture pH caused an increase in the proportion of acidic (sialylated) isoforms of EPO produced

by CHO cells, with an optimal range of pH 6.8-7.2 favoring sialylation

(Yoon et al., 2005)

TA

BLE

1.8

. List of the main results reported on the effects of different variables of cell culture processes on product glycosylation (continuation)


49

FACTORS

EFFECT

REFERENCE

CARBON DIOXIDE

Elevated carbon dioxide (CO

2) slightly decreased tPA sialylation in CHO cultures in serum-free media

(Kimura and Miller, 1997)

Increases in CO

2 decreased polysialylation in CHO cell cultures

(Zanghi et al., 1999)

TEMPERATURE

Lowering temperature from 37 ºC to 33 ºC and 30 ºC caused a decrease of the sialic acid ratio in a EPO-Fc

protein expressed in CHO cells

(Trummer et al., 2006a)

TECHNOLOGY PLATFORM

MODE OF CULTURE

Perfusion cultures increased overall sialylation compared to fed-batch cultures

(Lipscomb et al., 2005)

Perfusion culture of CHO cells immobilized in macroporous microcarriers, in a fluidized bed bioreactor, showed

an increasing level of N-glycan sialylation of IFN-γ during the culture

(Goldman et al., 1998)

The glycosylation pattern of transferrin varies during standard batch culture of HepG2

(Hahn and Goochee, 1992)

BIOREACTOR

Culture of CHO cells in a stirred tank bioreactor showed a decline of the N-glycan sialylation of IFN-γ late in

culture, coincident with the onset of cell death and lysis

(Goldman et al., 1998)

A stirred tank bioreactor provided higher levels of galactosylation of an IgG produced in SP2/0 cells than a

hollow fiber bioreactor

(Nahrgang et al., 1999)

MODE OF GROWTH

Microcarrier culture of CHO cells resulted in a recombinant human tissue kallikrein protein with lower

sialylation than suspension cultures

(Watson et al., 1994)

Stationary culture resulted in hypergalactosylation of an IgG

(Lund et al., 1993)

CELL RELATED FACTORS

CELL GROWTH

Slower growing cells facilitated a more glycosylated protein secreted by CHO cells

(Lipscomb et al., 2005)

CHO cells in the death phase produced less galactosylated antibodies

(Kaneko et al., 2010)

CELL DENSITY

Low density cultures gave rise to hypergalactosylation of a human IgG monoclonal antibody produced by

lymphoblastoid cells

(Kumpel et al., 1994)

Confluent HepG2 cells produced more active (biantennary) transferrin that subconfluent cultures

(Hahn and Goochee, 1992)

SPECIFIC PRODUCTIVITY

An increased specific productivity correlated to decreased levels of sialylation

(Trummer et al., 2006b)

EXPRESSION RATE

A lower rate of protein expression in MDCK cells increased the number of polylactosamine residues

(Nabi and Dennis, 1998)

A lower protein synthesis rate improved the glycosylation site occupancy of recombinant human prolactin

produced by C127 murine cells

(Shelikoff et al., 1994)

TA

BLE

1.8

. List of the main results reported on the effects of different variables of cell culture processes on product glycosylation (continuation)


50

MEDIUM AND NUTRIENTS

The cell culture medium determines the cell growth environment and physical conditions,

with a crucial influence on the course of the culture and on product quality (Zhu 2011; Walsh and

Jefferis 2006). The mammalian cell culture media is typically a mixture of 50-100 different

chemically-defined and sometimes undefined (e.g. peptones, yeast extracts, plant hydrolysates,

and serum) components (Hossler, Khattak, and Li 2009). The effects of some of these

components on protein glycosylation have been evaluated.

Glucose is the major energy source in mammalian cell culture (Burgener and Butler 2006;

Zhao and Keating 2007), and cell growth under glucose limitation seems to produce

abnormalities in the glycoprotein synthesis, such as the attachment of aberrant precursors to the

protein, reduced site occupancy, or the absence of glycosylation (Hayter et al. 1992; Hayter et al.

1993; Nyberg et al. 1999). These effects have been attributed to a shortage of glucose-derived

glycan precursors and/or to an intracellular-energy-depleted state (Kornfeld and Kornfeld 1985;

Rearick, Chapman, and Kornfeld 1981), but seem to be variable with the type of cell and product

(Hayter et al. 1993; Nahrgang et al. 1999; Wong et al. 2005; Cruz et al. 2000). The use of

galactose feeding as an alternative source of energy may facilitate a more fully galactosylated N-

glycan profile (Hossler, Khattak, and Li 2009; Andersen and Reilly 2004).

Glutamine is normally added to the culture medium to provide an energy source for cells

and is also an essential precursor for nucleotide synthesis. However, glutamine is also a major

source of ammonia accumulation in the medium, which inhibits cell growth and affects protein

glycosylation (Yang and Butler 2000; Andersen and Goochee 1994; Valley et al. 1999), with its

major effect being a decrease in terminal sialylation (Andersen and Goochee 1994; Gawlitzek et

al. 2000; Yang and Butler 2002). This effect on glycosylation could be explained by an increase

in the pH of the Golgi caused by ammonia, which could decrease the activity of some

glycosyltransferases (Gawlitzek et al. 2000; Butler 2004).

Serum is commonly used in cell culture to promote cell growth and improve shear

resistance, but due to safety and regulatory reasons it is now avoided in the biopharmaceutical

industry. This requires an adaptation procedure of the cell line to the new environmental

conditions, which can impact cell growth, product yield, and product glycosylation (LeFloch et al.

2006; Harcum 2006; Restelli and Butler 2002; Geigert 2004). The effects of serum and serum-

free medium on protein glycosylation have been attributed to the variable presence of proteolytic

activities of extracellular enzymes (e.g. sialydase, galactosydase) that may alter glycan


51

composition (Castro et al. 1995; Gawlitzek et al. 1995; Yang and Butler 2002). Furthermore,

these effects seem to be conflicting and very dependent on the type of cell and product (LeFloch

et al. 2006; Patel et al. 1992; Lifely et al. 1995), and may also vary with the different

formulations of serum-free media (Restelli and Butler 2002).

Other supplements used for the enhancement of production have been evaluated for their

potential effect on protein glycosylation, including lipids (Castro et al. 1995), amino acids,

manganese (Crowell et al. 2007), dimethylsulfoxide (DMSO), glycerol (Rodriguez et al. 2005),

and sodium butyrate (Sung et al. 2004; Mimura et al. 2001), with varying effects on

glycosylation, but typically affecting the level of sialylation.

The availability of nucleotide monosaccharide precursors may be a limiting factor for

glycosylation. The addition of some precursors to cell culture has been evaluated, including

glucosamine, cystidine, uridine, and N-acetyl mannosamine (ManNAc), with effects mainly on

sialylation and antennary levels. These effects have been variable with the type of product and

cells, leading to glycosylation enhancements (Kornfeld and Kornfeld 1985; Baker et al. 2001),

deterioration (Yang and Butler 2002), or no alteration (Baker et al. 2001).

CULTURE CONDITIONS

Several control parameters which are critical for the success of cell culture and product yield

have also been shown to affect protein glycosylation, including dissolved oxygen (DO), pH, carbon

dioxide (CO2), and temperature.

Oxygen plays a dominant role in the metabolism and viability of cells, but has low solubility

in the culture medium (Butler 2004). Therefore, it is important to continuously monitor and

control the DO level in cell culture, to maintain optimal metabolism and growth of producer cells

in bioprocesses (Zhu 2011; Hossler, Khattak, and Li 2009). DO has also shown to influence

glycosylation, apparently in a cell line- and/or protein-specific manner, but generally observed on

the levels of galactosylation (Kunkel et al. 1998) and sialylation (Chotigeat et al. 1994).

Culture pH is also an important parameter in cell culture and has been found to affect

protein glycosylation, specifically the levels of galactosylation (Müthing et al. 2003; Rothman et

al. 1989), sialylation (Müthing et al. 2003; Yoon et al. 2005), and glycan occupancy (Borys,

Linzer, and Papoutsakis 1993). The external culture pH may change the internal pH of the Golgi

apparatus, influencing the activities of key glycosylating enzymes and thereby changing the

glycan profile of the proteins produced (Butler 2004).


52

The effect of CO2 on glycosylation is also a very important bioprocess consideration, since

media pH is commonly maintained via the bicarbonate (HCO3-)/CO2 equilibrium. Some effects of

CO2 on sialylation levels have been reported (Kimura and Miller 1997; Zanghi et al. 1999).

Temperature shifts from the usual 37 ºC to lower values have been used as a means to

increase product titers (Moore et al. 1997; Yoon, Kim, and Lee 2003; Clark, Chaplin, and

Harcum 2004; Bollati-Fogolín et al. 2005). These shifts in some cases do not affect glycosylation

(Yoon, Song, and Lee 2003; Bollati-Fogolín et al. 2005), but there have also been reports of

changes in the sialylation levels (Trummer et al. 2006).

TECHNOLOGY PLATFORM

The technology platform (bioreactor and mode of culture) can have a pronounced effect on

the resulting glycoform profile (Kunkel et al. 2000). The three most common production modes

are batch, fed-batch, and perfusion (Hossler, Khattak, and Li 2009). In batch culture, due to

nutrient consumption and product accumulation that change the cellular environment, the

glycosylation pattern varies over the course of the culture (Hahn and Goochee 1992). For its turn,

fed-batch may be vulnerable to glycan degradation by extracellular enzymes secreted by the cells

or released upon cell lyses (Butler 2004; Gramer and Goochee 1993), which may result in

significant heterogeneity of glycoforms. The early extraction of the product from the medium

reduces the residence time of the glycoprotein in culture and may lessen glycoform heterogeneity

(Butler 2004). It has also been reported that the use of perfusion culture seems to increase

sialylation in comparison to fed-batch cultures (Lipscomb et al. 2005).

The type of reactor also appears to affect protein glycosylation, with differences found on the

levels of sialylation (Goldman et al. 1998) and galactosylation (Nahrgang et al. 1999).

Furthermore, mammalian cells can be grown as anchorage-dependent cells in T-flasks or

microcarriers, or can be alternatively adapted to suspension culture. This adaptation process

leads to changes in the expression of cell surface proteins and may also affect glycosylation, in

particular the sialylation levels (Watson et al. 1994). Additionally, static culture may produce

proteins with higher levels of galactosylation (Lund et al. 1993; Kumpel et al. 1994).

CELL RELATED FACTORS

The culture conditions mentioned above affect the cell metabolism, in particular the cell

growth rate, density, viability, and specific productivity. These cell-specific parameters have been


53

shown to modify the glycosylation pathway by affecting the levels of extracellular glycosidases in

the medium, which can step-wise remove monosaccharides from the glycans (Gramer and

Goochee 1993; Trummer et al. 2006; Gramer et al. 1995). They may also alter the residence

time of the protein in the Golgi, which might facilitate a more fully glycosylated structure (Wang et

al. 1991; Nabi and Dennis 1998; Shelikoff, Sinskey, and Stephanopoulos 1994), although some

studies have reported no effect (Bulleid et al. 1992).

1.6.6. GLYCOENGINEERING FOR IMPROVED THERAPEUTIC EFFICACY

As reflected in the previous sections, the control of glycosylation to a homogeneous

glycoform of desired characteristics is still a challenge that results in most of the glycoprotein-

based therapeutics currently being produced as mixtures of glycoforms. Such variability reduces

the therapeutic efficacy of the product, requiring its administration at higher doses and/or

frequency to obtain the desired effect in the patient (Beck et al. 2008; Wang and Lomino 2011).

This is a particularly relevant concern for therapeutic mAbs since their specific mode of action,

through binding to the target, is inhibited by the competition between undesired and desired

glycoforms, which further reduces the in vivo efficacy (Kanda et al. 2007; Iida et al. 2006; Satoh,

Iida, and Shitara 2006). To overcome these hurdles, a new approach has been explored in the

past decade, based on glycoengineering technologies that modify the glycans associated to the

proteins for an enhanced therapeutic efficacy. These technologies are applied to the glycoprotein

itself or to the host cell used for its production and have already achieved important

improvements on glycan quality (Wang and Lomino 2011; Shen et al. 2006; Sinclair and Elliott

2005; Yamane-Ohnuki and Satoh 2009).

CELL GLYCOENGINEERING

Perhaps the most intensely pursued glycoengineering approach in recent years has been the

engineering of the host cells through modifications of the glycosylation pathway to obtain a final

product with advantageous properties (Restelli and Butler 2002; Lim et al. 2010; Wang and

Lomino 2011). This approach has been used to improve the characteristics of the glycoproteins

produced by the currently preferred mammalian cell systems, but also to modify the glycosylation


54

mechanisms of non-mammalian cells to permit their application in therapeutic glycoprotein

production.

MAMMALIAN CELLS

Although considered as the most adequate cell host for the expression of therapeutic

glycoproteins, due to their mostly human-like glycosylation, mammalian cells still impart

glycoproteins with some differences compared to the native human counterpart. Consequently,

glycoengineering of mammalian cells has focused on modifying the glycosylation pathways of

mammalian cells to obtain more uniform and human-like glycoproteins, and to improve their

therapeutic efficacy. Different approaches have been explored to achieve this more defined

glycosylation, including:

(i) Knock-out mutagenesis, which consists of the knock-out of specific genes along the

glycosylation pathways. A major target of this method has been the α1,6-fucosyltransferase

gene (fut8), which encodes for the fucosyltransferase enzyme responsible for the

attachment of a core Fuc residue to the glycan. The successful knock-out of fut8 has led to

the development of a knockout CHO cell line that produces completely defucosylated

antibodies (Natsume, Niwa, and Satoh 2009; Yamane-Ohnuki et al. 2004), which exhibit

enhanced ADCC (Satoh, Iida, and Shitara 2006; Masuda et al. 2007) and induce high

cellular cytotoxicity against tumor cells that express low levels of antigen (Niwa et al. 2005),

therefore achieving therapeutic efficacy at low doses (Shinkawa et al. 2003; Yamane-Ohnuki

et al. 2004).

(ii) Ribonucleic acid (RNA) interference, which uses specific small-molecule inhibitors to block

selected enzymes in the biosynthetic pathway (Wang and Lomino 2011; Zhou et al. 2008).

For example, the introduction of small interfering RNA to silence the fut8 gene is one of the

methods used to defucosylate mammalian cells for the production of mAbs with enhanced

ADCC activity (Mori et al. 2004; Omasa et al. 2008). Other example of this methodology is

the antisense knockdown of the enzyme CMP-sialic acid hydroxylase, responsible for the

conversion of NeuNAc to NeuGc, which results in the reduction of the levels of the

potentially immunogenic NeuGc (Chenu et al. 2003).

(iii) Overexpression of specific glycoprocessing enzymes in the host cells, an approach that can

also change the glycosylation profile and enrich the production of desired glycoforms. A

successful example is the combined overexpression of the enzymes α2,3-sialyltransferase


55

and β1,4-galactosyltransferase, or of the enzymes α2,3-sialyltransferase and/or α2,6-

sialyltransferase, in the host cells, which resulted in increased sialylation of the glycoproteins

and consequently improved serum half-life (not applicable to mAbs) (Weikert et al. 1999;

Fukuta, Yokomatsu, et al. 2000; Jeong et al. 2008; Bork, Horstkorte, and Weidemann

2009). The overexpression of the enzymes GnT-I, GnT-IV, and GnT-V has also been

associated with improvements in sialylation due to an increase of the proportion of

branching structures and hence of the number of potential sites for sialylation (Fukuta, Abe,

et al. 2000; Goh et al. 2010). Another example of the use of this approach is the

overexpression of GnT-III to enhance the number of glycoforms containing bisecting GlcNAc

to improve the activity of mAbs via ADCC, since this residue blocks the attachment of core

Fuc, a monosaccharide known to be detrimental to ADCC (Davies et al. 2001; Ferrara,

Brünker, et al. 2006; Schuster et al. 2005). The combined overexpression of GnT-III and α-

mannosidase II has also been used to this effect (Matsumiya et al. 2007; Schuster et al.

2005; Ferrara, Brünker, et al. 2006).

NON-MAMMALIAN CELLS

Non-mammalian expression systems have several advantages over mammalian cells,

namely lower costs, high product yield, and simplicity of processes. However, due to their limited

capacity for glycosylation or to the introduction of potentially immunogenic residues in the

glycoproteins, they are not suitable for the production of therapeutic glycoproteins. Therefore,

glycoengineering emerges has a promising technology to overcome these limitations, by

providing means to humanize the glycoproteins produced in non-mammalian systems.

This humanization has been achieved using the same approaches described for mammalian

cells (gene knock-out, RNA interference, enzyme overexpression), but with different targets.

Indeed, in non-mammalian cells, the first objective is to avoid or eliminate the production of non-

human glycoforms, such as hypermannosylated glycans or glycans with immunogenic residues

(e.g. β1-2-Xyl and core α1-3-Fuc), which has been achieved by either gene knockout or RNA

interference methods (Chiba and Jigami 2007; Vervecken et al. 2004; Abe et al. 2009; Strasser

et al. 2008). After accomplishing this, a second objective of fully glycoprotein humanization may

be achieved through the functional transfer of mammalian glycan processing enzymes into the

non-mammalian cells, which essentially creates a mammalian biosynthetic pathway (De Pourcq,

De Schutter, and Callewaert 2010; Bobrowicz et al. 2004; Castilho et al. 2011).


56

One of the successful examples of the application of this strategy is the recent engineering

of the glycosylation pathways in the yeast Pichia pastoris that allowed the expression of

recombinant proteins with human-type complex glycans. This entailed knocking out four genes to

prevent yeast specific glycosylation, and introducing 14 additional glycosylation genes encoding

for glycosylation enzymes (Hamilton and Gerngross 2007).

PROTEIN GLYCOENGINEERING

Glycoengineering technologies have also been used to modify the glycan moiety itself, which

may be accomplished by different strategies:

(i) Incorporation of additional glycosylation sites into the protein backbone: in this strategy, new

N-glycosylation consensus sequences (Asn-X-Ser/Thr) are inserted into desired positions in

the protein backbone by site-directed mutagenesis. It has been used to create glycoproteins

with increased levels of glycosylation and consequently sialylation, leading to extended

serum half-life and improved in vivo activity (Elliott et al. 2003; Perlman et al. 2003).

(ii) Chemoselective and site-specific glycosylation of proteins: while the incorporation of

additional glycosylation sites into the protein backbone improves the therapeutic efficacy of

the protein by increasing the overall level of glycosylation, this approach acts in a more

specific manner. Here, the aim is to insert specific glycans into glycosylation sites that have

been pre-indentified as critical for proper biological activity. This is accomplished by the

introduction of specific tags (e.g. the natural cisteine or unnatural azide-, aldehyde-, alkyne-

and alkene-containing residues) at predetermined glycosylation sites by site-directed

mutagenesis. The tags are then selectively reacted with a modified glycan, containing an

appropriate functional group, via bio-orthogonal chemoselective ligation (Wang and Lomino

2011; Hirano et al. 2009; Wang, Winblade Nairn, et al. 2008). Since different tags can be

selectively introduced at different sites in a protein, it becomes possible to insert multiple

distinct glycans in a given protein through orthogonal chemoselective ligations (Wang and

Lomino 2011; van Kasteren et al. 2007), thereby controlling and improving the biological

efficacy of the glycoprotein.

(iii) Chemoenzymatic glycosylation remodeling: this strategy provides an attractive approach

toward glycan-defined glycoforms (Wang and Lomino 2011), and can be performed using

two strategies. In one of the strategies, the glycosylated recombinant proteins obtained in


57

the host cells are enzymatically trimmed down (deglycosylated) to the innermost GlcNAc,

giving a homogeneous GlcNAc-containing protein. The homogenous glycans are then

extended by selected enzymes such as glycosyltransferases and endoglycosidases to provide

a mature, glycan-defined glycoprotein (Wang and Lomino 2011; Wei et al. 2008; Zou et al.

2011). However, this strategy of sequential extension does not guarantee the homogeneity

of the end product, and mixtures of glycoforms may be produced (Wang and Lomino 2011).

An alternative to this sequential extension is the en block transfer of a presynthesized large

glycan to the protein in a single step under the catalysis of natural or mutant endo-β-N-

acetylglucosaminidase (ENGase) (Wang and Lomino 2011; Wang 2008; Walsh 2010; Zhu et

al. 2005). It should be noted that although efficient, these enzymatic strategies are generally

associated with cost issues (Yamane-Ohnuki and Satoh 2009).

1.6.7. STRATEGIES FOR GLYCOSYLATION ANALYSIS

The analysis of protein glycosylation for quality control, product optimization, and safety

reasons is of huge concern to the biotechnology industry, but presents technical challenges.

Indeed, owing to the complexity and heterogeneity of glycans, the chemical similarity in their

monosaccharide subunits, and the possibility of many different types of linkages between them, a

single analytical method rarely provides full compositional and structural analysis (Brooks 2009;

Dalpathado and Desaire 2008). On their own, each method provides helpful information, but for

a reliable and detailed analysis, several complimentary approaches need to be employed in

combination (Brooks 2009; Roth, Yehezkel, and Khalaila 2012). For this reason, regulatory

agencies require the use of at least two or more orthogonal analytical methods to fully

characterize therapeutic glycoproteins, such as mAbs (Harcum 2006; Raju 2008; Marino et al.

2010). The selection of the appropriate analytical techniques depends, among other things, on

the amount and purity of sample available for analysis, cost per sample, the resources and

technical expertise available, and the level of compositional and structural detail required (Brooks

2009; Harcum 2006).

Broadly, glycosylation analysis can be divided into four strategies: glycoprotein, glycopeptide,

glycan, and monosaccharide analysis (Figure 1.10) (Harcum 2006). All the strategies require

prior purification of the protein of interest, usually achieved by gel electrophoresis or affinity


58

chromatography (e.g. protein A or protein G for mAb purification) (Küster et al. 2001; Wang, Wu,

and Hancock 2006; Huhn et al. 2009).

FIGURE 1.10. Approaches for the glycosylation analysis of proteins. The glycoprotein can be analyzed intact or

fractioned into glycopeptides, glycans or monosaccharides for more detailed assessment. The analysis by one

approach can be followed by another, with the most and less common analytical pathways shown as full and dashed

lines, respectively.

GLYCOPROTEIN ANALYSIS

The analysis of intact glycoproteins is an easy, fast and inexpensive approach, but limited by

the relatively low level of resolution obtained (Harcum 2006). Consequently, this approach is

most commonly used to obtain preliminary information on molecular mass, to judge the outcome

of enzymatic reactions, to assess the sites of heterogeneity, and for purposes of purification and

fractioning prior to mass spectrometry analysis used in other approaches. Additionally, intact

glycoprotein analysis can be used for a fast monitoring during the optimization of manufacturing

processes and to assess the variability and lot-to-lot consistency (Taverna et al. 1998; Huhn et al.

2009).

Different techniques are used to analyze intact glycoproteins, based on their separation

according to molecular weight, charge, or affinity.

MOLECULAR WEIGHT

Proteins can be separated according to their molecular weight by conventional sodium

dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), and then identified as

GLYCOPROTEIN GLYCOPEPTIDE

GLYCAN

MONOSACCHARIDE


59

glycosylated proteins (glycoproteins) by staining the gel with the periodic acid Schiff (PAS)

reaction (Patton 2001; Osborne and Brooks 2006; Roth, Yehezkel, and Khalaila 2012). For a

more specific identification, the separated glycoproteins can be transferred onto a supporting

membrane to be probed for the binding of lectins or antibodies (see below for affinity-based

analysis) (Osborne and Brooks 2006).

CHARGE

The use of two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) allows the

separation of glycoproteins by both molecular weight and charge. This technique, also followed

by PAS staining, provides a more powerful approach than simple SDS-PAGE for the separation of

glycoproteins from complex and heterogeneous mixtures (Brooks 2009; Wilson et al. 2002; Dwek

and Rawlings 2006).

AFFINITY

Affinity-based procedures are more specific and allow the determination of the type of

glycosylation (e.g. N- or O-linked) (Roth, Yehezkel, and Khalaila 2012). They mostly rely on

lectins, which are proteins that bind highly selectively to all monosaccharides or glycans on the

glycoprotein, or sometimes on anti-glycan mAbs that are more specific for certain glycan

determinants and more expensive (Cummings and Etzler 2009; Brooks 2009). Probing

glycoproteins with a variety of lectins and antibodies (in a format similar to enzyme-linked

immunosorbent assays) is especially useful to determine whether two samples differ in

glycosylation (Cummings and Etzler 2009; Roth, Yehezkel, and Khalaila 2012).

Recently there has been interest in using lectins and antibodies in arrays or ‘chips’ for

qualitative and quantitative glycan profiling of purified glycoproteins or mixtures of glycoproteins

(Hsu and Mahal 2006; Pilobello, Slawek, and Mahal 2007). This technology provides an

indication of the range of glycan structures present and may become important for a fast, low-

tech, and high-throughput analysis (Brooks 2009; Patwa et al. 2010).

GLYCOPEPTIDE ANALYSIS

Glycopeptide analysis is the least established approach, but is currently showing rapid

development. The interest of this strategy is the possibility of correlating the glycan composition

to the attachment site, a characteristic that makes it often referred as a glycosylation site-specific


60

analysis. The assignment of a structure to a specific site may provide an indication of the function

of the glycan and shed light on the glycosylation profile and microheterogeneity of a protein, and

consequently on its activity (Dalpathado and Desaire 2008; Roth, Yehezkel, and Khalaila 2012).

Glycopeptide analysis can be applied to both N- and O-linked glycans, although the former is

more common, and involves different steps. Initially the glycoprotein is isolated from complex

biological mixtures, traditionally by gel electrophoresis or affinity chromatography (using

antibodies for protein-specific analysis or lectins to isolate a broad range of glycoproteins) (Küster

et al. 2001; Geng et al. 2001; Wang, Wu, and Hancock 2006; Dalpathado and Desaire 2008).

The isolated glycoprotein is then typically denatured (i.e. by reduction and alkylation) to improve

the efficiency of the following step of proteolysis (Dalpathado and Desaire 2008). Proteolysis is

performed to obtain both glycosylated and non-glycosylated peptides from the glycoprotein, which

can be achieved with a variety of enzymes (e.g. trypsin, which cleaves the protein at well-defined

sites, or non-specific proteinase K and pronase) (Dalpathado and Desaire 2008; Larsen, Højrup,

and Roepstorff 2005; Wuhrer et al. 2004). The peptides may then be purified for glycopeptide

enrichment, which consists on the separation of glycopeptides from non-glycosylated peptides

and is usually accomplished by reverse phase high-performance liquid chromatography (RP-

HPLC) on its own or combined with other purification methods (e.g. lectin affinity

chromatography or size exclusion chromatography), by hydrophilic interaction chromatography

(HILIC) or normal-phase HPLC (NP-HPLC) (Peterman and Mulholland 2006; Wuhrer et al. 2004;

Hemström and Irgum 2006). Finally, the glycopeptides are analyzed, typically by mass

spectrometry (MS) techniques (Dalpathado and Desaire 2008; Wuhrer et al. 2007).

GLYCAN ANALYSIS

Glycan analysis is the approach of choice for in-depth structural characterization, including

linkage information. The overall strategy for glycan analysis includes glycan release, derivatization

and sequencing, to obtain a final structure assignment.

Depending on its nature, glycans can be removed from the protein by chemical or enzymatic

means. The chemical removal can be applied to both N- and O-linked glycans, and includes the

methods of β-elimination, either alkaline (reductive, may cause glycan degradation) or ammonia-

based (non-reductive, prevents degradation), and the widely used and high-yielding hydrazinolysis

(Brooks 2009; Merry et al. 2002; Huang, Mechref, and Novotny 2001). On the other hand, the


61

enzymatic release relies on specific enzymes (glycanases) that cleave the glycans from the

glycoprotein, and is mostly used for N-glycans, with limited application in O-glycans. This is a

consequence of the inability to find a generic glycanase that acts on all O-glycan core structures

(only one specific enzyme was found against Core 1), making the chemical methods the current

best approach for their release (Marino et al. 2010; Merry et al. 2004). The most efficient

cleaving enzyme for N-glycan release is the peptide N-glycosidase F (PNGase F), which cleaves

the bond between the core GlcNAc and the Asn residue of all classes of N-glycans, with the

exception of those containing a core α1,3-Fuc (Tretter, Altmann, and März 1991; Roth, Yehezkel,

and Khalaila 2012). To cleave N-glycans containing the core α1,3-Fuc, PNGAse A should be used

(Royle et al. 2007; Tretter, Altmann, and März 1991). Alternatively, endoglycosidases that cleave

N-glycans between the two core GlcNAcs can be used, such as endo-H and endo-F1,2,3 (Trimble

and Tarentino 1991; Roth, Yehezkel, and Khalaila 2012; Marino et al. 2010).

After being released, glycans are commonly labeled using a fluorescent tag to enable their

detection and quantitation at the fentomole level. This labeling is performed by derivatization with

a fluorophore, with the most common including 2-aminobenzamide (2-AB), 2-aminopyridine (2-

AP), 2-aminoanthranilic acid (2-AA), 2-aminoacridone (AMAC), and 8-aminonaphtalene-1,3,6-

trisulfonic acid (ANTS) (Merry et al. 2004; Bigge et al. 1995; Domann et al. 2007). The labeled

glycans may then be separated by high-resolution techniques for sequence analysis by

chromatographic, electrophoretic and mass spectrometry methods (Merry et al. 2004).

CHROMATOGRAPHIC METHODS

Chromatography is one of the most widely used, and most accessible, approaches for the

analysis of glycans. Depending on the specific method, it is possible to separate the glycans

according to different physical characteristics, and it is common to use more than one approach

in parallel to obtain more complete compositional or structural data. Furthermore, to obtain a full

structural analysis, chromatographic methods are often used in conjunction with other

approaches like MS (see below) (Brooks 2009). The most commonly used chromatographic

methods in glycan analysis include: high performance anion exchange chromatography (HPAEC)

usually combined with pulse amperometric detection (HPAEC-PAD), for the analysis of

underivatized glycans (Townsend and Hardy 1991); weak anion exchange (WAX-HPLC), which

separates glycans by charge (Tran et al. 2000; Royle et al. 2002; Guile, Wong, and Dwek 1994);

RP-HPLC, which separates glycans on the basis of their hydrophobicity (Royle et al. 2002; Brooks


62

2009); and NP-HPLC, which separates glycans on amine columns where the retention time

mainly depends on the first approximation of the size of the glycan (Royle et al. 2002; Guile et al.

1996). This latter method is the most widely used for glycan analysis due a high resolution and

reproducibility. By using an external standard (e.g. hydrolyzed dextran), it is possible to convert

the retention times of the separated glycans to glucose units (GU), and compare them to those of

a database of GU values from known structures (e.g. Glycobase). The GU value of each individual

glycan structure is directly related to the number and linkage of its constituent monosaccharides,

so the GU values can be used to obtain a preliminary structure assignment that nevertheless

needs further confirmation (Guile et al. 1996; Brooks 2009; Campbell et al. 2008). For this,

digestion with exoglycosidases (enzymes that specifically cleave terminal monosaccharides from

the glycans, Table 1.9), are often used, either individually or in arrays, followed by profiling of the

digestion products by NP-HPLC (Royle et al. 2007).

TABLE 1.9. Exoglycosidase enzymes used for glycan analysis and respective monosaccharide and linkage specificity

(Royle et al. 2007)

ENZYME ABBREVIATION SPECIFICITY

Arthrobacter uerafaciens sialidase Abs Sialic acid α2-6>3,8

Streptococcus pneumoniae sialidase Nan1 Sialic acid α2-3,8

Bovine testis galactosidase Btg Gal β1-3,4>6

Streptococcus pneumoniae galactosidase Spg Gal β1-4

Coffee bean α galactosidase Cbg Gal α1-3,4,6

Bovine kidney focusidase Bkf Fuc β1-6>2>>3,4

Almond meal fucosidase Amf Fuc β1,3-4

Streptococcus pneumoniae N-acetylhexosaminidase Sph GlcNAc β1-2>3,4,6

Jack bean N-acetylhexosaminidase Jbh GlcNAc, GalNAc β1-2,3,4,6

Jack bean mannosidase Jbm Man α1-2,3,6

The existence of peak shifts after digestion with the exoglycosidases is indicative of the

specific activity of the enzymes and can be related to both the type and the number of

monosaccharides removed (Table 1.10), enabling detailed structural assignments (Guile et al.

1996; Brooks 2009; Royle, Dwek, and Rudd 2001).


63

TABLE 1.10. Incremental glucose unit (GU) values of different monosaccharides and linkages for the structural

assignment of 2-AB labeled N-glycans (Royle et al. 2007)

MONOSACCHARIDE LINKAGE TO GU INCREMENT

Man α1-2,3,6 Man 0.7-0.9

GlcNAc β1-2,4,6 α-Man 0.5

GlcNAc (bisecting) β1-4 β-Man 0.2-0.4

Gal α or β1-3,4 GlcNAc or Gal 0.8-0.9

Fuc (core) α1-6 Core GlcNAc 0.5

Fuc (outer arm) α1-3,4 GlcNAc 0.8

Fuc (outer arm) α1-2 Gal 0.5

NeuAc α2-3,6 Gal 0.7-1.2

Additionally, it is possible to obtain relative glycan quantification by calculating the peak area

of each glycan in relation to the total repertoire, when using the 2-AB fluorescent label, due to the

direct correlation between its fluorescence intensity and the number of moles of the labeled

glycans (Royle et al. 2007; Roth, Yehezkel, and Khalaila 2012).

ELECTROPHORETIC METHODS

Electrophoretic methods such as capillary electrophoresis (CE) efficiently separate charged

molecules. In particular, CE with laser-induced fluorescence detection (CE-LIF) provides a

promising method for glycan analysis, offering high sensitivity, high resolution, and a separation

time considerably shorter than HPLC (Beck et al. 2008; Kamoda and Kakehi 2006; Marino et al.

2010). The combination of CE and MS (CE-MS/CE-LIF-MS) also holds promise as a powerful tool

for the structural analysis of glycans (Kamoda and Kakehi 2006; Gennaro and Salas-Solano

2008).

MASS SPECTROMETRY METHODS

MS analysis is based on glycan ionization, fragmentation, and mass identification of the

fragments, providing a link between mass and composition of the glycans (Marino et al. 2010;

Roth, Yehezkel, and Khalaila 2012). It requires prior derivatization of the glycans due to their low

ionization efficiency (Kang, Mechref, and Novotny 2008; Marino et al. 2010). Two main types of

MS are currently used for glycan analysis, namely matrix assisted laser desorption ionization

(MALDI-MS) and electrospray (ESI-MS) (Merry et al. 2004; Marino et al. 2010; Roth, Yehezkel,

and Khalaila 2012). MALDI-MS gives some indication of glycan identity and is often used in


64

conjunction with time of flight analysis (MALDI-MS-TOF) (Wada et al. 1994). ESI-MS provides data

that can be interpreted to determine monosaccharide linkage and position (Fenn et al. 1989).

Although MALDI and ESI spectra can be used for the elucidation of glycan composition and

structure, full glycan characterization by MS is rather difficult due to common glycan quantity

limitations and, more importantly, the identical molecular weight of some monosaccharides that

hinders their specific identification based solely on the number of carbon atoms (Roth, Yehezkel,

and Khalaila 2012; Brooks 2009). Therefore, MS techniques are often used in combination with

HPLC or CE (Roth, Yehezkel, and Khalaila 2012; Kamoda and Kakehi 2006; Marino et al. 2010).

MONOSACCHARIDE ANALYSIS

Monosaccharides can be obtained directly from an intact glycoprotein or glycopeptide pool,

or even from released glycans (Figure 1.10). They are cleaved through acid hydrolysis and

identified by chromatography or electrophoresis; or alternatively released through methanolysis

and identified and quantified by gas chromatography (GC) combined with MS (GC-MS) (Manzi

and Varki 1993; Tarelli 2007; Marino et al. 2010).

This approach provides preliminary information on identity and composition of the

monosaccharides, which is a good indicator of the type and relative amount of glycans present in

the glycoprotein (Tarelli 2007; Brooks 2009; Harcum 2006). This information is valuable for the

rational planning of further investigations, and can be used as a very basic measure of product

consistency and extent of glycosylation for recombinant glycoprotein therapeutics at all stages of

production (Marino et al. 2010; Brooks 2009; Tarelli 2007). Additionally, the use of methylation

followed by GC-MS can provide information about the linkage positions of individual

monosaccharides (Tarelli 2007; Brooks 2009).

In this brief description of the strategies for glycosylation analysis, several analytical

techniques were mentioned. Each of these techniques measures glycosylation based on diverse

physical and chemical characteristics of the glycoprotein and offers different information (Harcum

2006). No technique on its own is currently able to provide full glycan characterization, and their

integration is often used to obtain more in-depth profiling. Table 1.11 summarizes the major

techniques used in glycosylation analysis, indicating their basic principle, approaches of analysis

in which they apply, information gathered, and main advantages and limitations.


65

ANALYTICAL METHOD

PRINCIPLE

APPROACH

INFORMATION / USE

MAIN ADVANTAGES AND LIMITATION

ELECTROPHORETIC

SDS-PAGE

Molecular weight

Glycoprotein, glycopeptide

Separation by molecular weight, degree of glycosylation

+ Cheap, fast, adaptable to high-throughput

- Limited resolution, labor-intensive

IEF-PAGE

Charge

Glycoprotein, Glycopeptide

Separation by isoelectric point (pI), glycoform profiling

+ Cheap, fast

2D-PAGE

Molecular weight,

charge

Glycoprotein

Separation by pI and molecular weight, glycoform profiling

+ Cheap, fast, high resolution

- Low reproducibility, inability to resolve too large/small/acidic/

basic, low abundant or hydrophobic proteins

CE / CE-LIF

Charge

Glycoprotein, Glycopeptide, Glycan

Separation, profiling, site occupancy

+ Cheap, fast, high specificity, versatile, adaptable to high-

throughput

CHROMATOGRAPHIC

GC

Partition

Monosaccharide

Monosaccharide composition and quantification

+ Robust and high resolution

HPAEC-PAD

Charge, linked

isomers

Glycan, Monosaccharide

Separation without need for derivatization, glycan profiling,

monosaccharide identification

+ It does not require derivatization, high resolution

- Aggressive eluents, relatively extensive analysis time

WAX-HPLC

Charge, size (to a

less extent)

Glycoprotein, Glycan

Separation, determination of acidic and neutral

monosaccharides

+ High sensitivity

RP-HPLC

Polarity

Glycopeptide, Glycan, Monosaccharide

Separation, glycan profiling and sequencing, monosaccharide

identification

+ High resolution, simple

- Relatively extensive analysis time

NP-HPLC

Polarity, size

Glycopeptide, Glycan, Monosaccharide

Separation, glycan profiling and sequencing, monosaccharide

identification and relative quantification

+ High resolution, simple

- Relatively extensive analysis time

HILIC

Polarity, size

Glycopeptide, Glycan

Improved separation of glycans, including resolution of

structural isomers

+ High sensitivity, high reproducibility

TA

BLE

1.1

1. Analytical techniques to characterize glycosylation, their basic principle, approaches in which they apply, information provided and main advantages and limitations (Roth,

Yehezkel, and Khalaila 2012; Harcum 2006; Brooks 2009; Marino et al. 2010; Dalpathado and Desaire 2008; Huhn et al. 2009)(Roth, Yehezkel, and Khalaila 2012; Harcum 2006;

Brooks 2009; Marino et al. 2010; Dalpathado and Desaire 2008; Huhn et al. 2009)(Roth, Yehezkel, and Khalaila 2012; Harcum 2006; Brooks 2009; Marino et al. 2010; Dalpathado and

Desaire 2008; Huhn et al. 2009)(Roth, Yehezkel, and Khalaila 2012; Harcum 2006; Brooks 2009; Marino et al. 2010; Dalpathado and Desaire 2008; Huhn et al. 2009)


66

ANALYTICAL METHOD

PRINCIPLE

APPROACH

INFORMATION / USE

MAIN ADVANTAGES AND LIMITATION

MASS SPECTROMETRY

ESI-MS

Mass


Relative proportion of glycoforms, glycan sequencing,

monosaccharide linkage and position

+ Fast, detailed information, easily coupled to HPLC and CE

- Laborious sample preparation, complex data analysis

MALDI-MS

Mass

All

Relative proportion of glycoforms, heterogeneity in mass,

glycan sequencing

+ Fast, adaptable to high-throughput

- Laborious sample preparation, Quantitation not very reliable

MALDI-TOF-MS

Mass

All

Relative proportion of glycoforms, heterogeneity in mass,

glycan sequencing, site occupancy

+ Fast, adaptable to high-throughput

- Laborious sample preparation

GC-MS

Partition, mass

Monosaccharide

Identification and quantification of monosaccharides,

determination of monosaccharide linkage positions

+ Reliable, powerful

- Laborious sample preparation, complex data analysis

LC-MS

Polarity, mass

Glycopeptide, Glycan

Glycopeptide/glycan mass and profiling

+ High sensitivity and selectivity, less sample preparation

requirements

- Expensive equipment, complex data analysis

CE-MS / CE-LIF-MS

Charge, mass


Mapping of glycosylation sites, glycan mass and sequencing

+ High sensitivity and resolution

OTHER

Excoglycosidase digestion


Glycan linkage and sequencing

+ Detailed glycan fingerprinting, versatile

- Time-consuming and limited to the scope of enzymes available

NMR

Electromagnetism

Glycan, Monosaccharide

Full (tri-dimensional) structural information of complex glycans,

including monosaccharide anomericity

+ Non destructive

- Expensive equipment, complex data analysis, requires large

amounts and highly purified glycans

Immunoblotting /

Affinity assay

Affinity

Glycoprotein

Partial glycan characterization: differentiation between O- and

N-glycosylation, detection of sialic acids

+ Simple, fast

TA

BLE

1.1

1. Analytical techniques to characterize glycosylation, their basic principle, approaches in which they apply, information provided and main advantages and limitations (Roth,

Yehezkel, and Khalaila 2012; Harcum 2006; Brooks 2009; Marino et al. 2010; Dalpathado and Desaire 2008; Huhn et al. 2009)(Roth, Yehezkel, and Khalaila 2012; Harcum 2006;

Brooks 2009; Marino et al. 2010; Dalpathado and Desaire 2008; Huhn et al. 2009)(Roth, Yehezkel, and Khalaila 2012; Harcum 2006; Brooks 2009; Marino et al. 2010; Dalpathado

and Desaire 2008; Huhn et al. 2009)(Roth, Yehezkel, and Khalaila 2012; Harcum 2006; Brooks 2009; Marino et al. 2010; Dalpathado and Desaire 2008; Huhn et al. 2009)


67

1.7. CONCLUDING REMARKS

In spite of recent advances in the field, the characterization and control of glycosylation

during the diverse stages of therapeutic protein production is currently hindered by the scarce

understanding on the effect of different parameters and culture conditions in this property, as

well as by the lack of adequate analytical methodologies available. It is expected that in the

future, a more in-depth knowledge on the relationship between glycosylation and protein

functionality and on the influence of culture parameters on glycosylation will not only allow a

proper control of this modification but also enable manipulation of culture conditions towards a

more desirable glycoform profile. In this context, glycoengineering appears as a promising

technology, enabling the engineering of the glycoprotein itself or the glycan biosynthetic pathway

in the host cell, to produce glycoproteins with improved functionality. In particular,

glycoengineering opens new horizons to the application of non-mammalian expression systems to

produce therapeutic glycoproteins, by humanizing glycan biosynthesis and/or eliminating

potentially immunogenic structures.

Nevertheless, further progresses in the field will largely depend of the development of fast,

simple, high sensitive, and high-throughput analytical methodologies, which provide complete

structural analysis for glycosylation monitoring during different stages of process development

and production. In particular, the protocols for sample preparation and purification will have to be

addressed since they are currently a major time-consuming step of glycosylation analysis.

Ultimately, these developments will lead to therapeutic products with improved efficacy,

which will reduce the dose and frequency of administration required and therefore improve

patient compliance with the therapy.


68

1.8. REFERENCES

Abe H, Takaoka Y, Chiba Y, Sato N, Ohgiya S, Itadani A, Hirashima M, Shimoda C, Jigami Y, and

Nakayama K (2009). Development of valuable yeast strains using a novel mutagenesis technique for

the effective production of therapeutic glycoproteins. Glycobiology 19 (4):428-436.

Abès R, and Teillaud J-L (2010). Impact of Glycosylation on Effector Functions of Therapeutic IgG.

Pharmaceuticals 3 (1):146-157.

Adams T, Noack U, Frick T, Greller G, and Fenge C (2011). Increasing Efficiency in Protein and Cell

Production by Combining Single-Use Bioreactor Technology and Perfusion. BioPharm International

Supplements 24 (5):s4-s11.

Aldrich TL, Viaje A, and Morris AE (2003). EASE Vectors for Rapid Stable Expression of Recombinant

Antibodies. Biotechnology Progress 19 (5):1433-1438.

Altamirano C, Paredes C, Illanes A, Cairó JJ, and Gòdia F (2004). Strategies for fed-batch cultivation of t-

PA producing CHO cells: substitution of glucose and glutamine and rational design of culture

medium. Journal of Biotechnology 110 (2):171-179.

Altmann F, Staudacher E, Wilson IBH, and März L (1999). Insect cells as hosts for the expression of

recombinant glycoproteins. Glycoconjugate Journal 16 (2):109-123.

Andersen DC, and Reilly DE (2004). Production technologies for monoclonal antibodies and their

fragments. Current Opinion in Biotechnology 15 (5):456-462.

Andersen DC, and Goochee CF (1994). The effect of cell-culture conditions on the oligosaccharide

structures of secreted glycoproteins. Current Opinion in Biotechnology 5 (5):546-549.

Baker K, Flatman S, and Birch J (2007). Product characterization from gene to therapeutic product. In

Medicines from Animal Cell Culture: John Wiley & Sons, Ltd.

Baker KN, Rendall MH, Hills AE, Hoare M, Freedman RB, and James DC (2001). Metabolic control of

recombinant protein N-glycan processing in NS0 and CHO cells. Biotechnology and Bioengineering

73 (3):188-202.

Bañó-Polo M, Baldin F, Tamborero S, Marti-Renom MA, and Mingarro I (2011). N-glycosylation efficiency

is determined by the distance to the C-terminus and the amino acid preceding an Asn-Ser-Thr

sequon. Protein Science 20 (1):179-186.

Bardor M, Faveeuw C, Fitchette A-C, Gilbert D, Galas L, Trottein F, Faye L, and Lerouge P (2003).

Immunoreactivity in mammals of two typical plant glyco-epitopes, core α(1,3)-fucose and core

xylose. Glycobiology 13 (6):427-434.

Barnes LM, Bentley CM, and Dickson AJ (2000). Advances in animal cell recombinant protein production:

GS-NS0 expression system. Cytotechnology 32 (2):109-123.


69

Barnes LM, Bentley CM, and Dickson AJ (2001). Characterization of the stability of recombinant protein

production in the GS-NS0 expression system. Biotechnology and Bioengineering 73 (4):261-270.

Basta M (2008). Ambivalent effect of immunoglobulins on the complement system: Activation versus

inhibition. Molecular Immunology 45 (16):4073-4079.

Bebbington CR, Renner G, Thomson S, King D, Abrams D, and Yarranton GT (1992). High-level

expression of a recombinant antibody from myeloma cells using a glutamine synthetase gene as an

amplifiable selectable marker. Nature Biotechnology 10 (2):169-175.

Beck A , Wagner-Rousset E, Bussat MC, Lokteff M, Klinguer-Hamour C, Haeuw JF, Goetsch L, Wurch T,

Van Dorsselaer A, and Corvaïa N (2008). Trends in glycosylation, glycoanalysis and glycoengineering

of therapeutic antibodies and Fc-fusion proteins. Current Pharmaceutical Biotechnology 9 (6):482-

501.

Belshaw PJ, Ho SN, Crabtree GR, and Schreiber SL (1996). Controlling protein association and

subcellular localization with a synthetic ligand that induces heterodimerization of proteins.

Proceedings of the National Academy of Sciences of the United States of America 93 (10):4604-

4607.

Benz I, and Schmidt MA (2002). Never say never again: protein glycosylation in pathogenic bacteria.

Molecular Microbiology 45 (2):267-276.

Berson RE, and Friederichs G (2008). A self-feeding roller bottle for continuous cell culture. Biotechnology

Progress 24 (1):154-157.

Bertozzi CR, and Rabuka D (2009). Structural basis of glycan diversity. In Essentials of Glycobiology,

edited by A Varki, R Cummings, J Esko, H Freeze, P Stanley, C Bertozzi, G Hart and M Etzler. Cold

Spring Harbor, NY: Cold Spring Harbor Laboratory Press.

Bigge JC, Patel TP, Bruce JA, Goulding PN, Charles SM, and Parekh RB (1995). Nonselective and efficient

fluorescent labeling of glycans using 2-amino benzamide and anthranilic acid. Analytical

Biochemistry 230 (2):229-238.

Bill RM, Revers L, and Wilson IBH (1998). Protein Glycosylation. Dordrecht, The Netherlands: Kluwer

Academic Publishers.

Birch JR, and Racher AJ (2006). Antibody production. Advanced Drug Delivery Reviews 58 (5-6):671-685.

Blaas L, Musteanu M, Eferl R, Bauer A, and Casanova E (2009). Bacterial artificial chromosomes improve

recombinant protein production in mammalian cells. BMC Biotechnology 9 (1):3.

Blüml G (2007). Microcarrier cell culture technology. In Animal Cell Biotechnology: Methods and

Protocols, Second Edition, edited by R. Pörtner. Totowa, NJ: Humana Press Inc.

Bobrowicz P, Davidson RC, Li H, Potgieter TI, Nett JH, Hamilton SR, Stadheim TA, Miele RG, Bobrowicz B,

Mitchell T, Rausch S, Renfer E, and Wildt S (2004). Engineering of an artificial glycosylation pathway


70

blocked in core oligosaccharide assembly in the yeast Pichia pastoris: production of complex

humanized glycoproteins with terminal galactose. Glycobiology 14 (9):757-766.

Böhm E, Voglauer R, Steinfellner W, Kunert R, Borth N, and Katinger H (2004). Screening for improved

cell performance: Selection of subclones with altered production kinetics or improved stability by cell

sorting. Biotechnology and Bioengineering 88 (6):699-706.

Bollati-Fogolín M, Forno G, Nimtz M, Conradt HS, Etcheverrigaray M, Kratje R (2005). Temperature

reduction in cultures of hGM-CSF-expressing CHO cells: effect on productivity and product quality.

Biotechnology Progress 21 (1):17-21.

Bork K, Horstkorte R, and Weidemann W (2009). Increasing the sialylation of therapeutic glycoproteins:

The potential of the sialic acid biosynthetic pathway. Journal of Pharmaceutical Sciences 98

(10):3499-3508.

Borys MC, Linzer DIH, and Papoutsakis ET (1993). Culture pH affects expression rates and glycosylation

of recombinant mouse placental lactogen proteins by Chinese hamster ovary (CHO) cells. Nature

Biotechnology 11 (6):720-724.

Boyd PN, Lines AC, and Patel AK (1995). The effect of the removal of sialic acid, galactose and total

carbohydrate on the functional activity of Campath-1H. Molecular Immunology 32 (17-18):1311-

1318.

Brockhausen I, Schachter H, and Stanley P (2009). O-GalNac glycans. In Essentials of Glycobiology. 2nd

edition., edited by A Varki, R Cummings, J Esko, H Freeze, G Hart and J Marth. Cold Spring Harbor

(NY): Cold Spring Harbor Laboratory Press.

Brooks SA, Dwek MV, and Schumacher U (2002). Modulation of glycan expression using inhibitors of N-

and O-linked glycoconjugate processing and transgenic technology. In Functional and Molecular

Glycobiology. Oxford, UK: Bios Scientific.

Brooks S (2004). Appropriate glycosylation of recombinant proteins for human use. Molecular


Brooks S (2009). Strategies for Analysis of the Glycosylation of Proteins: Current Status and Future

Perspectives. Molecular Biotechnology 43 (1):76-88.

Brown M, Renner G, Field R, and Hassell T (1992). Process development for the production of

recombinant antibodies using the glutamine synthetase (GS) system. Cytotechnology 9 (1):231-236.

Brunner D, Frank J, Appl H, Schöffl H, Pfaller W, and Gstraunthaler G (2010). Serum-free cell culture: the

serum-free media interactive online database. ALTEX 27 (1):53-62.

Bulleid NJ, Bassel-Duby RS, Freedman RB, Sambrook JF, and Gething MJ (1992). Cell-free synthesis of

enzymically active tissue-type plasminogen activator. Protein folding determines the extent of N-

linked glycosylation. The Biochemical Journal 286:275-280.


71

Burda P, and Aebi M (1999). The dolichol pathway of N-linked glycosylation. Biochimica et Biophysica

Acta (BBA) - General Subjects 1426 (2):239-257.

Burgener A, and Butler M (2006). Medium development. In Cell culture technology for pharmaceutical

and cell-based therapies, edited by S Ozturk and W-S Hu. Boca Raton, FL: CRC Press.

Burton DR, and Woof JM (1992). Human antibody effector function. Advances in Immunology 51:1-84.

Butler M (2004). Animal Cell Culture and Technology. Edited by M Butler. 2nd ed, The Basics: Taylor &

Francis.

Butler M (2005). Animal cell cultures: recent achievements and perspectives in the production of

biopharmaceuticals. Applied Microbiology and Biotechnology 68 (3):283-291.

Butters TD (2002). Control in the N-linked glycoprotein biosynthesis pathway. Chemistry and Biology 9

(12):1266-1268.

Campbell MP, Royle L, Radcliffe CM, Dwek RA, and Rudd PM (2008). GlycoBase and autoGU: tools for

HPLC-based glycan analysis. Bioinformatics 24 (9):1214-1216.

Canatella PJ, Karr JF, Petros JA, and Prausnitz MR (2001). Quantitative study of electroporation-mediated

molecular uptake and cell viability. Biophysical Journal 80 (2):755-764.

Carswell S, and Alwine JC (1989). Efficiency of utilization of the simian virus 40 late polyadenylation site:

effects of upstream sequences. Molecular and Cellular Biology 9 (10):4248-4258.

Castilho A, Gattinger P, Grass J, Jez J, Pabst M, Altmann F, Gorfer M, Strasser R, and Steinkellner H

(2011). N-Glycosylation engineering of plants for the biosynthesis of glycoproteins with bisected and

branched complex N-glycans. Glycobiology 21 (6):813-823.

Castillo LR (2008). Animal Cell Technology: from Biopharmaceuticals to Gene Therapy Edited by AM

Moraes, LR Castilho and EFP Augusto: Taylor & Francis, Inc.

Castro PM, Hayter PM, Ison AP, and Bull AT (1992). Application of a statistical design to the optimization

of culture medium for recombinant interferon-gamma production by Chinese hamster ovary cells.

Applied Microbiology and Biotechnology 38 (1):84-90.

Castro PM, Ison AP, Hayter PM, and Bull AT (1995). The macroheterogeneity of recombinant human

interferon-gamma produced by Chinese-hamster ovary cells is affected by the protein and lipid

content of the culture medium. Biotechnology and Applied Biochemistry 21:87-100.

Cavazzana-Calvo M, Thrasher A, and Mavilio F (2004). The future of gene therapy. Nature 427

(6977):779-781.

Cereghino JL, and Cregg JM (2000). Heterologous protein expression in the methylotrophic yeast Pichia

pastoris. FEMS Microbiology Reviews 24 (1):45-66.

Chenu S, Grégoire A, Malykh Y, Visvikis A, Monaco L, Shaw L, Schauer R, Marc A, and Goergen J-L

(2003). Reduction of CMP-N-acetylneuraminic acid hydroxylase activity in engineered Chinese


72

hamster ovary cells using an antisense-RNA strategy. Biochimica et Biophysica Acta (BBA) - General

Subjects 1622 (2):133-144.

Chiba Y, and Jigami Y (2007). Production of humanized glycoproteins in bacteria and yeasts. Current

Opinion in Chemical Biology 11 (6):670-676.

Chisholm V (1995). High efficiency gene transfer into mammalian cells. In DNA Cloning: A Practical

Approach. Vol. IV. Mammalian Systems, edited by DM Glover and BD Hames. Oxford: Oxford

University Press.

Chotigeat W, Watanapokasin Y, Mahler S, and Gray PP (1994). Role of environmental conditions on the

expression levels, glycoform pattern and levels of sialyltransferase for hFSH produced by

recombinant CHO cells. Cytotechnology 15 (1):217-221.

Chu L, and Robinson DK (2001). Industrial choices for protein production by large-scale cell culture.

Current Opinion in Biotechnology 12 (2):180-187.

Chung CH, Mirakhur B, Chan E, Le Q-T, Berlin J, Morse M, Murphy BA, Satinover SM, Hosen J, Mauro D,

Slebos RJ, Zhou Q, Gold D, Hatley T, Hicklin DJ, and Platts-Mills TAE (2008). Cetuximab-induced

anaphylaxis and IgE specific for galactose-α-1,3-galactose. New England Journal of Medicine 358

(11):1109-1117.

Chusainow J, Yang YS, Yeo JHM, Toh PC, Asvadi P, Wong NSC, and Yap MGS (2009). A study of

monoclonal antibody-producing CHO cell lines: What makes a stable high producer? Biotechnology

and Bioengineering 102 (4):1182-1196.

Clark KJR, Chaplin FWR, and Harcum SW (2004). Temperature effects on product-quality-related enzymes

in batch CHO cell cultures producing recombinant tPA. Biotechnology Progress 20 (6):1888-1892.

Clark MR (1997). IgG effector mechanisms. Chemical Immunology 65:88-110.

Crispin M, Bowden TA, Coles CH, Harlos K, Aricescu AR, Harvey DJ, Stuart DI, and Jones EY (2009).

Carbohydrate and domain architecture of an immature antibody glycoform exhibiting enhanced

effector functions. Journal of Molecular Biology 387 (5):1061-1066.

Crowell CK, Grampp GE, Rogers GN, Miller J, and Scheinman RI (2007). Amino acid and manganese

supplementation modulates the glycosylation state of erythropoietin in a CHO culture system.

Biotechnology and Bioengineering 96 (3):538-549.

Cruz, HJ, Freitas CM, Alves PM, Moreira JL, and Carrondo MJT (2000). Effects of ammonia and lactate on

growth, metabolism, and productivity of BHK cells. Enzyme and Microbial Technology 27 (1–2):43-

52.

Cummings RD, and Etzler ME (2009). Antibodies and lectins in glycan analysis. In Essentials of

Glycobiology, 2nd Edition, edited by A Varki, R Cummings and J Esko. Cold Spring Harbor, NY: Cold

Spring Harbor Laboratory Press.


73

Dalpathado DS, and Desaire H (2008). Glycopeptide analysis by mass spectrometry. Analyst 133 (6):731-

738.

Dasgupta S, Navarrete AM, Bayry J, Delignat S, Wootla B, André S, Christophe O, Nascimbeni M,

Jacquemin M, Martinez-Pomares L, Geijtenbeek TBH, Moris A, Saint-Remy J-M, Kazatchkine MD,

Kaveri SV, and Lacroix-Desmazes S (2007). A role for exposed mannosylations in presentation of

human therapeutic self-proteins to CD4+ T lymphocytes. Proceedings of the National Academy of

Sciences 104 (21):8965-8970.

Davies J, Jiang LY, Pan LZ, LaBarre MU, Andersen D, and Reff M (2001). Expression of GnTIII in

recombinant anti-CD20 CHO production cell line: expression of antibodies with altered glycoforms

leads to an increase in ADCC through higher affinity for FcyRIII. Biotechnology and Bioengineering

74:288-294.

De Pourcq K, Schutter K, and Callewaert N (2010). Engineering of glycosylation in yeast and other fungi:

current state and perspectives. Applied Microbiology and Biotechnology 87 (5):1617-1631.

Dempski REJ, and Imperiali B (2002). Oligosaccharyl transferase: gatekeeper to the secretory pathway.

Current Opinion in Chemical Biology 6 (6):844-850.

Derouazi M, Girard P, Van Tilborgh P, Iglesias K, Muller N, Bertschinger M, and Wurm FM (2004). Serum-

free large-scale transient transfection of CHO cells. Biotechnology & Bioengineering 87 (4):537-545.

Dingermann T (2008). Recombinant therapeutic proteins: production platforms and challenges.

Biotechnology Journal 3 (1):90-97.

Domann PJ, Pardos-Pardos AC, Fernandes DL, Spencer DIR, Radcliffe CM, Royle L, Dwek RA, and Rudd

PM (2007). Separation-based glycoprofiling approaches using fluorescent labels. Proteomics 7

(S1):70-76.

Dorka P, Fischer C, Budman H, and Scharer J (2009). Metabolic flux-based modeling of mAb production

during batch and fed-batch operations. Bioprocess and Biosystems Engineering 32 (2):183-196.

Doyle A., and Griffiths JB (1998). Cell and Tissue Culture: Laboratory Procedures in Biotechnology.

Chichester, England: John Wiley & Sons.

Durocher Y, and Butler M (2009). Expression systems for therapeutic glycoprotein production. Current

Opinion in Biotechnology 20 (6):700-707.

Dwek MV, Rawlings SL (2006). Breast cancer proteomics using two-dimensional electrophoresis: studying

the breast cancer proteome. Methods in Molecular Medicine 120:231-243.

Egrie JC, and Browne JK (2001). Development and characterization of novel erythropoiesis stimulating

protein (NESP). British Journal of Cancer 84 (S1):3-10.

Elbein AD (1987). Inhibitors of the biosynthesis and processing of N-linked oligosaccharide chains. Annual

Review of Biochemistry 56 (1):497-534.


74

Elbein AD (1991). Glycosidase inhibitors: inhibitors of N-linked oligosaccharide processing. The FASEB

Journal 5 (15):3055-63.

Ellgaard L, and Helenius A (2001). ER quality control: towards an understanding at the molecular level.

Current Opinion in Cell Biology 13 (4):431-437.

Elliott S, Lorenzini T, Asher S, Aoki K, Brankow D, Buck L, Busse L, Chang D, Fuller J, Grant J, Hernday

N, Hokum M, Hu S, Knudten A, Levin N, Komorowski R, Martin F, Navarro R, Osslund T, Rogers G,

Rogers N, Trail G, and Egrie J (2003). Enhancement of therapeutic protein in vivo activities through

glycoengineering. Nature Biotechnology 21 (4):414-421.

Erbayraktar S, Grasso G, Sfacteria A, Xie Q-W, Coleman T, Kreilgaard M, Torup L, Sager T, Erbayraktar Z,

Gokmen N, Yilmaz O, Ghezzi P, Villa P, Fratelli M, Casagrande S, Leist M, Helboe L, Gerwein J,

Christensen S, Geist MA, Pedersen LØ, Cerami-Hand C, Wuerth J-P, Cerami A, and Brines M (2003).

Asialoerythropoietin is a nonerythropoietic cytokine with broad neuroprotective activity in vivo.

Proceedings of the National Academy of Sciences 100 (11):6741-6746.

Eszterhas SK, Bouhassira EE, Martin DIK, and Fiering S (2002). Transcriptional interference by

independently regulated genes occurs in any relative arrangement of the genes and is influenced by

chromosomal integration position. Molecular and Cellular Biology 22 (2):469-479.

Europa AF, Gambhir A, Fu PC, and Hu WS (2000). Multiple steady states with distinct cellular metabolism

in continuous culture of mammalian cells. Biotechnology and Bioengineering 67 (1):25-34.

Fenge C, and Lüllau E (2006). Cell culture bioreactors. In Cell culture technology for pharmaceutical and

cell-based therapies, edited by SS Ozturk and WS Hu. New York: CRC.

Fenn JB, Mann M, Meng CK, Wong SF, and Whitehouse CM (1989). Electrospray ionization for mass

spectrometry of large biomolecules. Science 246 (4926):64-71.

Ferrara C, Brünker P, Suter T, Moser S, Püntener U, and Umaña P (2006). Modulation of therapeutic

antibody effector functions by glycosylation engineering: Influence of Golgi enzyme localization

domain and co-expression of heterologous beta1, 4-N-acetylglucosaminyltransferase III and Golgi

alpha-mannosidase II. Biotechnology and Bioengineering 93 (5):851-861.

Ferrara C, Stuart F, Sondermann P, Brünker P, and Umaña P (2006). The carbohydrate at FcγRIIIa Asn-

162. An element required for high affinity binding to non-fucosylated IgG glycoforms. Journal of

Biological Chemistry 281 (8):5032-5036.

Fox SR, Patel UA, Yap MGS, and Wang DIC (2004). Maximizing interferon-gamma production by Chinese

hamster ovary cells through temperature shift optimization: Experimental and modeling.


Froud SJ (1999). The development, benefits and disadvantages of serum-free media. Developments in

Biological Standardization 99:157-166.


75

Fukuta K, Abe R, Yokomatsu T, Kono N, Asanagi M, Omae F, Minowa MT, Takeuchi M, and Makino T

(2000). Remodeling of sugar chain structures of human interferon-γ. Glycobiology 10 (4):421-430.

Fukuta K, Yokomatsu T, Abe R, Asanagi M, and Makino T (2000). Genetic engineering of CHO cells

producing human interferon-γ by transfection of sialyltransferases. Glycoconjugate Journal 17

(12):895-904.

Fussenegger M, Bailey JE, Hauser H, and Mueller PP (1999). Genetic optimization of recombinant

glycoprotein production by mammalian cells Trends in Biotechnology 17 (1):35-42

Gagnon M, Hiller G, Luan Y-T, Kittredge A, DeFelice J, and Drapeau D (2011). High-end pH-controlled

delivery of glucose effectively suppresses lactate accumulation in CHO Fed-batch cultures.


Galili U, Rachmilewitz EA, Peleg A, and Flechner I (1984). A unique natural human IgG antibody with anti-

alpha-galactosyl specificity. The Journal of Experimental Medicine 160 (5):1519-1531.

Gao X, and Huang L (1995). Cationic liposome-mediated gene transfer. Gene Therapy 2 (10):710-722.

Garcia-Casado G, Sanchez-Monge R, Chrispeels MJ, Armentia A, Salcedo G, and Gomez L (1996). Role of

complex asparagine-linked glycans in the allergenicity of plant glycoproteins. Glycobiology 6 (4):471-

477.

Gawlitzek M, Ryll T, Lofgren J, and Sliwkowski MB (2000). Ammonium alters N-glycan structures of

recombinant TNFR-IgG: degradative versus biosynthetic mechanisms. Biotechnology and

Bioengineering 68 (6):637-646.

Gawlitzek M, Valley U, Nimtz M, Wagner R, and Conradt HS (1995). Characterization of changes in the

glycosylation pattern of recombinant proteins from BHK-21 cells due to different culture conditions.

Journal of Biotechnology 42 (2):117-131.

Geigert J (2004). The Challenge of CMC Regulatory Compliance for Biopharmaceuticals. New York:

Springer.

Geng M, Zhang X, Bina M, and Regnier F (2001). Proteomics of glycoproteins based on affinity selection

of glycopeptides from tryptic digests. Journal of Chromatography B: Biomedical Sciences and

Applications 752 (2):293-306.

Gennaro LA, and Salas-Solano O (2008). On-line CE−LIF−MS technology for the direct characterization of

N-Linked glycans from therapeutic antibodies. Analytical Chemistry 80 (10):3838-3845.

Girod P-A, Zahn-Zabal M, and Mermod N (2005). Use of the chicken lysozyme 5prime matrix attachment

region to generate high producer CHO cell lines. Biotechnology and Bioengineering 91 (1):1-11.

Glover DJ, Lipps HJ, and Jans DA (2005). Towards safe, non-viral therapeutic gene expression in humans.

Nature Reviews Genetics 6 (4):299-310.


76

Goh JSY, Zhang P, Chan KF, Lee MM, Lim SF, and Song Z (2010). RCA-I-resistant CHO mutant cells have

dysfunctional GnT I and expression of normal GnT I in these mutants enhances sialylation of

recombinant erythropoietin. Metabolic Engineering 12 (4):360-368.

Goldman MH, James DC, Rendall M, Ison AP, Hoare M, and Bull AT (1998). Monitoring recombinant

human interferon-gamma N-glycosylation during perfused fluidized-bed and stirred-tank batch culture

of CHO cells. Biotechnology and Bioengineering 60 (5):596-607.

Gomord V, and Faye L (2004). Posttranslational modification of therapeutic proteins in plants. Current

Opinion in Plant Biology 7 (2):171-181.

Goochee CF, Gramer MJ, Andersen DC, Bahr JP, and Rasmussen JR (1991). The oligosaccharides of

glycoproteins: bioprocess factors affecting oligosaccharide structure and their effect on glycoprotein

properties. Nature Biotechnology 9 (12):1347-1355.

Graham FL, and van der Eb AJ (1973). A new technique for the assay of infectivity of human adenovirus 5

DNA. Virology 52 (2):456-467.

Gramer MJ, and Goochee CF (1993). Glycosidase activities in chinese hamster ovary cell lysate and cell

culture supernatant. Biotechnology Progress 9 (4):366-373.

Gramer MJ, Goochee CF, Chock VY, Brousseau DT, and Sliwkowski MB (1995). Removal of sialic acid

from a glycoprotein in CHO cell culture supernatant by action of an extracellular CHO cell sialidase.

Bio/Technology 13 (7):692-698.

Griffiths B (2000). Chapter 2. Scaling-up of animal cell cultures. In Animal cell culture: A practical

approach, edited by JRW Masters. Oxford: Oxford University Press.

Guile GR, Wong SYC, and Dwek RA (1994). Analytical and preparative separation of anionic

oligosaccharides by weak anion-exchange high-performance liquid chromatography on an inert

polymer column. Analytical Biochemistry 222 (1):231-235.

Guile GR, Rudd PM, Wing DR, Prime SB, and Dwek RA (1996). A rapid high-resolution high-performance

liquid chromatographic method for separating glycan mixtures and analyzing oligosaccharide

profiles. Analytical Biochemistry 240 (2):210-226.

Hahn TJ, and Goochee CF (1992). Growth-associated glycosylation of transferrin secreted by HepG2 cells.

Journal of Biological Chemistry 267 (33):23982-7.

Haltiwanger RS (2004). Role of glycosylation in development. Annual Review of Biochemistry 73 (1):491-

537.

Hamilton SR, and Gerngross TU (2007). Glycosylation engineering in yeast: the advent of fully humanized

yeast. Current Opinion in Biotechnology 18 (5):387-392.

Harcum SW (2006). Protein glycosylation. In Cell culture technology for pharmaceutical and cell-based

therapies, edited by S Ozturk and W-S Hu. Boca Raton, FL: CRC Press.


77

Hartenbach S, Daoud-El Baba M, Weber W, and Fussenegger M (2007). An engineered L-arginine sensor

of Chlamydia pneumoniae enables arginine-adjustable transcription control in mammalian cells and

mice. Nucleic Acids Research 35 (20):e136.

Hayter PM, Curling EM, Baines AJ, Jenkins N, Salmon I, Strange PG, Tong JM, and Bull AT (1992).

Glucose-limited chemostat culture of chinese hamster ovary cells producing recombinant human

interferon-gamma. Biotechnology and Bioengineering 39 (3):327-335.

Hayter PM, Curling EMA, Gould ML, Baines AJ, Jenkins N, Salmon I, Strange PG, and Bull AT (1993). The

effect of the dilution rate on CHO cell physiology and recombinant interferon-γ production in glucose-

limited chemostat culture. Biotechnology and Bioengineering 42 (9):1077-1085.

Helenius A, and Aebi M (2001). Intracellular functions of N-glycans. Science 291:2364-2369.

Helenius J, and Aebi M (2002). Transmembrane movement of dolichol linked carbohydrates during N-

glycoprotein biosynthesis in the endoplasmic reticulum. Seminars in Cell & Developmental Biology

13 (3):171-178.

Hellwig S, Drossard J, Twyman RM, and Fischer R (2004). Plant cell cultures for the production of

recombinant proteins. Nature Biotechnology 22 (11):1415-1422.

Hemström P, and Irgum K (2006). Hydrophilic interaction chromatography. Journal of Separation Science

29 (12):1784-1821.

Hillor A (2009). Fast growth foreseen for protein therapeutics. GEN 29:153-155.

Hirano K, Macmillan D, Tezuka K, Tsuji T, and Kajihara Y (2009). Design and synthesis of a

homogeneous erythropoietin analogue with two human complex-type sialyloligosaccharides:

combined use of chemical and bacterial protein expression methods. Angewandte Chemie

International Edition 48 (50):9557-9560.

Hirtenstein M, Clark J, Lindgren G, and Vretblad P (1980). Microcarriers for animal cell culture: a brief

review of theory and practice. Developments in Biological Standardization 46:109-116.

Hodoniczky J, Zheng YZ, and James DC (2005). Control of recombinant monoclonal antibody effector

functions by Fc N-glycan remodeling in vitro. Biotechnology Progress 21 (6):1644-1652.

Holmen S, Vanbrocklin M, Eversole R, Stapleton S, and Ginsberg L (1995). Efficient lipid-mediated

transfection of DNA into primary rat hepatocytes. In Vitro Cellular & Developmental Biology - Animal

31 (5):347-351.

Hossler P (2012). Protein glycosylation control in mammalian cell culture: past precedents and

contemporary prospects. Advances in Biochemical Engineering/Biotechnology 127:187-219.

Hossler P, Khattak SF, and Li ZJ (2009). Optimal and consistent protein glycosylation in mammalian cell

culture. Glycobiology 19 (9):936-949.

Hsu K-L, and Mahal LK (2006). A lectin microarray approach for the rapid analysis of bacterial glycans.

Nature Protocols 1 (2):543-549.


78

Huang MTF, and Gorman CM (1990). Intervening sequences increase efficiency of RNA 3' processing and

accumulation of cytoplasmic RNA. Nucleic Acids Research 18 (4):937-947.

Huang Y, Li Y, Wang YG, Gu X, Wang Y, and Shen BF (2007). An efficient and targeted gene integration

system for high-level antibody expression. Journal of Immunological Methods 322 (1-2):28-39.

Huang Y-M, Hu W, Rustandi E, Chang K, Yusuf-Makagiansar H, and Ryll T (2010). Maximizing productivity

of CHO cell-based fed-batch culture using chemically defined media conditions and typical

manufacturing equipment. Biotechnology Progress 26 (5):1400-1410.

Huang Y, Mechref Y, and Novotny MV (2001). Microscale nonreductive release of O-linked glycans for

subsequent analysis through MALDI mass spectrometry and capillary electrophoresis. Analytical

Chemistry 73 (24):6063-6069.

Huggett B, Hodgson J, and Lahteenmaki R (2011). Public biotech 2010 - the numbers. Nature


Huhn C, Selman MHJ, Ruhaak LR, Deelder AM, and Wuhrer M (2009). IgG glycosylation analysis.

Proteomics 9 (4):882-913.

Iida S, Misaka H, Inoue M, Shibata M, Nakano R, Yamane-Ohnuki N, Wakitani M, Yano K, Shitara K, and

Satoh M (2006). Nonfucosylated therapeutic IgG1 antibody can evade the inhibitory effect of serum

immunoglobulin G on Antibody-Dependent Cellular Cytotoxicity through its high binding to FcγRIIIa.

Clinical Cancer Research 12 (9):2879-2887.

Jain E, and Kumar A (2008). Upstream processes in antibody production: evaluation of critical

parameters. Biotechnology Advances 26 (1):46-71.

James DC, and Baker KN (2002). Glycosylation of Recombinant Proteins. In Encyclopedia of Bioprocess

Technology: John Wiley & Sons, Inc.

Jayapal K, Wlaschin K, Hu WS, and Yap M (2007). Recombinant protein therapeutics from CHO cells - 20

years and counting. CHO Consortium: SBE Special Edition:40-47.

Jefferis R (2006). Criteria for selection of IgG isotype and glycoform of antibody therapeutics. BioProcess

International 4 (9):40-43.

Jefferis R, Lund J, and Pound J (1998). IgG-Fc-mediated effector functions: molecular definition of

interaction sites for effector ligands and the role of glycosylation. Immunological Reviews 163:59-76.

Jefferis R (2007). Antibody therapeutics: isotype and glycoform selection. Expert Opinion on Biological

Therapy 7 (9):1401-1413.

Jefferis R (2009). Glycosylation as a strategy to improve antibody-based therapeutics. Nature Reviews

Drug Discovery 8 (3):226-234.

Jefferis R, and Lund J (2002). Interaction sites on human IgG-Fc for FcγR: current models. Immunology

Letters 82 (1–2):57-65.


79

Jefferis R (2005). Glycosylation of recombinant antibody therapeutics. Biotechnology Progress 21 (1):11-

16.

Jefferis R (2009). Recombinant antibody therapeutics: the impact of glycosylation on mechanisms of

action. Trends in Pharmacological Sciences 30 (7):356-362.

Jeffs S (2007). Expression of recombinant biomedical products from continuous mammalian cell lines. In

Medicines from Animal Cell Culture: John Wiley & Sons, Ltd.

Jenkins N, and Curling EM (1994). Glycosylation of recombinant proteins: problems and prospects.

Enzyme and Microbial Technology 16 (5):354-364.

Jenkins N, Parekh RB, and James DC (1996). Getting the glycosylation right: implications for the

biotechnology industry. Nature Biotechnology 14 (8):975-981.

Jeong YT, Choi O, Lim HR, Son YD, Kim HJ, and Kim JH (2008). Enhanced sialylation of recombinant

erythropoietin in CHO cells by human glycosyltransferase expression. Journal of Microbiology and

Biotechnology 18 (12):1945.

Jin C, Altmann F, Strasser R, Mach L, Schähs M, Kunert R, Rademacher T, Glössl J, and Steinkellner H

(2008). A plant-derived human monoclonal antibody induces an anti-carbohydrate immune response

in rabbits. Glycobiology 18 (3):235-241.

Jones AJS, Papac DI, Chin EH, Keck R, Baughman SA, Lin YS, Kneer J, and Battersby JE (2007).

Selective clearance of glycoforms of a complex glycoprotein pharmaceutical caused by terminal N-

acetylglucosamine is similar in humans and cynomolgus monkeys. Glycobiology 17 (5):529-540.

Jones D, Kroos N, Anema R, Van Montfort B, Vooys A, van der Kraats S, van der Helm E, Smits S,

Schouten J, Brouwer K, Lagerwerf F, Van Berkel P, Opstelten D-J, Logtenberg T, and Bout A (2003).

High-level expression of recombinant IgG in the human cell line PER.C6. Biotechnology Progress 19

(1):163-168.

Jordan M, Schallhorn S, and Wurm FM (1996). Transfecting mammalian cells: optimization of critical

parameters affecting calcium-phosphate precipitate formation. Nucleic Acids Research 24 (4):596-

601.

Kamoda S, and Kakehi K (2006). Capillary electrophoresis for the analysis of glycoprotein

pharmaceuticals. Electrophoresis 27 (12):2495-2504.

Kanda Y, Yamada T, Mori K, Okazaki A, Inoue M, Kitajima-Miyama K, Kuni-Kamochi R, Nakano R, Yano K,

Kakita S, Shitara K, and Satoh M (2007). Comparison of biological activity among nonfucosylated

therapeutic IgG1 antibodies with three different N-linked Fc oligosaccharides: the high-mannose,

hybrid, and complex types. Glycobiology 17 (1):104-118.

Kaneko Y, Nimmerjahn F, and Ravetch JV (2006). Anti-Inflammatory activity of immunoglobulin G

resulting from Fc sialylation. Science 313 (5787):670-673.


80

Kang P, Mechref Y, and Novotny MV (2008). High-throughput solid-phase permethylation of glycans prior

to mass spectrometry. Rapid Communications in Mass Spectrometry 22 (5):721-734.

Kaufman RJ (1990). Selection and coamplification of heterologous genes in mammalian cells. Methods in

Enzymology 185:537-566.

Kaufmann H, and Fussenegger M (2003). Metabolic engineering of mammalian cells for higher protein

yield. In Gene transfer and expression in mammalian cells, edited by S. C. Makrides. Netherlands:

Elsevier Science B.V.

Keown WA, Campbell CR, and Kucherlapati RS (1990). Methods for introducing DNA into mammalian

cells. Methods in Enzymology 185:527-537.

Kibe T, Fujimoto S, Ishida C, Togari H, Wada Y, Okada S, Nakagawa H, Tsukamoto Y, and Takahashi N

(1996). Glycosylation and placental transport of immunoglobulin G. Journal of Clinical Biochemistry

and Nutrition 21:57-63.

Kim NS, Kim SJ, and Lee GM (1998). Clonal variability within dihydrofolate reductase-mediated gene

amplified Chinese hamster ovary cells : Stability in the absence of selective pressure. Biotechnology

and bioengineering 60 (6):679-688.

Kim TK, and Eberwine JH (2010). Mammalian cell transfection: the present and the future. Analytical and

Bioanalytical Chemistry 397:3173-3178.

Kim YK, Shin HS, Tomiya N, Lee YC, Betenbaugh MJ, and Cha HJ (2005). Production and N-glycan

analysis of secreted human erythropoietin glycoprotein in stably transfected Drosophila S2 cells.


Kimura R, and Miller WM (1997). Glycosylation of CHO-derived recombinant tPA produced under elevated

pCO2. Biotechnology Progress 13 (3):311-317.

Kingston RE (2003). Chapter 9. Introduction of DNA into mammalian cells. In Current Protocols in

Molecular Biology. New Jersey: John Wiley & Sons.

Kirkpatrick RB, Ganguly S, Angelichio M, Griego S, Shatzman A, Silverman C, and Rosenberg M (1995).

Heavy chain dimers as well as complete antibodies are efficiently formed and secreted from

Drosophila via a BiP-mediated pathway. Journal of Biological Chemistry 270 (34):19800-19805.

Knibbs RN, Dame M, Allen MR, Ding Y, Hillegas WJ, Varani J, and Stoolman LM (2003). Sustained high-

yield production of recombinant proteins in transiently transfected COS-7 cells grown on

trimethylamine-coated (Hillex) microcarrier beads. Biotechnology Progress 19 (1):9-13.

Kompella A, and Lee V (1991). Pharmacokinetics of peptide and protein drugs. In Peptide and Protein

Drug Delivery, edited by V Lee. New York: Marcel Dekker.

Kontoravdi C, Asprey SP, Pistikopoulos EN, and Mantalaris A (2007). Development of a dynamic model of

monoclonal antibody production and glycosylation for product quality monitoring. Computers and

Chemical Engineering 31 (5-6):392-400.


81

Kornfeld R, and Kornfeld S (1985). Assembly of asparagine-linked oligosaccharides. Annual Review of


Kost TA, Condreay JP, and Jarvis DL (2005). Baculovirus as versatile vectors for protein expression in

insect and mammalian cells. Nature Biotechnology 23 (5):567-575.

Kozak M (1995). Adherence to the first-AUG rule when a second AUG codon follows closely upon the first.

Proceedings of the National Academy of Sciences of the United States of America 92 (15):7134-.

Kozak M (1986). Point mutations define a sequence flanking the AUG initiator codon that modulates

translation by eukaryotic ribosomes. Cell 44 (2):283-292.

Kumpel BM, Rademacher TW, Rook GA, Williams PJ, and Wilson IB (1994). Galactosylation of human IgG

monoclonal anti-D produced by EBV-transformed B-lymphoblastoid cell lines is dependent on culture

method and affects Fc receptor-mediated functional activity. Human Antibodies and Hybridomas 5

(3-4):143-151.

Kunes YZ, Gion WR, Fung E, Salfeld JG, Zhu R-R, Sakorafas P, and Carson GR (2009). Expression of

antibodies using single-open reading frame vector design and polyprotein processing from

mammalian cells. Biotechnology Progress 25 (3):735-744.

Kunkel JP, Jan DC, Butler M, and Jamieson JC (2000). Comparisons of the glycosylation of a monoclonal

antibody produced under nominally identical cell culture conditions in two different bioreactors.


Kunkel JP, Jan DC, Jamieson JC, and Butler M (1998). Dissolved oxygen concentration in serum-free

continuous culture affects N-linked glycosylation of a monoclonal antibody. Journal of Biotechnology

62 (1):55-71.

Küster B, Krogh TN, Mørtz E, and Harvey DJ (2001). Glycosylation analysis of gel-separated proteins.

Proteomics 1 (2):350-361.

Kwaks THJ, Barnett P, Hemrika W, Siersma T, Sewalt RGAB, Satijn DPE, Brons JF, van Blokland R,

Kwakman P, Kruckeberg AL, Kelder A, and Otte AP (2003). Identification of anti-repressor elements

that confer high and stable protein production in mammalian cells. Nature Biotechnology 21 (5):553-

558.

Lam JS, Huang H, and Levitz SM (2007). Effect of differential N-linked and O-linked mannosylation on

recognition of fungal antigens by dendritic cells. PLoS ONE 2 (10):e1009.

Lao M-S, and Toth D (1997). Effects of Ammonium and Lactate on Growth and Metabolism of a

Recombinant Chinese Hamster Ovary Cell Culture. Biotechnology Progress 13 (5):688-691.

Larsen MR, Højrup P, and Roepstorff P (2005). Characterization of gel-separated glycoproteins using two-

step proteolytic digestion combined with sequential microcolumns and mass spectrometry.

Molecular and Cellular Proteomics 4 (2):107-119.


82

Leader, Benjamin, Quentin J. Baca, and David E. Golan. 2008. Protein therapeutics: a summary and

pharmacological classification. Nature Reviews Drug Discovery 7 (1):21-39.

LeFloch F, Tessier B, Chenuet S, Guillaume J-M, Cans P, Goergen J-L, and Marc A (2006). Related effects

of cell adaptation to serum-free conditions on murine EPO production and glycosylation by CHO

cells. Cytotechnology 52 (1):39-53.

Leitzgen K, Knittler MR, and Haas IG (1997). Assembly of immunoglobulin light chains as a prerequisite

for secretion. A model for oligomerization-dependent subunit folding. Journal of Biological Chemistry

272 (5):3117-3123.

Li F, Vijayasankaran N, Shen A, Kiss R, and Amanullah A (2010). Cell culture processes for monoclonal

antibody production. mAbs 2 (5):466-479.

Li F, Zhou JX, Yang X, Tressel T, and Lee B (2005). Current therapeutic antibody production and process

optimization. BioProcessing Journal 4 (5).

Li H, and d’Anjou M (2009). Pharmacological significance of glycosylation in therapeutic proteins. Current


Li H, Sethuraman N, Stadheim TA, Zha D, Prinz B, Ballew N, Bobrowicz P, Choi BK, Cook WJ, Cukan M,

Houston-Cummings NR, Davidson R, Gong B, Hamilton SR, Hoopes JP, Jiang Y, Kim N, Mansfield R,

Nett JH, Rios S, Strawbridge R, Wildt S, and Gerngross TU (2006). Optimization of humanized IgGs

in glycoengineered Pichia pastoris. Nature Biotechnology 24 (2):210-215.

Li J, Menzel C, Meier D, Zhang C, Dübel S, and Jostock T (2007). A comparative study of different vector

designs for the mammalian expression of recombinant IgG antibodies. Journal of Immunological

Methods 318 (1-2):113-124.

Liang M, Guttieri M, Lundkvist Å, and Schmaljohn C (1997). Baculovirus expression of a human G2-

specific, neutralizing IgG monoclonal antibody to Puumala Virus. Virology 235 (2):252-260.

Liao MMl, and Sunstrom N-A (2006). A transient expression vector for recombinant protein production in

Chinese hamster ovary cells. Journal of Chemical Technology and Biotechnology 81 (1):82-88.

Lifely MR, Hale C, Boyce S, Keen MJ, and Phillips J (1995). Glycosylation and biological activity of

CAMPATH-1H expressed in different cell lines and grown under different culture conditions.

Glycobiology 5 (8):813-822.

Lim Y, Wong NSC, Lee YY, Ku SCY, Wong DCF, and Yap MGS (2010). Engineering mammalian cells in

bioprocessing – current achievements and future perspectives. Biotechnology and Applied

Biochemistry 55:175-189.

Lin Z, Lo A, Simeone DM, Ruffin MT, and Lubman DM (2012). An N-glycosylation analysis of human

alpha-2-macroglobulin using an integrated approach. Journal of Proteomics and Bioinformatics 5

(5):124-134.


83

Lipscomb ML, Palomares LA, Hernández V, Ramírez OT, and Kompala DS (2005). Effect of production

method and gene amplification on the glycosylation pattern of a secreted reporter protein in CHO

cells. Biotechnology Progress 21 (1):40-49.

Lis H, and Sharon N (1993). Protein glycosylation. European Journal of Biochemistry 218 (1):1-27.

Liu C-H, and Chang T-Y (2006). Rational development of serum-free medium for Chinese hamster ovary

cells. Process Biochemistry 41 (11):2314-2319.

Liu C-H, Chu IM, and Hwang S-M (2001). Factorial designs combined with the steepest ascent method to

optimize serum-free media for CHO cells. Enzyme and Microbial Technology 28 (4–5):314-321.

Lloyd DR, Holmes P, Jackson LP, Emery AN, and Al-Rubeai M (2000). Relationship between cell size, cell

cycle and specific recombinant protein productivity. Cytotechnology 34 (1):59-70.

Lowe JB, and Marth JD (2003). A genetic approach to mammalian glycan function. Annual Review of


Lund J, Takahashi N, Nakagawa H, Goodall M, Bentley T, Hindley SA, Tyler R, and Jefferis R (1993).

Control of IgG/Fc glycosylation: a comparison of oligosaccharides from chimeric human/mouse and

mouse subclass immunoglobulin Gs. Molecular Immunology 30 (8):741-748.

Luong M, Lam JS, Chen J, and Levitz SM (2007). Effects of fungal N- and O-linked mannosylation on the

immunogenicity of model vaccines. Vaccine 25 (22):4340-4344.

Ma HM, and Chen G (2005). Gene transfer technique. Nature and Science 3 (1):25-31.

Manzi AE, and Varki A (1993). Compositional analysis of glycoproteins. In Glycobiology: a practical

approach, edited by M Fukuda and A Kobata. Oxford, United Kingdom: IRI Press.

Marchal I, Jarvis DL, Cacan R, and Verbert A (2001). Glycoproteins from insect cells: sialylated or not?

Biological Chemistry 382 (2):151-159.

Marino K, Bones J, Kattla JJ, and Rudd PM (2010). A systematic approach to protein glycosylation

analysis: a path through the maze. Nature Chemical Biology 6 (10):713-723.

Marth JD (1999). N-Glycans. In Essentials of Glycobiology, edited by A Varki, R Cummings, J Esko, H

Freeze, G Hart and J Marth. Cold Spring Harbor (NY): Cold Spring Harbor Laboratory Press.

Masuda K, Kubota T, Kaneko E, Iida S, Wakitani M, Kobayashi-Natsume Y, Kubota A, Shitara K, and

Nakamura K (2007). Enhanced binding affinity for FcγRIIIa of fucose-negative antibody is sufficient to

induce maximal antibody-dependent cellular cytotoxicity. Molecular Immunology 44 (12):3122-3131.

Matsumiya S, Yamaguchi Y, Saito J, Nagano M, Sasakawa H, Otaki S, Satoh M, Shitara K, and Kato J

(2007). Structural comparison of fucosylated and nonfucosylated Fc fragments of human

immunoglobulin G1. Journal of Molecular Biology 368 (3):767-779.

Melton DW, Ketchen AM, Kind AJ, McEwan C, Paisley D, and Selfridge J (2001). A one-step gene

amplification system for use in cultured mammalian cells and transgenic animals. Transgenic

Research 10 (2):133-142.


84

Merry AH, Neville DCA, Royle L, Matthews B, Harvey DJ, Dwek RA, and Rudd PM (2002). Recovery of

intact 2-aminobenzamide-labeled O-glycans released from glycoproteins by hydrazinolysis. Analytical


Merry TH, Royle L, Radcliffe CM, Astrautsova SA, Dwek RA, and Rudd PM (2004). Analysis of N- and O-

linked glycans of glycoproteins. In Encyclopedia of Genetics, Genomics, Proteomics and

Bioinformatics: John Wiley & Sons, Ltd.

Mijakovic I, Petranovic D, and Jensen PR (2005). Tunable promoters in systems biology. Current Opinion

in Biotechnology 16 (3):329-335.

Mimura Y, Jefferis R, Mimura-Kimura Y, Abrahams J, and Rudd PM (2009). Glycosylation of therapeutic

IgGs. In Therapeutic Monoclonal Antibodies: John Wiley & Sons, Inc.

Mimura Y, Lund J, Church S, Dong S, Li J, Goodall M, and Jefferis R (2001). Butyrate increases

production of human chimeric IgG in CHO-K1 cells whilst maintaining function and glycoform profile.

Journal of Immunological Methods 247 (1–2):205-216.

Mizuochi T, Taniguchi T, Shimizu A, and Kobata A (1982). Structural and numerical variations of the

carbohydrate moiety of immunoglobulin G. The Journal of Immunology 129 (5):2016-20.

Mohan CM, Kim Y-G, Koo J, and Lee GM (2008). Assessment of cell engineering strategies for improved

therapeutic protein production in CHO cells. Biotechnology Journal 3 (5):624-630.

Moore A, Mercer J, Dutina G, Donahue CJ, Bauer KD, Mather JP, Etcheverry T, and Ryll T (1997). Effects

of temperature shift on cell cycle, apoptosis and nucleotide pools in CHO cell batch cultues.

Cytotechnology 23 (1):47-54.

Morell AG, Gregoriadis G, Scheinberg IH, Hickman J, and Ashwell G (1971). The role of sialic acid in

determining the survival of glycoproteins in the circulation. Journal of Biological Chemistry 246

(5):1461-1467.

Morgan A, Jones ND, Nesbitt AM, Chaplin L, Bodmer MW, and Emtage JS (1995). The N-terminal end of

the CH2 domain of chimeric human IgG1 anti-HLA-DR is necessary for C1q, Fc gamma RI and Fc

gamma RIII binding. Immunology 86 (2):319-324.

Mori K, Kuni-Kamochi R, Yamane-Ohnuki N, Wakitani M, Yamano K, Imai H, Kanda Y, Niwa R, Iida S,

Uchida K, Shitara K, and Satoh M (2004). Engineering Chinese hamster ovary cells to maximize

effector function of produced antibodies using FUT8 siRNA. Biotechnology and Bioengineering 88

(7):901-908.

Mortensen RM, and Kingston RE (2009). Selection of transfected mammalian cells. In Current Protocols

in Molecular Biology: John Wiley & Sons, Inc.

Müthing J, Kemminer SE, Conradt HS, Sagi D, Nimtz M, Kärst U, and Peter-Katalinić J (2003). Effects of

buffering conditions and culture pH on production rates and glycosylation of clinical phase I anti-


85

melanoma mouse IgG3 monoclonal antibody R24. Biotechnology and Bioengineering 83 (3):321-

334.

Nabi IR, and Dennis JW (1998). The extent of polylactosamine glycosylation of MDCK LAMP-2 is

determined by its Golgi residence time. Glycobiology 8 (9):947-953.

Nahrgang S, Kragten E, De Jesus M, Bourgeois M, Dejardin S, Von Stockar U, and Marison IW (1999).

The effect of cell line, transfection procedure and reactor conditions on the glycosylation of

recombinant human anti-rhesus D IGG1. In Animal Cell Technology: Products from Cells, Cells as

Products. Proceedings of the ESACT Meeting, 16th, Lugano, Switzerland, Apr. 25-29, edited by A

Bernard.

Narhi LO, Arakawa T, Aoki KH, Elmore R, Rohde MF, Boone T, and Strickland TW (1991). The effect of

carbohydrate on the structure and stability of erythropoietin. The Journal of Biological Chemistry 266

(34):23022-23026.

Natsume A, Niwa R, and Satoh M (2009). Improving effector functions of antibodies for cancer treatment:

enhancing ADCC and CDC. Drug Design, Development and Therapy 3:7-16.

Nehlsen K, Schucht R, da Gama-Norton L, Kromer W, Baer A, Cayli A, Hauser H, and Wirth D (2009).

Recombinant protein expression by targeting pre-selected chromosomal loci. BMC Biotechnology 9

(1):100.

Niklas J, Schräder E, Sandig V, Noll T, and Heinzle E (2011). Quantitative characterization of metabolism

and metabolic shifts during growth of the new human cell line AGE1.HN using time resolved

metabolic flux analysis. Bioprocess and Biosystems Engineering 34 (5):533-545.

Nilsang S, Nandakumar KS, Galaev IY, Rakshit SK, Holmdahl R, Mattiasson B, and Kumar A (2007).

Monoclonal antibody production using a new supermacroporous cryogel bioreactor. Biotechnology

Progress 23 (4):932-939.

Nimmerjahn F, and Ravetch JV (2008). Fcγ receptors as regulators of immune responses. Nature Reviews

Immunology 8 (1):34-47.

Niwa R, Sakurada M, Kobayashi Y, Uehara A, Matsushima K, Ueda R, Nakamura K, and Shitara K (2005).

Enhanced natural killer cell binding and activation by low-fucose IgG1 antibody results in potent

Antibody-Dependent Cellular Cytotoxicity induction at lower antigen density. Clinical Cancer Research

11 (6):2327-2336.

Noe W, Bux R, Werner RG, and Berthold W (1992). Aspects of continuous culture versus batch mode for

animal cells in industrial processes. In Animal cell technology: Developments, processes and

products, edited by RE Spier, JB Griffiths and C MacDonald. Oxford: Butterworth-Heinemann.

Nyberg GB, Balcarcel RR, Follstad BD, Stephanopoulos G, and Wang DI (1999). Metabolic effects on

recombinant interferon-gamma glycosylation in continuous culture of Chinese hamster ovary cells.



86

Oguchi S, Saito H, Tsukahara M, and Tsumura H (2006). pH condition in temperature shift cultivation

enhances cell longevity and specific hMab productivity in CHO culture. Cytotechnology 52 (3):199-

207.

Omasa T, Tanaka R, Doi T, Ando M, Kitamoto Y, Honda K, Kishimoto M, and Ohtake H (2008). Decrease

in antithrombin III fucosylation by expressing GDP-fucose transporter siRNA in Chinese hamster

ovary cells. Journal of Bioscience and Bioengineering 106 (2):168-173.

Osborne C, and Brooks SA (2006). SDS-PAGE and Western blotting to detect proteins and glycoproteins of

interest in breast cancer research. Methods in Molecular Medicine 120:217-229.

Özdemir Ö (1998). The Construction of a mammalian transfection vector for expression of cytosine-5

specific dna methyltransferase gene M.Msp1 in cultured cells. Turkish Journal of Biology 22:161-

170.

Ozturk S, Kasebo G, Mahaworasilpa T, and Coster HG (2003). Adaptation of cell lines to serum-free

culture medium. Hybridoma and Hybridomics 22 (4):267-272.

Ozturk SS, and Hu W-S (2006). Cell Culture Technology for Pharmaceutical and Cell-Based Therapies.

New York: CRC Press.

Padler-Karavani V, Yu H, Cao H, Chokhawala H, Karp F, Varki N, Chen X, and Varki A (2008). Diversity in

specificity, abundance, and composition of anti-Neu5Gc antibodies in normal humans: potential

implications for disease. Glycobiology 18 (10):818-830.

Page MJ, and Sydenham MA (1991). High level expression of the humanized monoclonal antibody

CAMPATH-1H in Chinese hamster ovary cells. Nature Biotechnology 9 (1):64-68.

Parodi AJ (2000). Protein glucosylation and its role in protein folding. Annual Review of Biochemistry

69:69-93.

Patel TP, Parekh RB, Moellering BJ, and Prior CP (1992). Different culture methods lead to differences in

glycosylation of a murine IgG monoclonal antibody. Biochemical Journal 285:839-845.

Patton WF (2001). Fluorescence detection of glycoproteins in gels and on electroblots. In Current

Protocols in Cell Biology: John Wiley & Sons, Inc.

Patwa T, Li C, Simeone DM, and Lubman DM (2010). Glycoprotein analysis using protein microarrays and

mass spectrometry. Mass Spectrometry Reviews 29 (5):830-844.

Perlman S, van den Hazel B, Christiansen J, Gram-Nielsen S, Jeppesen CB, Andersen KV, Halkier T,

Okkels S, and Schambye HT (2003). Glycosylation of an N-terminal extension prolongs the half-life

and increases the in vivo activity of follicle stimulating hormone. Journal of Clinical Endocrinology &

Metabolism 88 (7):3227-3235.

Peterman S, and Mulholland J (2006). A novel approach for identification and characterization of

glycoproteins using a hybrid linear ion trap/FT-ICR mass spectrometer. Journal of The American

Society for Mass Spectrometry 17 (2):168-179.


87

Pfeifer A, and Verma IM (2001). Gene terapy: promises and problems. Annual Review of Genomics and

Human Genetics 2 (1):177-211.

Pierce LN, and Shabram PW (2004). Scalability of a disposable bioreactor from 25L - 500L run in

perfusion mode with CHO-based cell line: A tech review. BioProcessing Journal 3 (4):51-56.

Pilobello KT, Slawek DE, and Mahal LK (2007). A ratiometric lectin microarray approach to analysis of the

dynamic mammalian glycome. Proceedings of the National Academy of Sciences 104 (28):11534-

11539.

Potter H, and Heller R (2010). Transfection by electroporation. In Current Protocols in Molecular Biology:

John Wiley & Sons, Inc.

Preithner S, Elm S, Lippold S, Locher M, Wolf A, da Silva AJ, Baeuerle PA, and Prang NS (2006). High

concentrations of therapeutic IgG1 antibodies are needed to compensate for inhibition of antibody-

dependent cellular cytotoxicity by excess endogenous immunoglobulin G. Molecular Immunology 43

(8):1183-1193.

Raju TS (2008). Terminal sugars of Fc glycans influence antibody effector functions of IgGs. Current

Opinion in Immunology 20 (4):471-478.

Raju TS (2003). Glycosylation variations with expression systems and their impact on biological activity of

therapeutic immunoglobulins. BioProcess International 1 (4):44-53.

Raju TS, and Scallon B (2007). Fc glycans terminated with N-acetylglucosamine residues increase

antibody resistance to papain. Biotechnology Progress 23 (4):964-971.

Rearick JI, Chapman A, and Kornfeld S (1981). Glucose starvation alters lipid-linked oligosaccharide

biosynthesis in Chinese hamster ovary cells. Journal of Biological Chemistry 256 (12):6255-6261.

Recillas-Targa F (2006). Multiple strategies for gene transfer, expression, knockdown, and chromatin

influence in mammalian cell lines and transgenic animals. Molecular Biotechnology 34 (3):337-354.

Reff ME, Hariharan K, and Braslawsky G (2002). Future of monoclonal antibodies in the treatment of

hematologic malignancies. Cancer Control 9 (2):152-166.

Restelli V, and Butler M (2002). The effect of cell culture parameters on protein glycosylation. In

Glycosylation, edited by M Al-Rubeai. Dordrecht: Kluwer Publishers.

Reuter G, and Gabius HJ (1999). Eukaryotic glycosylation: whim of nature or multipurpose tool? Cellular

and Molecular Life Sciences 55 (3):368-422.

Rodriguez J, Spearman M, Huzel N, and Butler M (2005). Enhanced production of monomeric interferon-

beta by CHO cells through the control of culture conditions. Biotechnology Progress 21 (1):22-30.

Rose DW (2007). Genetic manipulation of mammalian cells by microinjection. Cold Spring Harbor

Protocols 2007 (5):pdb.prot4754.


88

Rosser MP, Xia W, Hartsell S, McCaman M, Zhu Y, Wang S, Harvey S, Bringmann P, and Cobb RR (2005).

Transient transfection of CHO-K1-S using serum-free medium in suspension: a rapid mammalian

protein expression system Protein Expression and Purification 40 (2):237-243.

Roth Z, Yehezkel G, and Khalaila I (2012). Identification and quantification of protein glycosylation.

International Journal of Carbohydrate Chemistry 2012:1-10.

Rothman RJ, Warren L, Vliegenthart JFG, and Hard KJ (1989). Clonal analysis of the glycosylation of

immunoglobulin G secreted by murine hybridomas. Biochemistry 28 (3):1377-1384.

Royle L, Dwek RA, and Rudd PM (2001). Determining the structure of oligosaccharides N- and O-linked to

glycoproteins. In Current Protocols in Protein Science: John Wiley & Sons, Inc.

Royle L, Mattu TJ, Hart E, Langridge JI, Merry AH, Murphy N, Harvey DJ, Dwek RA, and Rudd PM (2002).

An analytical and structural database provides a strategy for sequencing o-glycans from microgram

quantities of glycoproteins. Analytical Biochemistry 304 (1):70-90.

Royle L, Radcliffe CM, Dwek RA, and Rudd PM (2007). Detailed structural analysis of N-glycans released

from glycoproteins in SDS-PAGE gel bands using HPLC combined with exoglycosidase array

digestions. In Glycobiology protocols, edited by I Brockhausen: Humana Press.

Rudolph G, Lindner P, Gierse A, Bluma A, Martinez G, Hitzmann B, and Scheper T (2008). Online

monitoring of microcarrier based fibroblast cultivations with in situ microscopy. Biotechnology and


Runkel L, Meier W, Pepinsky RB, Karpusas M, Whitty A, Kimball K, Brickelmaier M, Muldowney C, Jones

W, and Goelz SE (1998). Structural and functional differences between glycosylated and non-

glycosylated forms of human Interferon-β (IFN-β). Pharmaceutical Research 15 (4):641-649.

Ryll T, Dutina G, Reyes A, Gunson J, Krummen L, and Etcheverry T (2000). Performance of small-scale

CHO perfusion cultures using an acoustic cell filtration device for cell retention: characterization of

separation efficiency and impact of perfusion on product quality. Biotechnology and Bioengineering

69 (4):440-449.

Sandadi S, Ensari S, and Kearns B (2005). Heuristic optimization of antibody production by Chinese

hamster ovary cells. Biotechnology Progress 21 (5):1537-1542.

Sareneva T, Cantell K, Pyhälä L, Pirhonen J, and Julkunen I (1993). Effect of carbohydrates on the

pharmacokinetics of human interferon-gamma. Journal of Interferon Research 13 (4):267-269.

Sareneva T, Mørtz E, Tölö H, Roepstorff P, and Julkunen I (1996). Biosynthesis and N-glycosylation of

human Interferon-γ. European Journal of Biochemistry 242 (2):191-200.

Sasaki H, Bothner B, Dell A, and Fukuda M (1987). Carbohydrate structure of erythropoietin expressed in

Chinese hamster ovary cells by a human erythropoietin cDNA. Journal of Biological Chemistry 262

(25):12059-12076.


89

Satoh M, Iida S, and Shitara K (2006). Non-fucosylated therapeutic antibodies as next-generation

therapeutic antibodies. Expert Opinion on Biological Therapy 6 (11):1161-1173.

Scallon BJ, Tam SH, McCarthy SG, Cai AN, and Raju TS (2007). Higher levels of sialylated Fc glycans in

immunoglobulin G molecules can adversely impact functionality. Molecular Immunology 44

(7):1524-1534.

Schäffer C, Graninger M, and Messner P (2001). Prokaryotic glycosylation. Proteomics 1 (2):248-261.

Schenborn ET, and Goiffon V (2000). DEAE-dextran transfection of mammalian cultured cells. Methods in

Molecular Biology 130:147-153.

Schiedner G, Hertel S, Bialek C, Kewes H, Waschutza G, and Volpers C (2008). Efficient and reproducible

generation of high-expressing, stable human cell lines without need for antibiotic selection. BMC

Biotechnology 8 (1):13.

Schlatter S, Stansfield SH, Dinnis DM, Racher AJ, Birch JR, and James DC (2005). On the optimal ratio of

heavy to light chain genes for efficient recombinant antibody production by CHO cells. Biotechnology

Progress 21 (1):122-133.

Schräder E, Scholz S, Niklas J, Rath A, Barradas O, Jandt U, Sandig V, Rose T, Pörtner R, Reichl U, Zeng

A-P, Heinzle E, and Noll T (2011). Characterisation of cultivation of the human cell line

AGE1.HN.AAT. BMC Proceedings 5 (0):1-2.

Schumann W, and Ferreira LCS (2004). Production of recombinant proteins in Escherichia coli. Genetics

and Molecular Biology 27:442-453.

Schuster M, Umana P, Ferrara C, Brünker P, Gerdes C, Waxenecker G, Wiederkum S, Schwager C,

Loibner H, Himmler G, and Mudde GC (2005). Improved effector functions of a therapeutic

monoclonal Lewis Y-specific antibody by glycoform engineering. Cancer Research 65 (17):7934-

7941.

Sears P, and Wong CH (1998). Enzyme action in glycoprotein synthesis. Cellular and Molecular Life

Sciences 54 (3):223-252.

Serrato JA, Palomares LA, Meneses-Acosta A, and Ramírez OT (2004). Heterogeneous conditions in

dissolved oxygen affect N-glycosylation but not productivity of a monoclonal antibody in hybridoma

cultures. Biotechnology and Bioengineering 88 (2):176-188.

Seth G, Hossler, Yee J, and Hu W-S (2006). Engineering Cells for Cell Culture Bioprocessing –

Physiological Fundamentals, Cell Culture Engineering, edited by W-S Hu: Springer Berlin/Heidelberg.

Sethuraman N, and Stadheim TA (2006). Challenges in therapeutic glycoprotein production. Current


Sheeley DM, Merrill BM, and Taylor LCE (1997). Characterization of monoclonal antibody glycosylation:

comparison of expression systems and identification of terminal α-linked galactose. Analytical



90

Shelikoff M, Sinskey AJ, and Stephanopoulos G (1994). The effect of protein synthesis inhibitors on the

glycosylation site occupancy of recombinant human prolactin. Cytotechnology 15 (1):195-208.

Shen AY, Van de Goor J, Zheng L, Reyes A, and Krummen LA (2006). Recombinant DNA technology and

cell line development. In Cell culture technology for pharmaceutical and cell-based therapies, edited

by S Ozturk and W-S Hu. Boca Raton, Florida: CRC Press.

Sheridan C (2010). Fresh from the biologic pipeline - 2009. Nature Biotechnology 28 (4):307-310.

Shields RL, Lai J, Keck R, O'Connell LY, Hong K, Meng YG, Weikert SHA, and Presta LG (2002). Lack of

fucose on human IgG1 N-linked oligosaccharide improves binding to human FcγRIII and Antibody-

dependent Cellular Toxicity. Journal of Biological Chemistry 277 (30):26733-26740.

Shinkawa T, Nakamura K, Yamane N, Shoji-Hosaka E, Kanda Y, Sakurada M, Uchida K, Anazawa H,

Satoh M, Yamasaki M, Hanai N, and Shitara K (2003). The absence of fucose but not the presence

of galactose or bisecting N-acetylglucosamine of human IgG1 complex-type oligosaccharides shows

the critical role of enhancing antibody-dependent cellular cytotoxicity. Journal of Biological Chemistry

278 (5):3466-3473.

Sinacore M, Drapeau D, and Adamson S (2000). Adaptation of mammalian cells to growth in serum-free

media. Molecular Biotechnology 15 (3):249-257.

Sinclair AM, and Elliott S (2005). Glycoengineering: the effect of glycosylation on the properties of

therapeutic proteins. Journal of Pharmaceutical Sciences 94 (8):1626-1635.

Smalling R, Foote M, Molineux G, Swanson SJ, and Elliott S (2004). Drug-induced and antibody-mediated

pure red cell aplasia: a review of literature and current knowledge. Biotechnology Annual Review

10:237-250.

Soderquist RG, and Lee JM (2005). Enhanced production of recombinant proteins from plant cells by the

application of osmotic stress and protein stabilization. Plant Cell Reports 24 (2):127-132.

Solá RJ, and Griebenow K (2010). Glycosylation of therapeutic proteins: an effective strategy to optimize

efficacy. BioDrugs 24 (1):9-21.

Solá RJ, and Griebenow K (2009). Effects of glycosylation on the stability of protein pharmaceuticals.

Journal of Pharmaceutical Sciences 98 (4):1223-1245.

Sondermann P, Kaiser J, and Jacob U (2001). Molecular basis for immune complex recognition: a

comparison of Fc-receptor structures. Journal of Molecular Biology 309 (3):737-749.

Spearman M, Rodriguez J, Huzel N, and Butler M (2005). Production and glycosylation of recombinant β-

interferon in suspension and cytopore microcarrier cultures of CHO cells. Biotechnology Progress 21

(1):31-39.

Spiro RG (2002). Protein glycosylation: nature, distribution, enzymatic formation, and disease implications

of glycopeptide bonds. Glycobiology 12 (4):43R-56R.


91

Stiens LR, Büntemeyer H, Lütkemeyer D, Lehmann J, Bergmann A, and Weglöhner W (2000).

Development of serum-free bioreactor production of recombinant human thyroid stimulating

hormone receptor. Biotechnology Progress 16 (5):703-709.

Strasser R, Stadlmann J, Schähs M, Stiegler G, Quendler H, Mach L, Glössl J, Weterings K, Pabst M, and

Steinkellner H (2008). Generation of glyco-engineered Nicotiana benthamiana for the production of

monoclonal antibodies with a homogeneous human-like N-glycan structure. Plant Biotechnology

Journal 6 (4):392-402.

Strohl WR (2009). Therapeutic Monoclonal Antibodies: Past, Present, and Future. In Therapeutic

Monoclonal Antibodies: John Wiley & Sons, Inc.

Strohl WR, and Knight DM (2009). Discovery and development of biopharmaceuticals: current issues.


Sung YH, Song YJ, Lim SW, Chung JY, and Lee GM (2004). Effect of sodium butyrate on the production,

heterogeneity and biological activity of human thrombopoietin by recombinant Chinese hamster

ovary cells. Journal of Biotechnology 112 (3):323-335.

Swiech K, Picanço-Castro V, and Covas DT (2012). Human cells: new platform for recombinant

therapeutic protein production. Protein Expression and Purification 84 (1):147-153.

Tait AS, Brown CJ, Galbraith DJ, Hines MJ, Hoare M, Birch JR, and James DC (2004). Transient

production of recombinant proteins by Chinese hamster ovary cells using polyethyleneimine/DNA

complexes in combination with microtubule disrupting anti-mitotic agents. Biotechnology and


Tarelli E (2007). Glycosylation of Medicinal Products. In Medicines from Animal Cell Culture: John Wiley &

Sons, Ltd.

Tarp MA, and Clausen H (2008). Mucin-type O-glycosylation and its potential use in drug and vaccine

development. Biochimica et Biophysica Acta - General Subjects 1780 (3):546-563.

Taverna M, Tran NT, Merry T, Horvath E, and Ferrier D (1998). Electrophoretic methods for process

monitoring and the quality assessment of recombinant glycoproteins. Electrophoresis 19 (15):2572-

2594.

Taylor ME, and Drickamer K (2006). Introduction to Glycobiology. Oxford, USA: Oxford University Press.

Tharmalingam T, Sunley K, Spearman M, and Butler M (2011). Enhanced production of human

recombinant proteins from CHO cells grown to high densities in macroporous microcarriers.

Molecular Biotechnology 49 (3):263-276.

Townsend R, and Stahl P (1981). Isolation and characterization of a mannose/N-

acetylglucosamine/fucose-binding protein from rat liver. The Biochemical Journal 194 (1):209-214.

Townsend RR, and Hardy MR (1991). Analysis of glycoprotein oligosaccharides using high-pH anion

exchange chromatography. Glycobiology 1 (2):139-147.


92

Tran NT, Taverna M, Merry AH, Chevalier M, Morgant G, Valentin C, and Ferrier D (2000). A sensitive

mapping strategy for monitoring the reproducibility of glycan processing in an HIV vaccine, RGP-160,

expressed in a mammalian cell line. Glycoconjugate Journal 17 (6):401-406.

Tretter V, Altmann F, and März L (1991). Peptide-N4-(N-acetyl-β-glucosaminyl)asparagine amidase F

cannot release glycans with fucose attached α1 → 3 to the asparagine-linked N-acetylglucosamine

residue. European Journal of Biochemistry 199 (3):647-652.

Trill JJ, Shatzman AR, and Ganquly S (1995). Production of monoclonal antibodies in COS and CHO cells.

Current Opinion in Biotechnology 6:553-560.

Trimble RB, and Tarentino AL (1991). Identification of distinct endoglycosidase (endo) activities in

Flavobacterium meningosepticum: endo F1, endo F2, and endo F3. Endo F1 and endo H hydrolyze

only high mannose and hybrid glycans. Journal of Biological Chemistry 266 (3):1646-1651.

Trummer E, Fauland K, Seidinger S, Schriebl K, Lattenmayer C, Kunert R, Vorauer-Uhl K, Weik R, Borth N,

Katinger H, and Müller D (2006). Process parameter shifting: Part I. Effect of DOT, pH, and

temperature on the performance of Epo-Fc expressing CHO cells cultivated in controlled batch

bioreactors. Biotechnology and Bioengineering 94 (6):1033-1044.

Umana P, Jean-Mairet J, Moudry R, Amstutz H, and Bailey JE (1999). Engineered glycoforms of an

antineuroblastoma IgG1 with optimized antibody-dependent cellular cytotoxic activity. Nature


Upreti RK, Kumar M, and Shankar V (2003). Bacterial glycoproteins: functions, biosynthesis and

applications. Proteomics 3 (4):363-379.

Valley U, Nimtz M, Conradt HS, and Wagner R (1999). Incorporation of ammonium into intracellular UDP-

activated N-acetylhexosamines and into carbohydrate structures in glycoproteins. Biotechnology and


van den Eijnden DH, and Joziasse DH (1993). Enzymes associated with glycosylation. Current Opinion in

Structural Biology 3 (5):711-721.

Van den Steen P, Rudd PM, Dwek RA, and Opdenakker G (1998). Concepts and principles of O-linked

glycosylation. Critical Reviews in Biochemistry and Molecular Biology 33 (3):151-208.

van Kasteren SI, Kramer HB, Jensen HH, Campbell SJ, Kirkpatrick J, Oldham NJ, Anthony DC, and Davis

BG (2007). Expanding the diversity of chemical protein modification allows post-translational

mimicry. Nature 446 (7139):1105-1109.

Vervecken W, Kaigorodov V, Callewaert N, Geysens S, De Vusser K, and Contreras R (2004). In vivo

synthesis of mammalian-like, hybrid-type N-glycans in Pichia pastoris. Applied and Environmental

Microbiology 70 (5):2639-2646.


93

Voisard, D., F. Meuwly, P. A. Ruffieux, G. Baer, and A. Kadouri. 2003. Potential of cell retention

techniques for large-scale high-density perfusion culture of suspended mammalian cells.


Vugmeyster Y, and Howell K (2004). Rituximab-mediated depletion of cynomolgus monkey B cells in vitro

in different matrices: possible inhibitory effect of IgG. International Immunopharmacology 4

(8):1117-1124.

Wada Y, Gu J, Okamoto N, and Inui K (1994). Diagnosis of carbohydrate-deficient glycoprotein syndrome

by matrix-assisted laser desorption time-of-flight mass spectrometry. Biological Mass Spectrometry

23 (2):108-109.

Waegeman H, and De Mey M (2012). Increasing recombinant protein production in E. coli by an

alternative method to reduce acetate. In Advances in Applied Biotechnology, edited by M Petre.

InTech.

Wagner-Rousset E, Bednarczyk A, Bussat MC, Colas O, Corvaïa N, Schaeffer C, Van Dorsselaer A, and

Beck A (2008). The way forward, enhanced characterization of therapeutic antibody glycosylation:

comparison of three level mass spectrometry-based strategies. Journal of Chromatography B 872 (1-

2):23-37.

Walsh G (2003). Biopharmaceutical benchmarks 2003. Nature Biotechnology 21 (8):865-870.

Walsh G (2010). Biopharmaceutical benchmarks 2010. Nature Biotechnology 28 (9):917-924.

Walsh G, and Jefferis R (2006). Post-translational modifications in the context of therapeutic proteins.

Nature Biotechnology 24 (10):1241-1252.

Wang A, Nairn NW, Johnson RS, Tirrell DA, and Grabstein K (2008). Processing of N-terminal unnatural

amino acids in recombinant human interferon-β in Escherichia coli. Chembiochem 9 (2):324-330.

Wang L-X (2008). Chemoenzymatic synthesis of glycopeptides and glycoproteins through endoglycosidase-

catalyzed transglycosylation. Carbohydrate Research 343 (10–11):1509-1522.

Wang L-X, and Lomino JV (2011). Emerging technologies for making glycan-defined glycoproteins. ACS

Chemical Biology 7 (1):110-122.

Wang T-Y, Yang R, Qin C, Wang L, and Yang X-J (2008). Enhanced expression of transgene in CHO cells

using matrix attachment region. Cell Biology International 32 (10):1279-1283.

Wang WC, Lee N, Aoki D, Fukuda MN, and Fukuda M (1991). The poly-N-acetyllactosamines attached to

lysosomal membrane glycoproteins are increased by the prolonged association with the Golgi

complex. Journal of Biological Chemistry 266 (34):23185-90.

Wang Y, Wu S-I, and Hancock WS (2006). Monitoring of glycoprotein products in cell culture lysates using

lectin affinity chromatography and capillary HPLC coupled to electrospray linear ion trap-fourier

transform mass spectrometry (LTQ/FTMS). Biotechnology Progress 22 (3):873-880.


94

Washbourne P, and McAllister AK (2002). Techniques for gene transfer into neurons. Current Opinion in

Neurobiology 12 (5):566-573.

Watson E, Shah B, Leiderman L, Hsu Y-R, Karkare S, Lu HS, and Lin F-K (1994). Comparison of N-linked

oligosaccharides of recombinant human tissue kallikrein produced by Chinese hamster ovary cells

on microcarrier beads and in serum-free suspension culture. Biotechnology Progress 10 (1):39-44.

Wei Y, Li C, Huang W, Li B, Strome S, and Wang L-X (2008). Glycoengineering of human IgG1-Fc through

combined yeast expression and in vitro chemoenzymatic glycosylation. Biochemistry 47 (39):10294-

10304.

Weigel PH, and Yik JHN (2002). Glycans as endocytosis signals: the cases of the asialoglycoprotein and

hyaluronan/chondroitin sulfate receptors. Biochimica et Biophysica Acta - General Subjects 1572

(2–3):341-363.

Weikert S, Papac D, Briggs J, Cowfer D, Tom S, Gawlitzek M, Lofgren J, Mehta S, Chisholm V, Modi N,

Eppler S, Carroll K, Chamow S, Peers D, Berman P, and Krummen L (1999). Engineering Chinese

hamster ovary cells to maximize sialic acid content of recombinant glycoproteins. Nature


Wileman T, Harding C, and Stahl P (1985). Receptor-mediated endocytosis. Biochemical Journal 232:1-

14.

Wilson NL, Schulz BL, Karlsson NG, and Packer NH (2002). Sequential analysis of N- and O-linked

glycosylation of 2D-PAGE separated glycoproteins. Journal of Proteome Research 1 (6):521-529.

Wilson SP, and Smith LA (1997). Addition of glycerol during DNA exposure enhances calcium phosphate

transfection. Analytical Biochemistry 246 (1):148-150.

Wong CFD, Wong TKK, Goh TL, Heng KC, and Yap GSM (2005). Impact of dynamic online fed-batch

strategies on metabolism, productivity and N-glycosylation quality in CHO cell cultures. Biotechnology


Wong C-H (2005). Protein glycosylation: new challenges and opportunities. The Journal of Organic

Chemistry 70 (11):4219-4225.

Wong DCF, Wong NSC, Goh JSY, May LM, and Yap MGS (2010). Profiling of N-glycosylation gene

expression in CHO cell fed-batch cultures. Biotechnology and Bioengineering 107 (3):516-528.

Woolley JF, and Al-Rubeai M (2009). The application of SELDI-TOF mass spectrometry to mammalian cell

culture. Biotechnology Advances 27 (2):177-184.

Wopereis S, Lefeber DJ, Morava É, and Wevers RA (2006). Mechanisms in protein O-glycan biosynthesis

and clinical and molecular aspects of protein O-glycan biosynthesis defects: a review. Clinical

Chemistry 52 (4):574-600.


95

Wright A, Sato Y, Okada T, Chang KH, Endo T, and Morrison SL (2000). In vivo trafficking and catabolism

of IgG1 antibodies with Fc associated carbohydrates of differing structure. Glycobiology 10

(12):1347-1355.

Wuhrer M, Catalina MI, Deelder AM, and Hokke CH (2007). Glycoproteomics based on tandem mass

spectrometry of glycopeptides. Journal of Chromatography B 849 (1–2):115-128.

Wuhrer M, Koeleman CAM, Hokke CH, and Deelder AM (2004). Protein glycosylation analyzed by normal-

phase nano-liquid chromatography−mass spectrometry of glycopeptides. Analytical Chemistry 77

(3):886-894.

Wurm FM, Johnson A, Lie YS, Etcheverry MT, and Anderson KP (1992). Host cell derived retroviral

sequences enhance transfection and expression efficiency in CHO cells. In Animal Cell technology:

Developments, Processes and Products, edited by RE Spier, JB Griffiths, G MacDonnald, Oxford, Uk:

Butterworth--Heinemann.

Wurm FM (2004). Production of recombinant protein therapeutics in cultivated mammalian cells. Nature


Yahata K, Kishine H, Sone T, Sasaki Y, Hotta J, Chesnut JD, Okabe M, and Imamoto F (2005). Multi-gene

gateway clone design for expression of multiple heterologous genes in living cells: conditional gene

expression at near physiological levels. Journal of Biotechnology 118 (2):123-134.

Yamane-Ohnuki N, Kinoshita S, Miho I-U, Kusunoki M, Iida S, Nakano R, Wakitani M, Niwa R, Sakurada M,

Uchida K, Shitara K, and Satoh M (2004). Establishment of FUT8 knockout Chinese hamster ovary

cells: an ideal host cell line for producing completely defucosylated antibodies with enhanced

antibody-dependent cellular cytotoxicity. Biotechnology and Bioengineering 87 (5):614-622.

Yamane-Ohnuki N, and Satoh M (2009). Production of therapeutic antibodies with controlled fucosylation.

MAbs 1 (3):230-236.

Yang J-D, Angelillo Y, Chaudhry M, Goldenberg C, and Goldenberg DM (2000). Achievement of high cell

density and high antibody productivity by a controlled-fed perfusion bioreactor process.


Yang M, and Butler M (2000). Effects of ammonia on CHO cell growth, erythropoietin production, and

glycosylation. Biotechnology and Bioengineering 68 (4):370-380.

Yang M, and Butler M (2002). Effects of ammonia and glucosamine on the heterogeneity of erythropoietin

glycoforms. Biotechnology Progress 18 (1):129-138.

Yoon SK, Choi SL, Song JY, and Lee GM (2005). Effect of culture pH on erythropoietin production by

Chinese hamster ovary cells grown in suspension at 32.5 and 37.0°C. Biotechnology and



96

Yoon SK, Hong JK, Choo SH, Song JY, Park HW, and Lee GM (2006). Adaptation of Chinese hamster

ovary cells to low culture temperature: cell growth and recombinant protein production. Journal of


Yoon SK, Kim SH, and Lee GM (2003). Effect of low culture temperature on specific productivity and

transcription level of anti-4-1BB antibody in recombinant Chinese hamster ovary cells. Biotechnology

Progress 19 (4):1383-1386.

Yoon SK, Song JY, and Lee GM (2003). Effect of low culture temperature on specific productivity,

transcription level, and heterogeneity of erythropoietin in Chinese hamster ovary cells. Biotechnology


Yuansheng Y, Mariati, Ho SC, and Yap MG (2009). Mutated polyadenylation signals for controlling

expression levels of multiple genes in mammalian cells. Biotechnology and bioengineering 102

(4):1152-1160.

Yuen CT, Storring PL, Tiplady RJ, Izquierdo M, Wait R, Gee CK, Gerson P, Lloyd P, and Cremata JA

(2003). Relationships between the N-glycan structures and biological activities of recombinant

human erythropoietins produced using different culture conditions and purification procedures.

British Journal of Haematology 121 (3):511-526.

Zahn-Zabal M, Kobr M, Girod PA, Imhof M, Chatellard P, de Jesus M, Wurm F, and Mermod N (2001).

Development of stable cell lines for production or regulated expression using matrix attachment

regions. Journal of Biotechnology 87 (1):29-42.

Zanghi JA, Schmelzer AE, Mendoza TP, Knop RH, and Miller WM (1999). Bicarbonate concentration and

osmolality are key determinants in the inhibition of CHO cell polysialylation under elevated pCO2 or

pH. Biotechnology and Bioengineering 65 (2):182-191.

Zhang, J (2010). Mammalian cell culture for biopharmaceutical production. In Manual of Industrial

Microbiology and Biotechnology. 3rd edition, edited by RH Baltz.

Zhang J, and Robinson D (2005). Development of animal-free, protein-free and chemically-defined media

for NS0 cell culture. Cytotechnology 48 (1):59-74.

Zhao F-Q, and Keating AF (2007). Functional properties and genomics of glucose transporters. Current

Genomics 8 (2):113-128.

Zhou Q, Shankara S, Roy A, Qiu H, Estes S, McVie-Wylie A, Culm-Merdek K, Park A, Pan C, and Edmunds

T (2008). Development of a simple and rapid method for producing non-fucosylated oligomannose

containing antibodies with increased effector function. Biotechnology and Bioengineering 99 (3):652-

665.

Zhu J (2011). Mammalian cell protein expression for biopharmaceutical production. Biotechnology

Advances 30 (5):1158-1170.


97

Zhu MM, Mollet M, and Hubert RS (2007). Industrial production of therapeutic proteins: cell lines, cell

culture, and purification. In Kent and Riegel’s Handbook of Industrial Chemistry and Biotechnology,

edited by JA Kent: Springer US.

Zhu Y, Li X, Mcvie-Wylie A, Jiang C, Thurberg BL, Raben N, Mattaliano RJ, and Cheng SH (2005).

Carbohydrate-remodelled acid α-glucosidase with higher affinity for the cation-independent mannose

6-phosphate receptor demonstrates improved delivery to muscles of Pompe mice. The Biochemical

Journal 389 (Pt 3):619-628.

Ziao C, Huang Z, Li W, Hu X, Qu W, Gao L, and Liu G (2002). High density and scale-up cultivation of

recombinant cho cell line and hybridomas with porous microcarrier cytopore. In New Developments

and New Applications in Animal Cell Technology, edited by O-W Merten, P Perrin and B Griffiths:

Lavoisier.

Zou G, Ochiai H, Huang W, Yang Q, Li C, and Wang L-X (2011). Chemoenzymatic synthesis and Fcγ

receptor binding of homogeneous glycoforms of antibody Fc domain. presence of a bisecting sugar

moiety enhances the affinity of Fc to FcγIIIa receptor. Journal of the American Chemical Society 133

(46):18975-18991.


The OSCARTM system for the transfection

of Chinese hamster ovary cells for

monoclonal antibody production

CC

THE PRESENT CHAPTER WAS ADAPTED FROM THE FOLLOWING RESEARCH PAPER

ANA RITA COSTA, Maria Elisa Rodrigues, Mariana Henriques, David Melton, Philip

Cunnah, Rosário Oliveira, Joana Azeredo (2012). Evaluation of the OSCARTM system

for the production of monoclonal antibodies by CHO-K1 cells. International Journal of

Pharmaceutics, 430:42-46.

101

iopharmaceutical production of complex recombinant protein therapeutics currently

relies on mammalian cell systems. The development of high-yielding stable cell lines requires

processes of transfection, selection and adaptation. With several technologies available, selection

has been most frequently based on dihydrofolate reductase or glutamine synthetase systems,

which can be very time-consuming. Due to the pressure to reduce development costs and speed

up time to market, new technologies are emerging, as the promising OSCARTM expression system

that could provide more rapid development of high-yielding stable cell lines than the traditional

systems. However, further evaluation of its application in a wider range of cell types and media is

still necessary. In this chapter, the application of OSCARTM for the transfection of a Chinese

hamster ovary (CHO-K1) cell line with a monoclonal antibody (mAb) was evaluated. OSCARTM was

reasonably fast and simple, without negative impact on cell growth characteristics. However,

minigene selection proved to be critical, with only the most disabled minigene (pDWM128)

working for the cell line assessed. Furthermore, the initial relatively high levels of production

decreased significantly in the first few weeks of culture, remaining relatively stable although with

low yield thereafter. This study also suggests a strong impact of the methodology of clone

selection on the outcome of the process, recommending the use of an accurate method such as

productivity assessment. Additionally, glycosylation analysis of the mAb produced was performed

by high performance liquid chromatography and indicated a similar product quality among the

clones obtained after transfection. As a whole, the results of this work prove that the OSCARTM

expression system has significant value to the biopharmaceutical industry.

Keywords: Chinese hamster ovary cells; Transfection; OSCARTM expression system; Monoclonal

antibody; Glycosylation

BB

CHAPTER 2 │ TRANSFECTION WITH THE OSCARTM EXPRESSION SYSTEM

103

2.1. INTRODUCTION

Therapeutic proteins, and monoclonal antibodies (mAbs) in particular, have assumed an

increasing importance in the last years, which is corroborated by their significant market growth

(Li and d’Anjou 2009; Durocher and Butler 2009). They are applied in a wide range of areas that

include treatment of diseases such as cancer, diabetes, anaemia, haemophilia, allergies, and

blood clotting disorders, as well as research applications (Bhopale and Nanda 2005). For their

commercial production, several expression systems can be used, with mammalian cells as the

current preference (Mohan et al. 2008; Melton et al. 2001; Chu and Robinson 2001; Andersen

and Reilly 2004) due to their ability to perform in-vivo correct refolding and post-translational

modifications (Werner et al. 1998), which guarantee a proper biological function of the protein.

In order to maximize the protein yield in mammalian cells, gene expression systems are

routinely used, which through selection and amplification can provide relatively high expressing

stable cell lines to use in manufacture (Page and Sydenham 1991; Wurm 2004; Bebbington et

al. 1992). The most commonly used gene expression systems include the dihydrofolate

reductase (DHFR) and the glutamine synthetase (GS) systems that exploit different metabolic

pathways (nucleotide metabolism for DHFR and glutamine metabolism for GS) (Wurm 2004;

Lonza), but similarly require multiple rounds of amplification after transfection through the use of

increasing concentrations of specific drugs (methotrexate in DHFR and methionine sulphoximine

in GS) (Chusainow et al. 2009; Lonza ; Andersen and Reilly 2004). This results in extended

development times and has high costs associated with the use of specialized media and toxic

chemicals.

Recently, a novel expression system known as OSCARTM has been developed at the

University of Edinburgh (Melton et al. 2001). This system relies on partially disabled minigene

vectors that encode for hypoxanthine phosphoribosyltransferase (HPRT), essential for purine

synthesis via the normal cellular salvage pathway (Barnes, Bentley, and Dickson 2001; Melton et

al. 2001). HPRT-deficient mammalian cells transfected with one of these minigenes and a gene

of interest are placed in a selective Hypoxanthine Aminopterin Thymidine (HAT) medium that

blocks the de novo purine synthesis. This makes cell survival reliant on the salvage pathway

using a disabled HPRT enzyme (Melton et al. 2001). Since large amounts of this enzyme are

required for cell survival, selection and amplification in the OSCARTM system takes place in a


104

single step which potentially provides a time advantage over the traditional DHFR and GS

systems. In addition to being more expedite, OSCARTM is also said to provide cells with higher (7-

fold) and stable (Melton et al. 2001) expression yields at lower (7-fold) costs of goods

(Nicholson), due to the absence of specialized media and toxic chemicals that are required by

traditional systems. Furthermore, by providing HPRT minigenes with varying degrees of

expression disability, OSCARTM can be used in a wide range of cell types with different

requirements for HPRT expression, and enables the optimization of the production of different

proteins.

Although with many potential advantages, the recentness of OSCARTM demands for a deeper

evaluation of its applicability in a wider range of cell types and media, as well as its scalability

and performance in bioreactor cultures (Costa et al. 2010). Thus, the work described in the

present chapter evaluates the application of OSCARTM for the transfection of a CHO-K1 cell line

expressing a mAb. This covered the impact of the expression system on cell growth

characteristics, the levels and stability of mAb production achieved, as well as the ease of the

implementation/execution of the methodology. Additionally, the glycosylation profile of the mAb

obtained from the highest-producer cell clones was compared to evaluate possible differences in

product quality among the clones obtained using the OSCARTM expression system.


105

2.2. MATERIALS AND METHODS

2.2.1. CELLS, PLASMIDS AND MINIGENES

The CHO-K1 cell line was obtained from the American Type Culture Collection (ATCC, CCL-

61, Spain). The HPRT-deficient cells CHO-K1 TG-2 were kindly provided by the University of

Edinburgh (Scotland), which also supplied the disabled HPRT minigenes pDWM131, pDWM129

and pDWM128 required for transfection with their proprietary OSCARTM system. The

characteristics of these minigenes in comparison with the fully functional HPRT gene (pBT/PGK-

HPRT[RI]) are shown in Table 2.1. The minigenes were linearized with BamHI by the company

Biotecnol SA (Lisbon, Portugal), which also provided the gene/plasmid of their monoclonal

antibody CAB051 for cell transfection.

TABLE 2.1. Characteristics of the disabled HPRT minigenes used in the OSCARTM expression system, in comparison

with the fully functional HPRT gene pBT/PGK-HPRT[RI]

HPRT GENE/MINIGENE LEVEL OF EXPRESSION DISABILITY

pBT/PGK-HPRT[RI] Fully functional gene, contains nine introns

pDWM131 Least disabled minigene, contains truncated introns 1 and 2, and introns 7 and 8

pDWM129 Minigene contains truncated intron 1 and a more truncated intron 2

pDWM128 Most disabled minigene, contains only the truncated intron 1

2.2.2. SELECTION AND VALIDATION OF HPRT-DEFICIENT CLONES

CHO-K1 cells were seeded into petri dishes (Frilabo, Portugal) at 5x104 cells/mL in Growth

medium composed of Dulbecco’s Modified Eagle’s Medium (DMEM, Sigma-Aldrich, United

Kingdom) supplemented with 10 % fetal bovine serum (FBS, Sigma-Aldrich) and 200 mM L-

glutamine (Sigma-Aldrich), and grown overnight at 37 ºC and 5 % carbon dioxide (CO2). The

medium was replaced by Growth medium supplemented with 20 µg/mL 6-thioguanine (Sigma-

Aldrich) and the cells incubated for additional 14-15 days. Then, individual clones of HPRT-


106

deficient cells were selected and isolated using either cloning rings or cloning disks (Sigma-

Aldrich), according to manufacturer’s instructions. Briefly, the bottom part of the cloning rings

was coated with silicone grease (Sigma-Aldrich) and placed over individual cell colonies. Cells

isolated within the ring were detached using trypsin (Sigma-Aldrich), transferred to 24-well plates

(Frilabo) and sequentially expanded to 6-well plates, 25 cm2 and 75 cm2 culture flasks (Frilabo).

Cloning disks, for their turn, were soaked in trypsin and placed directly over individual cell

colonies. After 5-10 minutes, the disks containing the detached cells were transferred to 24-well

plates and expanded as described above for the cloning rings.

The clones isolated were validated for HPRT-deficiency and mutation stability, by seeding

into petri dishes at 1.3x105 cells/mL in Growth medium supplemented with 50x HAT (final

working concentration of 100 µM hypoxanthine, 0.4 µM aminopterin, and 16 µM thymidine)

(Sigma-Aldrich) (hereafter named HAT medium), at 37 ºC and 5 % CO2. A control was performed

with the parental CHO-K1 cells using the same conditions. Clones were validated by their inability

to grow in the HAT medium in contrast with the normal growth of the control cells.

2.2.3. TRANSFECTION

The validated HPRT-deficient clones and the HPRT-deficient CHO-K1 TG-2 cells were seeded

into petri dishes at 1.3x105 cells/mL in Growth medium, and incubated overnight at 37 ºC and 5

% CO2 for transfection by calcium phosphate co-precipitation.

The CAB051 gene/plasmid was co-transfected with each linearized HPRT minigene. For

this, 1 mL of 1x hydroxyethyl piperazineethanesulfonic acid (HEPES)-buffered salt solution (HBS,

137 mM sodium chloride (NaCl), 4.96 mM potassium chloride (KCl), 417 mM sodium phosphate

dibasic dihydrate (Na2HPO4.2H2O), 5.55 mM D-glucose, and 17.6 mM HEPES, pH 7.05) (Sigma-

Aldrich) was added to a deoxyribonucleic acid (DNA) mixture of 20 µg HPRT minigene and 20 µg

CAB051 plasmid (less than 100 µL of DNA). Then, 62 µL of 2 M calcium chloride (CaCl2, Sigma-

Aldrich) was added dropwise to the mixture and left for 45 min at room temperature until a

precipitate was formed. The precipitate was slowly dropped over the cells in the petri dishes after

removal of the culture medium. After incubation at room temperature for 20 min with frequent

spreading of the precipitate by tilting, Growth medium was added and the cells were incubated

overnight at 37 ºC and 5 % CO2.


107

Cells were then seeded into 24-well plates (at least two for each transfection) at a density of

10,000, 20,000, and 40,000 cells/well for pDWM131, and at 40,000 cells/well for pDWM129

and -128, in Growth medium, to provide an average of one HAT-resistant colony per well. After

24 h the medium was replaced by HAT medium to allow for the selective growth of the cells

successfully transfected, and the cells were incubated for 3-4 weeks with medium changes after

the first 2-3 days and every 10-14 days thereafter.

2.2.4. SELECTION OF THE HIGHEST MONOCLONAL ANTIBODY PRODUCERS

The individual well colonies obtained after transfection were assessed for expansion of the

highest mAb-producers using two methodologies: (i) measurement of absorbance, and (ii)

determination of productivity. For the first methodology, samples of the supernatant were taken

from each well containing a colony, and the absorbance read at 450 nm (Biotek Synergy HT

multi-mode microplate reader, Izasa, Portugal). In this case, a direct relationship between

absorbance values and mAb concentration is assumed. For the methodology based on

productivity, supernatant samples were taken from each well containing a colony to determine

mAb concentration by enzyme-linked immunosorbent assay (ELISA). Cell concentration was also

assessed by enzymatic release of the cells with trypsin and cell counting in a haemocytometer

using the trypan blue (Sigma-Aldrich) exclusion method, which allowed the subsequent

determination of mAb productivity, as described in section 2.2.5.

Clones selected were expanded into 25 cm2 culture flasks under selective conditions (HAT

medium). The stability of production of these clones was assessed periodically for several weeks,

as well as after two years (with the culture subjected to periods of both continuous splitting and

cryopreservation), following the methodology of productivity assessment described above.

2.2.5. ANTIBODY QUANTIFICATION BY ENZYME-LINKED IMMUNOSORBENT ASSAY

Samples were analyzed for mAb productivity by ELISA following an optimized procedure

described in the confidential and proprietary standard operating procedure (SOP) 2008-01 ANL

from Biotecnol SA. Briefly, 96 well plates (CoStar, United States of America) were coated with a


108

capture antibody (Sigma-Aldrich) overnight. After a 45 min blocking at room temperature with 1 %

bovine serum albumin in phosphate buffered saline (BSA-PBS, 11.5 µM BSA in 137mM NaCl,

2.68 mM KCl, 1.76 mM potassium dihydrogen phosphate (KH2PO4), and 12.72 mM sodium

phosphate dibasic (Na2HPO4), pH 7.4) (Sigma-Aldrich), sample dilutions in 0.2 % BSA-PBS (3.03

mM BSA in PBS), a standard of known concentration and a quality control (Biotecnol SA), were

added to the plates and incubated for 2 h at 37 ºC. The detection antibody was added and the

plates incubated for additional 2 h at room temperature. A 3,3’,5,5’-Tetramethylbenzidine (TMB,

Sigma-Aldrich) substrate solution was added and allowed to react for 10 min at room

temperature. After stopping the reaction with Phosphoric acid 75 % (Phosphoric acid 85 % in

water) (Frilabo), the absorbance was read at 450 nm and mAb production determined. For this,

the calibration curve of the ELISA assay was determined using the R software (version 2.6.2, The

R Foundation for Statistical Computing) to obtain the values of the four parameter logistic of

Equation 2.1.

Abs��=d+a-d

1+ �C�c �b

EQUATION 2.1. Four parameter logistic equation used as the calibration curve for the determination of monoclonal

antibody concentration from the absorbance values obtained by enzyme-linked immunosorbent assay. Abs450 -

absorbance read at 450 nm; a - estimated response at zero concentration; b - slope factor; c - mid-range

concentration; d - estimated response at infinite concentration; and CmAb - mAb concentration in µg/mL.

MAb concentration (production, CmAb) of the samples was then determined using the

functional inverse of Equation 2.1, as represented in Equation 2.2.

C� = �� − 1�

�� × c

EQUATION 2.2. Determination of monoclonal antibody concentration (CmAb) in µg/mL. a - estimated response at zero

concentration; b - slope factor; c - mid-range concentration; d - estimated response at infinite concentration; and

Abs450 - absorbance read at 450 nm.

Additionally, the specific mAb production (productivity, qmAb) of each sample was calculated

according to Equation 2.3, using the cell concentration (Ccell) determined by Equation 2.4.


109

�� = �� Ccell⁄t

× 10#

EQUATION 2.3. Determination of the monoclonal antibody productivity (qmAb) in pg/cell/day. CmAb - mAb concentration in

µg/mL; Ccell - cell concentration in cells/mL; and t - time of production in days.

C%&'' = ()*++� × 10� × F

EQUATION 2.4. Determination of cell concentration (Ccell) in cells/mL. Ncell - total number of cells counted in the

haemocytometer; and F - dilution factor of the sample.

Data of clone productivity and stability of production were statistically analyzed using the

Statistical Package for the Social Sciences (SPSS) software (International Business Machines,

IBM, USA), using one-way analysis of variance (ANOVA) with Bonferroni test, with a confidence

level of 95 %.

2.2.6. CELL GROWTH CHARACTERISTICS

The growth characteristics, specifically the cell doubling time (tD), of the original CHO-K1

cells and of two transfected clones (selected clones 18 and 32) were evaluated and compared.

For this, cells were seeded at 1x105 cells/mL into 6-well plates and allowed to grow at 37 ºC and

5 % CO2, for 2, 4, 6, 10, 24 and 48 h before direct cell counting with an haemocytometer and

trypan blue staining. The tD values were determined according to Equation 2.5.

t. = /t0 − t12 × '34/02'34/�5 ��⁄ 2

EQUATION 2.5. Determination of the cell doubling time (tD) in hours, assuming a constant growth rate (exponential

phase of cell growth). C1 and C2 - cell concentrations, in cells/mL, at times t1 and t2, in hours, respectively.

2.2.7. ANALYSIS OF THE GLYCOSYLATION PROFILE

The glycosylation profile of the mAb produced by Clone 18 and Clone 32 was assessed as

described below. To note that all water used in the following procedures was type 1 ultra-pure,


110

resistivity of 18 MΩ.cm, particle-free (> 0.22 µm), and total organic content of less than ten parts

per billion.

IGG PURIFICATION

Prior to IgG purification, samples were concentrated to volumes of 200 µL using 15 mL spin

concentrators of 5 KDa molecular weight cutoff (Agilent Technologies, Ireland). Then, the IgG

from the concentrated samples was purified with Protein A spin plates (Thermo Scientific,

Ireland) according to manufacturer instructions, with slight modifications. Briefly, the wells of a

spin plate were equilibrated with two washes of 200 µL of 1x PBS (pH 7.2, Merck, UK). Samples

were diluted two-fold in 1x PBS and added (200 µL) to the plate for a period of incubation of 30

min with moderate agitation. The resin was washed three times with 500 µL of PBS, and the

purified IgG eluted three times with 300 µL of 0.5 M acetic acid (pH 2.5, AnalaR, UK) and

neutralized with 20 µL of 1 M ammonium bicarbonate (pH 7-8.5, Sigma-Aldrich). The neutralized

elutions were dried overnight in a vacuum centrifuge.

IN GEL BLOCK IMMOBILIZATION

The dried samples were reduced with a solution containing 13.88 mM sodium dodecyl

sulfate (SDS, BDH, UK), 12.5 mM Tris (pH 6.6, AnalaR) and 0.05 M dithiothreitol (DTT, Sigma-

Aldrich) for 15 min at 65 ºC; and alkylated with 100 mM iodoacetamide (IAA, Sigma-Aldrich) for

30 min in the dark. The gel blocks were formed by adding a mixture of 22.5 µL of Protogel (30 %,

National Diagnostics, UK), 11.25 µL of 1.5 mM Tris, 1 µL of 35 mM SDS, and 1 µL of 34.7 mM

ammonium peroxisulphate (APS, AnalaR), and finally adding 1 µL of tetramethylethylenediamine

(TEMED, Sigma-Aldrich) to the samples, letting set for 15 min.

N-GLYCAN RELEASE AND FLUORESCENT LABELING

Each gel block was cut into small pieces (≈ 2 mm2), washed alternately with acetonitrile

(Sigma-Aldrich) and 20 mM sodium bicarbonate (pH 7, Merck), vortexed and mixed for 10 min

after each wash, and dried in a vacuum centrifuge.


111

For the release of N-linked glycans, the dried gel pieces were incubated for 15 min in a mix

of 2 µL of Peptide N-glycosidase F (PNGase F, Prozyme, USA) and 48 µL of 20 mM sodium

bicarbonate (pH 7), and a further 12-16 h at 37 ºC after adding extra 50 µL 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 sonication for each

wash. The supernatants obtained were collected and dried in a vacuum centrifuge. The dried

samples were redissolved in 20 µL of formic acid (1 %, Sigma-Aldrich), incubated for 40 min at

room temperature and dried in a vacuum centrifuge.

The released N-glycans were labeled for fluorescent detection using the LudgerTagTM 2-

aminobenzamide (2-AB) Glycan Labeling Kit (Ludger, UK), according to manufacturer

instructions. Briefly, 5 µL of the 2-AB labeling mix were added to each dried glycan sample,

incubated 30 min at 65 ºC, vortexed, spun down, and incubated for further 90 min.

Excess 2-AB label from labeled N-glycans was cleaned-up using Normal Phase 1 PhyNexus

tips (PhyNexus, USA). Briefly, samples were diluted in 95 µL of water and 900 µL of acetonitrile.

The PhyNexus tips were prepared by washing with ten 500 µL uptakes of 95 % acetonitrile (v/v in

water), ten 500 µL uptakes of 20 % acetonitrile (v/v in water), and another ten 500 µL uptakes of

95 % acetonitrile. Samples were loaded into the PhyNexus tips through ten 1 mL in-out cycles,

followed by washing of the tips with ten 1 mL uptakes of 95 % acetonitrile. The glycans were then

eluted with five uptakes of 200 µL of 20 % acetonitrile, and the elutions collected and dried in a

vacuum centrifuge.

NORMAL PHASE-HIGH PERFORMANCE LIQUID CHROMATOGRAPHY

For normal phase-high performance liquid chromatography (NP-HPLC), each dried sample

was ressuspended in 20 µL of water and 80 µL of acetonitrile. NP-HPLC was performed using a

TSKgel Amide-80 3 µm (150 x 4.6 mm) column (Tosoh Bioscience, UK) for 60 min runs, at 30

ºC, using 50 mM ammonium formate (Sigma-Aldrich) as Solvent A and acetonitrile as Solvent B.

The runs were performed on a 2695 Alliance separations module (Waters, Ireland) with a

2475 multi-wavelength fluorescence detector (Waters), with excitation and emission 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


112

% 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) alongside the sample runs.

PROCESSING OF SAMPLES

A fifth-order polynomial distribution curve was fitted to the dextran 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 available in GlycoBase (http://glycobase.nibrt.ie/), and using the known

immunoglobulin G1 (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.


113

2.3. RESULTS AND DISCUSSION

2.3.1. ANALYSIS OF THE TWO MAIN PHASES OF THE OSCARTM TECHNOLOGY

In the first phase of the OSCARTM technology, consisting of the selection of HPRT-deficient

clones, two methodologies for clone isolation were assessed, specifically cloning rings and

cloning disks. Their comparison in terms of simplicity, speed, and efficiency of cell recovery is

shown in Table 2.2.

TABLE 2.2. Comparison of cloning rings and cloning disks for the isolation of HPRT-deficient clones obtained during

the first phase of the OSCARTM expression system, considering simplicity, speed and cell recovery

MONOSACCHARIDE CLONING RINGS CLONING DISKS

Simplicity - +

Speed - +

Cell recovery + -

In terms of procedure, cloning disks had the simplest and fastest procedure. However,

considering efficiency, cloning rings were clearly superior by allowing the recovery of a higher

number of cells. Therefore, cloning rings proved to be the best choice when maximum cell

recovery is desired and/or crucial, such as in the present work where the amount of cells

isolated was very low.

For the second phase of the OSCARTM expression system – the transfection itself – three

HPRT minigenes with different degrees of expression disability are provided, to determine the

best choice for each cell line and/or product. These minigenes were evaluated for the co-

transfection of the CHO-K1 cells with the CAB051 plasmid, with the results obtained shown in

Table 2.3.


114

TABLE 2.3. Comparison of the three HPRT minigenes with varying degrees of expression disability for the co-

transfection of the CAB051 plasmid into the CHO-K1 cells using the OSCARTM expression system

HPRT GENE/MINIGENE LEVEL OF EXPRESSION DISABILITY TRANSFECTION SUCCESS

pDWM131 Least disabled +

pDWM129 -

pDWM128 Most disabled -

The successful transfection of the CHO-K1 cells for mAb production was only possible using

the most disabled minigene - pDWM128, with the others not providing any cell colonies. This was

an unexpected result, since the most disabled minigene is the one that requires extra

amplification, which may not allow the survival of cells with more fastidious requirements for

HPRT expression; so the most obvious outcome would be the inverse of what was observed in

this study. These results highlight the importance of testing all minigenes, since its success

seems to be dependent on the type of cells used.

Indeed, the possibility of selection from different minigenes has the potential advantage of

making the OSCARTM expression system adaptable to the requirements of different cell lines, and

allows tailoring of the system to the desired levels of expression (the most disabled minigene

usually provides the highest yields). However, testing different minigenes requires additional lab

work and increases costs. These problems may be attenuated over time with further studies that

will result in the build-up of information that may allow the knowledge beforehand of the

applicability of each minigene to different cell lines and clones, products and/or plasmids. In this

sense, the construction of a database to collect all the information would be of great value.

2.3.2. LEVEL AND STABILITY OF CLONE PRODUCTIVITY

After transfection, the clones obtained were assessed for the selection of the most

productive. This initial selection is typically based on the absorbance values of the clone

supernatants, a fast process but not the most accurate approach. A more precise, although time-

consuming and labor-intensive, alternative consists on determining productivity by measuring

both cell and antibody concentrations. Both methodologies were used and compared in the

present work, as shown in Figure 2.1.


115

FIG

UR

E 2

.1. Initial clone selection after transfection based on two approaches: (a) absorvance values of supernatant read at 450 nm (ABS 4

50), and (b) productivity (q m

Ab). The

interrupted line denotes the cutoff limit established for clone selection: (a) ABS 450 of 1, and (b) q

mAb of 10 pg/cell/day; and the dark color highlights the selected clones.

0.0

0.5

1.0

1.5

2.0

2.5

3.0

0

2

4

6

8

10

12

14

16

18

20

22

24

26

28

30

32

34

36

38

40

42

44

46

48

50

52

54

56

58

60

62

64

66

68

70

72

74

76

78

80

82

84

86

88

90

92

94

96

98

100

102

104

106

108

110

112

ABS450

Clo

ne

0

10

20

30

40

50

60

0

2

4

6

8

10

12

14

16

18

20

22

24

26

28

30

32

34

36

38

40

42

44

46

48

50

52

54

56

58

60

62

64

66

68

70

72

74

76

78

80

82

84

86

88

90

92

94

96

98

100

102

104

106

108

110

112

qmAb(pg/cell/day)

Clo

ne

(b)

(a)


116

Differences between the two methodologies are clearly observed, primarily in the higher

number of clones that seem to be high-producers when the selection is based in absorbance.

However, more noteworthy are the divergences in which clones are selected with each method.

Indeed, setting the selection criteria on absorbance above 1 (clones 27, 30, 34, 49, 65, 73, 75,

and 93) and productivity above 10 pg/cell/day (clones 6, 18, 27, 32, and 73), there are only two

clones in common for both approaches (clones 27 and 73). Additionally, in cases such as clones

6 and 93, the results obtained with each methodology are opposite. These results reveal a strong

impact of the methodology on the outcome of the process of selection. Therefore, although the

determination of productivity for each clone is less straight-forward and more time-consuming

than the assessment of absorbance, the fact that it contemplates both cell and mAb

concentrations makes it more accurate and therefore the favored method to base clone selection.

Considering these findings, the clones with productivities above 10 pg/cell/day (clones 6,

18, 27, 32, and 73) were selected to further assess the stability of production. Additionally, a

clone with lower productivities (clone 95) was also selected to enable the comparison between

high- and low-producing clones. The stability of production was evaluated by periodic monitoring

of clone productivity during 6 weeks, with additional measurements after two years. The first

measurement was performed after 3 weeks in culture (splitting three times a week), and the

productivity levels obtained are shown in Figure 2.2.

FIGURE 2.2. Values of productivity (qmAb) of the selected clones obtained initially and after three weeks in culture.

0

10

20

30

40

50

6 18 27 32 73 95

qm

Ab

(pg

/ce

ll/d

ay)

Clone

Initial

3 weeks


117

The evolution of productivity with time was evaluated for additional 3 weeks, as shown in

Figure 2.3. It should be noted that since clones 6 and 73 have shown very low levels of

production already at 3 weeks, they were not considered in this evaluation.

FIGURE 2.3. Evolution of the productivity (qmAb) levels of clones 18, 27, 32, and 95 during 6 weeks in culture.

In the first evaluation of the stability of production (Figure 2.2), a substantial decrease of

production, between 70 and 99 %, was noticeable for all clones. Further assessment (Figure 2.3)

shows that, although not so abruptly as before, the levels of production continue to decrease over

time and, after 6 weeks in culture, become significantly (p < 0.05) inferior to those of the first

measurements.

Comparing the clones, an attenuation of the differences in the productivity levels is observed

over time (Figure 2.3). Indeed, although clone 18 was the highest-producer (p < 0.05) in the first

weeks, at the end of the period of evaluation all clones showed similar levels of productivity (p >

0.05).

To further assess production stability, clones 18 and 32 were cultured for two years, after

which additional measurements of productivity were performed. The selection of these clones

was based on the globally better (although not significantly) levels of productivity demonstrated in

the first six weeks in culture. The results obtained after two years are shown in Table 2.4.

0.00

0.25

0.50

0.75

1.00

1.25

1.50

1.75

2.00

2.25

18 27 32 95

qm

Ab

(pg

/ce

ll/d

ay)

Clone

3 weeks

4 weeks

5 weeks

6 weeks


118

TABLE 2.4. Values of productivity (qmAb) obtained for Clone 18 and Clone 32 after two years in culture

CELLS qmAb (pg/cell/day)

Clone 18 0.21

Clone 32 0.31

The levels obtained were similar (p > 0.05) to those of the previous period of analysis (6

weeks) indicating a tendency to stabilize production. Nevertheless, there is a considerable loss of

the ability of the clones to produce mAb after two years in culture in comparison with the initial

measurements. Therefore, in opposition to what has been observed in studies with other cell

lines/proteins (Melton et al. 2001), the OSCARTM system does not provide stability of high yield

production when applied to the CHO-K1 cells producing the CAB051 mAb.

It should be noted that the clones with better levels of production and stability (clones 18 and

32) would not have been selected if the absorbance method of clone selection had been used,

highlighting the importance of the methodology of clone selection for the outcome of the process.

2.3.3. EFFECT OF TRANSFECTION ON CELL GROWTH CHARACTERISTICS

The growth characteristics of the CHO-K1 cells before and after transfection with OSCARTM

were evaluated. For this, the doubling times of the original CHO-K1 cells and of the mAb-

producing clones 18 and 32 were determined after two years in culture, as shown in Table 2.5.

TABLE 2.5. Doubling times (tD) of the original CHO-K1 cells and of Clones 18 and 32, after two years in culture

CELLS tD (h)

CHO-K1 20.97 ± 3.76

Clone 18 21.86 ± 4.30

Clone 32 22.69 ± 0.98

The OSCARTM expression system is said not to compromise the growth rate of cells, even if

generating a high gene copy number (Melton et al. 2001). This is a common problem after

transfection, since the ability to produce foreign proteins can be a metabolic burden to the cells and

therefore affect their growth (Gu et al. 1992; Gu, Todd, and Kompala 1995). However, in this work,

the presence of the gene does not appear to significantly affect the doubling times of the


119

transfected clones (Table 2.5). This may be an advantage of the OSCARTM system over more

traditional methodologies, but may also be caused by the clones not being expressing at particularly

high levels and therefore being capable to maintain growth rates similar to those of the parent cells.

2.3.4. ANALYSIS OF THE MONOCLONAL ANTIBODY GLYCOSYLATION PROFILE

The different mAb-producing clones obtained after transfection with the OSCARTM expression

system have shown different capacities of production. Likewise, they may also produce mAbs

with divergent patterns of glycosylation, which has potential effects on product quality. To

evaluate this possibility, the glycan profile of the mAb produced by clones 18 and 32 was

assessed. The peaks obtained and the corresponding GU values, tentative structure assignment,

and relative area are shown in Table 2.6.

TABLE 2.6. Data of mean glucose unit (GU) value, tentative structure assignment, and relative area (%) of the peaks

obtained by high performance liquid chromatography for the monoclonal antibody produced by Clones 18 and 32

PEAK MEAN GU 1 TENTATIVE ASSIGNMENT 2 % AREA

CLONE 18 CLONE 32

1 4.91 ± 0.02 A1 1.86 2.09

2 5.34 ± 0.01 FA1 / A2 2.05 2.01

3 5.82 ± 0.02 FA2 40.59 40.99

4 6.18 ± 0.01 M5 1.75 1.50

5 6.64 ± 0.02 FA2G1[6] 23.28 23.52

6 6.75 ± 0.03 FA2G1[3] 7.56 9.11

7 7.16 ± 0.03 A2G2 / M6 1.55 1.65

8 7.61 ± 0.02 FA2G2 8.19 8.40

9 7.98 ± 0.01 FA2G1S1 3.01 2.73

10 8.46 ± 0.02 A2G2S1 2.56 0.91

11 8.89 ± 0.01 FA2G2S(6)1 - 0.98

12 9.59 ± 0.09 A2G2S2 2.34 1.99

13 9.99 ± 0.17 FA2G2S2 1.46 0.95

14 10.45 ± 0.02 FA3G3S2[6] 1.24 1.14

15 10.67 ± 0.01 FA3G3S2[3,6] 0.95 0.92

Main structures (FA2+FA2G1+FA2G2) 79.62 82.02

Other structures 20.39 17.98

1 Average of the GU value of all samples containing the peak.

2 A – N-Acetylglucosamine, F – fucose, G – galactose, M – mannose, and S – sialic acid.


120

As it can be observed, Clones 18 and 32 produce mAbs with similar glycan profiles, with

one additional peak present at small percentage for Clone 32 (peak 11). Both profiles are in

accordance with the known human IgG1 glycan pattern (Serrato et al. 2004; Takahashi and

Tomiya 1992; Masuda et al. 2000; Raju et al. 2000; Saba et al. 2002), having the three typical

main structures present in the usual order of prevalence: core fucosylated agalactosylated (peak

3, FA2), core fucosylated monogalactosylated (peaks 5 and 6, two isoforms of FA2G1), and core

fucosylated digalactosylated (peak 8, FA2G2) glycans (Routier et al. 1997; Serrato et al. 2007;

Jenkins, Parekh, and James 1996; Jefferis and Lund 1997). Furthermore, the relation between

the two FA2G1 isoforms (peaks 5 and 6) matches the ratio of 3:1 usually displayed in

endogenous human IgG1 (Flynn et al. 2010; Jefferis et al. 1990).

The results of the glycan analysis indicate that the clones selected after transfection with the

OSCARTM system do not differ in the quality of the mAb produced, although having divergent

capacities of production.


121

2.4. CONCLUSION

The OSCARTM expression system appears as a feasible alternative to the more traditional

methods. It is potentially faster for cell line development, by combining selection and

amplification in one step, and provides cells with higher and more stable expression yields,

without significant impact on cell growth characteristics. In this study, its applicability for the

expression of a mAb in CHO-K1 cells was tested. Globally, the OSCARTM technology proved to be

reasonably fast and simple, with mAb-producing cells obtained after 6-8 weeks. However, the

selection of the minigene proved critical, with only the most disabled minigene (pDWM128)

working for the CHO-K1 cells assessed in this work. Furthermore, it was found that the growth

characteristics of the cells are not hindered by the process. However, the relatively high

expression level of 10 pg/cell/day initially obtained, rapid and sharply decayed in the first few

weeks of culture to less than 1 pg/cell/day, and remained low but stable when evaluated over an

extended period of two years.

This study also underlines the strong impact of the methodology of clone selection on the

outcome of the process, advising for the use of productivity analysis instead of the simple and

common comparison of the absorbance values of culture supernatants.

Finally, the glycosylation analysis of the mAb produced by the transfected cells indicates a

similar quality of the product among the clones obtained after transfection with the OSCARTM

expression system.


122

2.5. REFERENCES

Andersen DC, and Reilly DE (2004). Production technologies for monoclonal antibodies and their

fragments. Current Opinion in Biotechnology 15 (5):456-462.

Barnes LM, Bentley CM, and Dickson AJ (2001). Characterization of the stability of recombinant protein

production in the GS-NS0 expression system. Biotechnology and Bioengineering 73 (4):261-270.

Bebbington CR, Renner G, Thomson S, King D, Abrams D, and Yarranton GT (1992). High-level

expression of a recombinant antibody from myeloma cells using a glutamine synthetase gene as an

amplifiable selectable marker. Nature Biotechnology 10 (2):169-175.

Bhopale GM, and Nanda RK (2005). Recombinant DNA expression products for human therapeutic use.

Current Science 89 (4):614-622.



Chusainow J, Yang YS, Yeo JHM, Toh PC, Asvadi P, Wong NSC, and Yap MGS (2009). A study of

monoclonal antibody-producing CHO cell lines: what makes a stable high producer? Biotechnology


Costa AR, Rodrigues ME, Henriques M, Azeredo J, and Oliveira R (2010). Guidelines to cell engineering

for monoclonal antibody production. European Journal of Pharmaceuticals and Biopharmaceutics

74:127–138.

Durocher Y, and Butler M (2009). Expression systems for therapeutic glycoprotein production. Current


Flynn GC, Chen X, Liu YD, Shah B, and Zhang Z (2010). Naturally occurring glycan forms of human

immunoglobulins G1 and G2. Molecular Immunology 47 (11–12):2074-2082.

Gu MB, Todd P, and Kompala DS (1995). Metabolic burden in recombinant CHO cells: effect ofdhfr gene

amplification and lacZ expression. Cytotechnology 18 (3):159-166.

Gu M, Kern J, Todd P, and Kompala D (1992). Effect of amplification of dhfr and lacZ genes on growth

and β-galactosidase expression in suspension cultures of recombinant CHO cells. Cytotechnology 9

(1):237-245.

Jefferis R, and Lund J (1997). Glycosylation of antibody molecules: structural and functional significance.

Chemical Immunology 65 (111-128).

Jefferis R, Lund J, Mizutani H, Nakagawa H, Kawazoe Y, Arata Y, and Takahashi N (1990). A comparative

study of the N-linked oligosaccharide structures of human IgG subclass proteins. The Biochemical

Journal 268 (3):529-537.


123

Jenkins, Nigel, Raj B. Parekh, and David C. James. 1996. Getting the glycosylation right: Implications for

the biotechnology industry. Nature Biotechnology 14 (8):975-981.

Li H, and d’Anjou M (2009). Pharmacological significance of glycosylation in therapeutic proteins. Current


Lonza (2009). GS Gene Expression System [cited April 14 2009]. Available from http://www.lonza.com/g

roup/en/products_services/custommanufactoring/mammalian/geneexpressions.html.

http://www.lonza.com/custom-manufacturing/development-technologies/gs-gene-expression-system.aspx

Masuda K, Yamaguchi Y, Kato K, Takahashi N, Shimada I, and Arata Y (2000). Pairing of oligosaccharides

in the Fc region of immunoglobulin G. FEBS letters 473 (3):349-357.

Melton DW, Ketchen A-N, Kind AJ, McEwan C, Paisley D, and Selfridge J (2001). A one-step gene

amplification system for use in cultured mammalian cells and transgenic animals. Transgenic

Research 10 (2):133-142.

Mohan CM, Kim Y-G, Koo J, and Lee GM (2008). Assessment of cell engineering strategies for improved

therapeutic protein production in CHO cells. Biotechnology Journal 3 (5):624-630.

Nicholson W. OSCARTM: High yielding and rapid protein expression system. OSCARTM gene expression

system: Information Pack.

Page MJ, and Sydenham MA (1991). High level expression of the humanized monoclonal antibody

CAMPATH-1H in Chinese hamster ovary cells. Nature Biotechnology 9 (1):64-68.

Raju TS, Briggs JB, Borge SM, and Jones AJS (2000). Species-specific variation in glycosylation of IgG:

evidence for the species-specific sialylation and branch-specific galactosylation and importance for

engineering recombinant glycoprotein therapeutics. Glycobiology 10 (5):477-486.

Routier F, Davies M, Bergemann K, and Hounsell E (1997). The glycosylation pattern of a humanized IgGl

antibody (D1.3) expressed in CHO cells. Glycoconjugate Journal 14 (2):201-207.

Saba JA, Kunkel JP, Jan DCH, Ens WE, Standing KG, Butler M, Jamieson JC, and Perreault H (2002). A

study of immunoglobulin G glycosylation in monoclonal and polyclonal species by electrospray and

matrix-assisted laser desorption/ionization mass spectrometry. Analytical Biochemistry 305 (1):16-

31.




Serrato JA, Hernández V, Estrada-Mondaca S, Palomares LA, and Ramírez OT (2007). Differences in the

glycosylation profile of a monoclonal antibody produced by hybridomas cultured in serum-

supplemented, serum-free or chemically defined media Biotechnology and Applied Biochemistry

47:113-124.


124

Takahashi N, and Tomiya N (1992). Analysis of N-linked oligosaccharides: application of glycoamidase A.

In Handbook of Endoglycosidases and Glycoamidases edited by N Takahashi and T Muramatsu.

Boca Raton, FL: CRC Press.

Werner RG, Noe W, Kopp K, and Schlüter M (1998). Appropriate mammalian expression systems for

biopharmaceuticals. Arzneimittel-Forschung 48 (8):870-880

Wurm FM (2004). Production of recombinant protein therapeutics in cultivated mammalian cells. Nature



Impact of the process of cell adaptation to

serum-free conditions on product quality

CC

THE PRESENT CHAPTER WAS ADAPTED FROM THE FOLLOWING RESEARCH PAPERS

ANA RITA COSTA, Joanne Withers, Maria Elisa Rodrigues, Niaobh McLaughlin,

Mariana Henriques, Rosário Oliveira, Pauline Rudd, Joana Azeredo (2012). The

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

monoclonal antibody produced by CHO cells. Submitted to New Biotechnology.

Maria Elisa Rodrigues, ANA RITA COSTA, Mariana Henriques, Philip Cunnah, David

Melton, Joana Azeredo, Rosário Oliveira (2012). Advances and drawbacks of the

adaptation to serum-free culture of CHO-K1 cells for monoclonal antibody

production. Submitted to Applied Biochemistry and Biotechnology.

127

lycosylation is one of the most critical parameters affecting the biological activity of

therapeutic proteins, and should be closely monitored and controlled to guarantee a consistent

and high-quality product in biopharmaceutical processes. This requires understanding the factors

influencing glycosylation during production. Among these, the adaptation of producer cells,

typically mammalian, to culture in the serum-free conditions required for regulatory and safety

reasons is one of the most challenging steps, with potential effects on product quality.

Considering this, the work described in this chapter evaluated a process of gradual adaptation of

monoclonal antibody (mAb)-producing Chinese hamster ovary (CHO-K1) cells to serum-free

culture using different combinations of medium supplements, and assessed its impact on the

glycosylation pattern of the product.

The selection of an adequate combination of supplements proved to be critical for the

success of adaptation. Furthermore, analysis by high-performance liquid chromatography

revealed important changes in the glycosylation patterns of the mAb for all the steps of serum

reduction, which could be grouped into intermediate (2.5 % to 0.15 % serum) and final (0.075 %

and 0 % serum) stages. The intermediate levels of serum demonstrated the advantageous

increase of galactosylation and decrease of fucosylation, but also an undesirable increase in

sialylation. The inverse was obtained at the final stages of adaptation. These divergences could

be related to the reduction of serum supplementation, to variations observed in the levels of cell

density and viability achieved at these stages, as well as to the fact that cells naturally shifted

their mode of growth from adherent to suspended. The divergent glycan profiles obtained in this

study demonstrate a strong influence of the process of adaptation on mAb glycosylation,

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

stages of process development.

Keywords: Monoclonal antibody; Chinese hamster ovary cells; Serum-free; Adaptation;

Glycosylation

GG

CHAPTER 3 │ IMPACT OF SERUM-FREE ADAPTATION ON PRODUCT QUALITY

129

3.1. INTRODUCTION

Immunoglobulin G (IgG) monoclonal antibodies (mAbs) are currently one of the most

successful drug classes in the biopharmaceutical industry (Primack, Flynn, and Pan 2011; Reichert

et al. 2005; Carter 2006; Reichert and Valge-Archer 2007). For the production of these therapeutic

mAbs, mammalian cells are the preferred system due to their innate ability to perform post-

translational processing that closely resembles that found in humans (Rodrigues et al. 2009;

LeFloch et al. 2006; Hossler, Khattak, and Li 2009; Primack, Flynn, and Pan 2011). Of the post-

translational modifications, glycosylation is one of the most critical for protein quality (Hossler,

Khattak, and Li 2009). In particular, the glycosylation of the Fc region of the IgG plays essential

roles in structural integrity, effector functions, immunogenicity, plasmatic clearance, solubility,

and resistance toward proteases (Hossler, Khattak, and Li 2009; Lowe and Marth 2003; Mimura

et al. 2000; Davies et al. 2001; Krapp et al. 2003; Magdelaine-Beuzelin et al. 2007).

These functions and, consequently, the therapeutic efficacy of the mAb are also dependent

on the specific glycoforms present (Elliott et al. 2003; Primack, Flynn, and Pan 2011), which are

highly variably due to the different types of monosaccharides available and their combinational

pairing (Hashimoto et al. 2006; Fernandes 2005). For example, terminal galactose, core fucose,

sialic acid, bisecting N-acetylglucosamine (GlcNAc), and high mannose species are known as

important glycosylation elements that affect the effector functions of the antibody (Tsuchiya et al.

1989; Shields et al. 2002; Goetze et al. 2012; Kanda et al. 2007). Indeed, higher levels of

galactosylation (Abès and Teillaud 2010; Serrato et al. 2007; Jefferis and Lund 1997;

Hodoniczky, Zheng, and James 2005) and bisecting GlcNAc (Ferrara et al. 2006; Fernandes

2005), as well as reduced core fucosylation (Sibéril et al. 2006; Shields et al. 2002; Shinkawa et

al. 2003) and sialylation (Scallon et al. 2007; Kaneko, Nimmerjahn, and Ravetch 2006; Anthony

and Ravetch 2010; Anthony et al. 2008), 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 killing. This is particularly relevant for core

fucosylation, whose absence can improve ADCC activity by up to 100-fold (Shields et al. 2002).

Additionally, the presence of high mannose structures has been associated with reduced ADCC

and CDC (Kanda et al. 2007), and with higher clearance rates (Abès and Teillaud 2010).

The glycan profile (glycoforms) of glycosylated therapeutics can be affected by several factors


130

that include: the host cell expression system (Jenkins, Parekh, and James 1996; Sethuraman and

Stadheim 2006; Raju 2003); environmental factors such as dissolved oxygen (Chotigeat et al.

1994; Serrato et al. 2004), pH (Müthing et al. 2003; Yoon et al. 2005), and temperature (Yoon,

Kim, and Lee 2003; Trummer et al. 2006); as well as the cell culture media and mode of culture

(e.g. batch, fed-batch, and continuous) (Gu et al. 1997; Patel et al. 1992; Wong et al. 2005; Kunkel

et al. 2000; Jenkins, Parekh, and James 1996; Andersen and Goochee 1994).

The cell culture media is typically a mixture of 50-100 chemically defined components and a

handful of undefined components that are supplemented (Hossler, Khattak, and Li 2009). Of

these, serum is the most common for its positive effects on mammalian cell growth. However, its

ill-defined composition and animal origin raises several concerns, including lot-to-lot variability,

complex purification of the product, risk of transmission of diseases to humans, and high costs

(Rodrigues et al. 2009; Brunner et al. 2010; LeFloch et al. 2006; Zhang and Robinson 2005; Liu

and Chang 2006). Therefore, due to regulatory and safety issues, chemically-defined culture

conditions, without supplementation of serum or any other animal-derived components, are

considered fundamental to the manufacture of highly purified products of consistent quality

(Serrato et al. 2007; Ozturk and Hu 2006; Geigert 2004). Consequently, mammalian cells that

typically grow in serum-containing media need to be adapted to serum-free conditions, a process

that usually consists of progressive reductions of serum concentration in the medium to consent

the self-adjustment of cells to the new environment and increase the possibility of a successful

adaptation (Ozturk and Palsson 1991; Doyle and Griffiths 1998; Kovář 1989; Griffiths 1987).

Additional supplementation of the culture medium with commercial serum substitutes, defined

proteins, or small molecules such as amino acids and trace elements can also be attempted to

support adaptation (Doyle and Griffiths 1998). This process of adaptation is very time-consuming

and can affect cell growth, product yield, and product quality (e.g. glycosylation) (Crowell et al.

2007; Doyle and Griffiths 1998; Ozturk et al. 2003; Griffiths 1987). Therefore, to assure a

consistent quality of glycosylated therapeutics that meet the requirements of the regulatory

agencies, it is essential to understand how these processes at early phases of development affect

glycosylation (Raju 2003). In the present work, mAb-producing CHO-K1 cells were adapted to

culture under serum-free conditions using a gradual methodology of adaptation that involved the

additional test of different combinations of supplements. The effect of each step of the process

on mAb glycosylation was assessed by high performance liquid chromatography, and the

potential effects on biological activity were discussed.


131


3.2.1. CELL LINE AND CULTURE CONDITIONS

MAb-producing Chinese hamster ovary (CHO) cells, previously obtained (Costa et al. 2012)

through the transfection of a CHO-K1 cell line (CCL-61, American Type Culture Collection, Spain)

for the production of a recombinant human mAb (CAB051, Biotecnol SA, Portugal) were used in

this work. These cells were grown in Dulbecco’s Modified Eagle’s Medium (DMEM, Sigma-Aldrich,

United Kingdom) supplemented with 10 % fetal bovine serum (FBS, Sigma-Aldrich) and 1x

hypoxanthine aminopterin thymidine (HAT, Sigma-Aldrich) at 37 ºC and 5 % carbon dioxide (CO2).

3.2.2. CELL ADAPTATION TO SERUM-FREE CONDITIONS

The mAb-producing CHO-K1 cells were gradually adapted from growth in serum-containing

DMEM to serum-free conditions. For this, cells were seeded into 24-well plates or 25 cm2 culture

flasks (Frilabo, Portugal) at a viable density of 2x105 cells/mL or 4x105 cells/mL, respectively, in

DMEM with 10 % FBS, 1x HAT and additional supplementation with one of five combinations of

insulin and trace elements (Sigma-Aldrich), whose composition is described in Table 3.1.

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 DHFR- medium (Sigma-Aldrich). The remaining steps of

adaptation were performed in EXCELL. A control of mAb-producing CHO-K1 cells subjected to the

same process of adaptation without the support of a combination of supplements was included in

the study.

During this process of adaptation, cell concentration and viability were assessed using a

haemocytometer and the trypan blue (Sigma-Aldrich) exclusion method.


132

TABLE 3.1. Composition of the five combinations of supplements assessed for the adaptation of monoclonal

antibody-producing CHO-K1 cells to growth in serum-free conditions

SUPPLEMENT FUNCTION COMBINATION

1 2 3 4 5

Recombinant human insulin (rhInsulin) Growth Factor X X X X X

Copper sulfate (CuSO4) Proliferation X X X X

Zinc sulfate (ZnSO4) Proliferation X X X X

Sodium selenite (Na2SeO3) Antioxidant X X X X

Ammonium iron citrate (FeC6H5O7.NH4OH) Proliferation X X

Ferrous ammonium sulfate (NH4Fe(SO4)2) Proliferation X X

Ammonium metavanadate (NH4VO3) Unknown X X X

Nickel chloride (NiCl2) Unknown X X X

Stannous chloride (SnCl2) Unknown X X X


During the adaptation to serum-free conditions several samples were taken to assess the

glycosylation profile of the mAb produced. A description of these samples including the stage of

serum, media and combination of supplements, and sampling day, is provided in Table 3.2.

Combinations C1 and C4 were not included in this analysis due to an early exclusion from the

study (see section 3.3).

To note that in the following procedures all water used was of type 1 ultra-pure, resistivity of

18 MΩ.cm, particle-free (> 0.22 µm), and total organic content of less than ten parts per billion.


133

TABLE 3.2. Description of the samples taken at different stages of the process of adaptation of monoclonal antibody-

producing CHO-K1 cells to serum-free medium, for the evaluation of the glycosylation profile of the monoclonal

antibody produced

STAGE OF SERUM MEDIUM COMBINATIONS OF SUPPLEMENTS SAMPLING DAY SAMPLE NAME

10 % 1 DMEM None 1 D 10 %

5 % DMEM C2, C3, C5 3 -

2.5 % DMEM

C2

12

C1 D 2.5 %

C3 C2 D 2.5 %

C5 C3 D 2.5 %

1.25 % DMEM C2, C3, C5 20 -

0.625 %

DMEM C2, C5 - -

C3 24 C3 D 0.625 %

75% DMEM: 25% EXCELL C2, C3, C5 - -

50% DMEM: 50% EXCELL C2, C3, C5 - -

25% DMEM: 75% EXCELL C2, C3, C5 - -

EXCELL

C2

66

C1 E 0.625 %

C3 C2 E 0.625 %

C5 C3 E 0.625 %

0.31 % 3 EXCELL

C2, C5 2 - -

C3 86 C3 E 0.31 % i

167 C3 E 0.31 % f

0.15 % EXCELL C3 194 C3 E 0.15 %

0.075 % EXCELL C3 237 C3 E 0.075 %

0 % EXCELL C3 280 C3 E 0 %

1 Normal serum-supplemented culture conditions.

2 Cells growing in combinations C2 and C5 died at this stage.

3 This stage was the most time-consuming, so samples at the initial (i) and final (f) phases were assessed.

ANTIBODY PURIFICATION

Prior to IgG purification, samples were concentrated to volumes of 200 µL using 15 mL spin

concentrators of 5 KDa molecular weight cutoff (Agilent Technologies, Ireland). Then, the IgG

from the concentrated samples was purified with Protein A spin plates (Thermo Scientific,

Ireland) according to manufacturer instructions, with slight modifications. Briefly, the wells of a

spin plate were equilibrated with two washes of 200 µL of 1x PBS (pH 7.2, Merck, UK). Samples

were diluted two-fold in 1x PBS and added (200 µL) to the plate for a period of incubation of 30

min with moderate agitation. The resin was washed three times with 500 µL of PBS, and the


134

purified IgG eluted three times with 300 µL of 0.5 M acetic acid (pH 2.5, AnalaR, UK) and

neutralized with 20 µL of 1 M ammonium bicarbonate (pH 7-8.5, Sigma-Aldrich). The neutralized

elutions were dried overnight in a vacuum centrifuge.











(Sigma-Aldrich) and 20 mM sodium bicarbonate (pH 7, Merck, UK), vortexed and mixed for 10

min after each wash, and dried in a vacuum centrifuge.


of 2 µL of PNGase F (Prozyme, USA) and 48 µL of 20 mM sodium bicarbonate (pH 7), and a

further 12-16 h at 37 ºC after adding extra 50 µL 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 sonication for each wash. The supernatants

obtained were collected and dried in a vacuum centrifuge. The dried samples were redissolved in

20 µL formic acid (1 %, Sigma-Aldrich), incubated for 40 min at room temperature and dried in a

vacuum centrifuge.






135


tips (PhyNexus, United States of America). Briefly, samples were diluted in 95 µL of water and

900 µL of acetonitrile. The PhyNexus tips were prepared by washing with ten 500 µL uptakes of

95 % acetonitrile (v/v in water), ten 500 µL uptakes of 20 % acetonitrile (v/v in water), and

another ten 500 µL uptakes of 95 % acetonitrile. Samples were loaded into the PhyNexus tips

through ten 1 mL in-out cycles, followed by washing of the tips with ten 1 mL uptakes of 95 %

acetonitrile. The glycans were then eluted with five uptakes of 200 µL of 20 % acetonitrile, and

the elutions collected and dried in a vacuum centrifuge.



was resuspended in 20 µL of water and 80 µL of acetonitrile. NP-HPLC was performed using a
















IgG1 profile as a guide. A confirmation of the structures through exoglycosidase digestion was not


136

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 relative percentage of N-linked glycans was calculated for

agalactosylated (G0), monogalactosylated (G1), digalactosylated (G2), trigalactosylated (G3),

galactosylated (G1+G2+G3), core fucosylated (F), core fucosylated agalactosylated (FG0), core

fucosylated monogalactosylated (FG1), core fucosylated digalactosylated (FG2), core fucosylated

trigalactosylated (FG3), monosialylated (S1), disialylated (S2), trisialylated (S3), sialylated

(S1+S2+S3), monosialylated monogalactosylated (S1G1), monosialylated digalactosylated

(S1G2), disialylated digalactosylated (S2G2), disialylated trigalactosylated (S2G3), high mannose,

and complex glycan structures. The relative percentages were determined by summing the

percentage of area of the peaks pertaining to each of the abovementioned structures.


137


In this work, mAb-producing CHO-K1 cells were adapted to growth in serum-free conditions

and the impact of the process of adaptation on the glycosylation profile of the mAb produced was

evaluated.

For the adaptation to serum-free conditions, a gradual methodology was implemented. A

first attempt of this process was not successful, with cells dying before reaching the expected

serum-free growth. Indeed, it was observed that cells growing in medium supplemented with the

combinations of supplements C1 and C4 showed a decrease in viability very early in the process

(at 2.5 % serum) and died on day 29 of culture at the stage of 0.625 % serum. For the remaining

combinations of supplements (C2, C3, and C5) cells were able to reach the next level of serum

(0.31 %), but eventually died on day 45 of culture. The early cell death observed for combinations

C1 and C4 is most likely related to the trace element ammonium iron citrate that is present in

these combinations and absent in C2, C3, and C5.

The initial failed attempt at adaptation was associated with the following problems: (i) the

use of a low initial cell concentration (2x105 cells/mL) and of small culture flasks (24-well plates),

which resulted in a low number of cells surviving at each step of the demanding process of

adaptation; (ii) employing subculture procedures that could be damaging to the cells, such as

centrifugation and the use of enzymes (e.g. trypsin), which resulted in decreased cell viability and

morphological alterations (smaller and irregular cells); (iii) insufficient amount of time given at

each step of the process, which prevented a full cell adaptation.

The abovementioned problems were amended in a second attempt at cell adaptation to

growth in the absence of serum. So, higher initial cell concentration (4x105 cells/mL) and larger

culture flasks (25 cm2 T-flasks) were used, and enough time was given for the recovery of a

healthy appearance (viability) and proliferation of the cells at each step of serum reduction.

Furthermore, combinations of supplements C1 and C4 were not used in this second attempt due

to their observed worse effects on cell adaptation.

After these changes in the procedure, it was possible to successfully obtain cells fully

adapted to growth in serum-free conditions. The time spent at each step of the process and the

main events observed are shown in Table 3.3.


138

TABLE 3.3. Timeline of the process of adaptation of the monoclonal antibody-producing CHO-K1 cells to serum-free

culture conditions

SERUM MEDIUM DAYS OF CULTURE OBSERVATIONS

10 % DMEM 0 – 1 (1)

5 % DMEM 1 – 3 (2)

2.5 % DMEM 3 – 12 (9)

1.25 % DMEM 12 – 20 (8)

0.625 % 20 – 66 (46) total

100 % DMEM 20 – 24 (4)

75 % DMEM: 25% EXCELL 24 – 26 (2)

50 % DMEM: 25 % EXCELL 26 – 28 (2)

25 % DMEM: 75 % EXCELL 28 – 30 (2)

100 % EXCELL 30 – 66 (36)

0.31 % EXCELL 66 – 167 (101)

72 Cells started growing detached

146 Death of cells growing in C2 and C5, as well

as of the control without supplementation

0.15 % EXCELL 167 – 194 (27)

0.075 % EXCELL 194 – 237 (43)

0 % EXCELL 237 – 280 (43) Cells became fully adapted

Approximately 280 days were necessary to reach the stage of growth in full serum-free

conditions. The first four stages of adaptation, (10, 5, 2.5, and 1.25 %) were the least

contributors for this extended period of time, with cells maintaining their growth and

morphological characteristics. Likewise, at the stage of 0.625 % serum, the shift from growth in

DMEM to growth in the chemically-defined EXCELL medium occurred smoothly and was achieved

in just 10 days. After this point, the process of adaptation became more time-consuming,

requiring a greater effort from the cells to adapt. This was particularly critical at the stage of 0.31

% serum, which lasted for 101 days, mainly due to the changes occurring in the mode of growth

of the cells. Indeed, the percentage of serum supplemented in the medium at this point was

probably not providing enough specific factors that aid cell adhesion (e.g. fibronectin, vitronectin,

and laminin), which resulted in the shift of the cell growth mode from adherent to suspended

(without rocking). This was further enhanced by the use of EXCELL, a medium that is specifically


139

formulated for the culture of suspended cells. This change in cell behavior had the advantage of

allowing the expansion to new culture flasks without having to resort to the typical procedures of

splitting required in adherent culture (using trypsin and centrifugation) that have already shown to

be harmful to the cells during adaptation.

It was also at this stage of 0.31 % serum that differences between the combinations of

supplements were noticeable. Cells growing in medium supplemented with C2 and C5

demonstrated a clear decline of proliferation and viability over time, ultimately dying at day 146.

Cells growing in medium without addition of any combination of supplements (control, data not

shown) had a similar behavior and also died at this point, showing that the use of combinations

C2 and C5 had no advantage for adaptation. In contrast, combination C3 proved to be

advantageous for the adaptation of the mAb-producing CHO-K1 cells to serum-free culture,

allowing a full adaptation. It is therefore clear that the composition and concentration of the

supplements used in the growth medium strongly influence the ability of the cells to adapt to

serum-free conditions, as already seen in previous studies with PER.C6, HEK293, HeLa, CHO-S,

and C6 rat glioma cells (Jacobia et al. 2006; Ho and Ames 2002).

It is somewhat unexpected that only combination C3 allowed the full adaptation of the cells,

since the trace elements copper, zinc, selenite, and iron, present in the remaining combinations,

are often reported in the literature as the most important trace elements to use in media

optimization (Jacobia et al. 2006; Ho and Ames 2002; Haase and Beyersmann 1999; Bai et al.

2011; Zhang, Robinson, and Salmon 2006). However, for the cells in study, the less researched

trace elements of nickel, stannous, and metavanadate, whose particular functions are not yet

understood, were the most valuable. A possible explanation may be the fact that the influence of

the trace elements in the culture is not simply additive, so the unexpected results may be a

consequence of synergistic and antagonistic effects caused by their interaction (Jacobia et al.

2006).

After adapting to the challenging stage of 0.31 % serum, the cell viability increased and

persisted at high levels during the remaining steps of the process, resulting in an easier and

faster adaptation, as proved by the shorter time spent in the latter stages. Furthermore, a visible

improvement of cell proliferation was observed at the end of the step of 0.075 % serum and

during 0 %, reaching levels similar to those observed during cultivation in the typical 10 % serum-

supplemented conditions, which indicates a complete adaptation of the cells to the new culture

conditions.


140

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

mAb produced by the CHO-K1 cells was affected during the adaptation to serum-free conditions.

In Figure 3.1 it is possible to observe the profile obtained at the different stages of this process.

To note that for the combinations of supplements C2 and C5, only two samples were assessed

(D 2.5 % and E 0.625 %) due to the early death of cells at the stage of 0.31 % serum. In addition,

for combination C3, the extended time of adaptation to stage 0.31% serum led to the analysis of

samples from both the initial (E 0.31% i) and final (E 0.31% f) phases of this step.

FIGURE 3.1. N-glycosylation profile of the monoclonal antibody produced by CHO-K1 cells cultured in different

conditions during a process of gradual adaptation to serum-free medium. D – DMEM, E – EXCELL, Cx – combination

of supplements x (x = 2, 3, and 5), % referent to the percentage of serum supplemented in the culture medium.

The N-glycosylation profile of the mAb produced in the normal serum-supplemented culture

conditions (D 10 %) was in accordance with the known human IgG1 profile (Serrato et al. 2004;

C3 D 0.625 %

C3 E 0.625 %

C3 E 0.31 % i

C3 E 0.31% f

C3 E 0.15 %

C3 E 0.075 %

C3 E 0 %

15 20 25 30 35 40 45 50

6

C2 D 2.5 %

C2 E 0.625 %

C5 D 2.5 %

C5 E 0.625 %

C3 D 2.5 %

D 10 %

1 2

4

5

8

9

11 23222112

13 14 16 17 18 20

19

3

10

7

15

15 20 25 30 35 40 45 50

Time (minutes)

Time (minutes)


141

Takahashi and Tomiya 1992; Masuda et al. 2000; Raju et al. 2000; Saba et al. 2002), with the

three main structures present: core fucosylated agalactosylated (peak 4, FG0),

monogalactosylated (peaks 8 and 9, two isoforms of FG1), and digalactosylated glycans (peak

12, FG2) (Routier et al. 1997; Serrato et al. 2007; Jenkins, Parekh, and James 1996; Jefferis

and Lund 1997). Moreover, the relation between the two FG1 isoforms (peaks 8 and 9) matches

the ratio usually displayed in endogenous human IgG1 (approximately 3:1) (Flynn et al. 2010;

Jefferis et al. 1990). However, it is clear from Figure 3.1 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 (Mimura et al. 2001; Kanda et al. 2007).

Table 3.4 presents more detailed data for the analysis of the modifications observed in Figure

3.1, including the glucose unit (GU) values, tentative structure assignments, and percentage of

area of the peaks obtained at the different stages of the process of cell adaptation.


142

TA

BLE

3.4

. Data of mean glucose unit (GU) value, tentative assignment, and percentage of area of the different peaks detected on the monoclonal antibody produced by CHO-

K1 cells during the process of adaptation to serum-free conditions. D – DMEM, E – EXCELL, GU – glucose units, Cx – combination of supplements x (x = 2, 3, 5), % referent

to the percentage of serum supplemented in the culture media

PEAK

MEAN GU 1

TENTATIVE

ASSIGNMENT

2

% AREA

D

10 %

C2 D

2.5 %

C2 E

0.625 %

C5 D

2.5 %

C5 E

0.625 %

C3 D

2.5 %

C3 D

0.625 %

C3 E

0.625 %

C3 E

0.31 % i

C3 E

0.31 % f

C3 E

0.15 %

C3 E

0.075 %

C3 E

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

FA

2 4

0.99

41

.51

15

.34

1

8.14

1

9.2

4

29.3

1

37.

53

19.3

6

22.6

1

25.1

2

14.1

9

37.6

9

50.2

1

5 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

FA

2G1

[6]

23.

52

24.2

1

17.8

6

14.

96

15.

31

18

.50

2

3.42

13

.75

20

.31

21

.67

10

.10

26

.27

23

.27

9 6.75 ± 0.03

FA

2G1

[3]

9.11

6.

41

23.8

9

27.

84

8.66

6

.21

1

0.38

6.2

9

9.20

9.

31

8.4

2

8.3

0

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

1.94

12

7.61 ± 0.02

FA

2G2

8.

40

8.55

7.03

11.15

10.

45

7.26

9.24

8.04

12.7

9

9.27

8.50

10.0

0

7.34

13

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

11.30




143

The results shown in Table 3.4 confirm the variation on the proportions of the main

structures observed in Figure 3.1 for some samples. Indeed, cells growing in C5 E 2.5 % produce

a mAb where FA2G1 is the main glycan. Furthermore, in some cases the typically less abundant

structures become more prevalent than the usual three main glycans, as is the case of A1G1 for

samples C5 D 2.5 %, C3 D 2.5 %, C2 E 0.625 %, and C3 0.625 %, all at the expense of the

FA2G2 structure. For C3 E 0.15 %, FA3G3S2 and A2G2S2 substitute FA2G1 and FA2G2 as the

main glycans.

It is curious to note that the reduction on the global 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, as mentioned, 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. The low cell viabilities may cause the accumulation of extracellular glycosidases in

the medium, and potentially alter the glycan profile of the mAb produced (Gramer and Goochee

1993, 1994). 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.

It should be noted that structures with higher GU values, corresponding to glycans

containing sialic acids, are absent in some of the 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, core fucose, sialic acid and

bisecting N-acetylglucosamine (GlcNAc) (Cabrera et al. 2005; Kanda et al. 2007). Concerning the

latter, it is known that CHO cells lack the enzyme N-acetylglucosaminyl transferase III (GnT III)

that mediates the transfer of bisecting GlcNAc to complex glycans, so no particular attention will

be given to this element. In 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 3.5 shows the levels of the abovementioned functional elements

composing the mAb obtained at the different stages of the process of adaptation to serum-free

conditions.


144

TA

BLE

3

.5. 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-free conditions. D – DMEM, E – EXCELL, Cx – combination of supplements x (x = 2, 3, 5), % referent to the percentage of serum supplemented

in the culture media

GLYCAN STRUCTURE

% AREA

D

10 %

C2 D

2.5 %

C2 E

0.625 %

C5 D

2.5 %

C5 E

0.625 %

C3 D

2.5 %

C3 D

0.625 %

C3 E

0.625 %

C3 E

0.31 % i

C3 E

0.31 % f

C3 E

0.15 %

C3 E

0.075 %

C3 E

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

8.31

Trigalactosylated (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 fucosylated agalactosylated (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 fucosylated monogalactosylated (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 fucosylated digalactosylated (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 fucosylated trigalactosylated (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

2.16

Monosialylated (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

Monosialylated monogalactosylated (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

Monosialylated digalactosylated (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

Disialylated digalactosylated (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

Disialylated trigalactosylated (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

96.99

High mannose structures

2.32

2.52

2.73

1.04

1.55

5.31

6.21

5.91

4.13

5.01

2.46

2.50

3.01


145

GALACTOSYLATION

Galactosylation is one of the most studied glycosylation features of any glycoprotein (Pučić et

al. 2011). 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 (Abès

and Teillaud 2010; Serrato et al. 2007), since agalactosylated (G0) structures in antibodies are

correlated with several pathologies (Parekh et al. 1985). Antibodies with higher galactose content

could have a more effective action through improvements of CDC (Hodoniczky, Zheng, and

James 2005; Boyd, Lines, and Patel 1995).

In the present study, the levels of galactosylation of the mAb obtained in the normal culture

conditions (D 10 %, 53 % galactosylation) are comparable to those obtained previously for normal

human IgG (Serrato et al. 2007; Routier et al. 1998). These levels are increased for serum levels

between 0.625 % and 0.15 % for the three combinations of supplements assessed (C2, C3, and

C5). 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 (Lund et al. 1993) or low density static culture (Kumpel et al.

1994); 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 (Davies et al. 2001; Hodoniczky, Zheng, and James

2005; Boyd, Lines, and Patel 1995). However, with further reduction of serum percentage in the

medium (C3 E 0.075% and C3 E 0%), galactosylation decreases to levels below those observed at

the initial culture conditions. 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 accordance with previous works concerning CHO (Lifely et al. 1995) and hybridoma

cells (Serrato et al. 2007), and may reduce the efficacy of the mAb.

FUCOSYLATION

Core fucosylation of IgG mAbs has been intensively studied as a consequence of its role in

ADCC (Pučić et al. 2011). Fucose interferes with binding of the IgG to the FcyRIIIa receptor and

dampens its ability to destroy target cells through ADCC (Hossler, Khattak, and Li 2009;

Nimmerjahn and Ravetch 2008; Kanda et al. 2007; Kanda et al. 2006). Consequently, lack of

core fucose enhances the clinical efficacy of mAbs that exert their therapeutic effect by ADCC


146

mediated killing (Sibéril et al. 2006; Shields et al. 2002; Shinkawa et al. 2003). Indeed, it has

been shown that IgGs deficient in core fucose have ADCC activity enhanced by up to 100-fold

(Shields et al. 2002).

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 % (serum-free culture) 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 (Raju

2003).

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

(FG0/FG1/FG2) structures has been observed (Jones et al. 2003; Beck et al. 2005).

Interestingly, this ratio is not observed in the mAb produced in the initial culture conditions (D 10

%), 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 %.

SIALYLATION

The effects of sialylation of IgG have recently attracted much attention, since it was shown

that the addition of sialic acid monosaccharides dramatically changes the physiological role of

IgGs by converting them from pro-inflammatory into anti-inflammatory agents (Kaneko,

Nimmerjahn, and Ravetch 2006; Anthony and Ravetch 2010; Anthony et al. 2008). Reports have

shown that higher sialylation of IgG associates with low ADCC, due to reduced binding to specific

Fc receptors such as FcγRIII (Scallon et al. 2007).

In general, the degree of sialylation reported for IgG has been below 20 % (Pučić et al. 2011;

Arnold et al. 2007; Serrato et al. 2007). These values are in accordance with those obtained in

this 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). In contrast,

sialylated structures are less present at C2 E 0.625 %, C5 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 (S1) as

the most prevalent sialylated structure. These results of lower mAb sialylation in serum-free


147

conditions when compared to serum-containing media are in accordance to previous studies with

hibridoma cells (Serrato et al. 2007), and suggest an improvement of mAb efficacy through

ADCC.

OTHER ELEMENTS

In 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 (Kanda et al. 2007), as well as with a rapid IgG clearance from

serum (Abès and Teillaud 2010). 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.

Although the presence of IgG from fetal bovine serum in the samples could have influenced

the glycosylation profiles obtained, the likelihood of such contamination in this work was reduced

since the methodology used for IgG purification had a high specificity for human IgG while

providing weak purification for bovine IgG. 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 FA2G2 peak) (Serrato et al. 2007). Therefore, the

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


148

3.4. CONCLUSION

The establishment of processes for the large-scale production of therapeutic proteins, like

mAbs, 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 therapeutic activity is not affected. One of the most challenging and time-consuming

steps of process development is the adaptation of producer cells to serum-free conditions, with

potential impact on product quality. In this work, the effects of such a process of adaptation on

the glycosylation profile of a mAb produced by CHO-K1 cells were evaluated.

Concerning the adaptation itself, the results of this work highlight the importance of selecting

adequate medium supplements to support cell adaptation to serum-free culture, with only one of

the five combinations assessed being successful (C3). Additionally, the amount of time given at

each step of the process has also proved to be a determinant factor.

As for glycosylation, it was clear that cell adaptation to serum-free conditions had a strong

impact on mAb glycosylation. The initial profile (DMEM, 10 % serum) was in accordance with the

typical IgG1 profile, but the prevalence of the main structures (fucosylated agalacto-, monogalacto-

and digalactosylated) 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 different behaviors of mAb

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

increased and fucosylation reduced, both having a positive impact on the clinical efficacy of the

mAb through improvements of ADCC and CDC activity. Conversely, sialylation increased and

exceeded the levels usually reported (below 20 %), particularly for combination C3 with EXCELL at

0.15 % serum (45.52 % sialic acid content), potentially reducing ADCC and, consequently, the mAb

efficiency. At the final steps of adaptation, in serum-free medium, fucosylation was similar to that of

the initial serum-supplemented conditions, but galactosylation and sialylation were decreased, with

divergent potential effects (negative and positive, respectively) on the biological activity of the mAb.

The divergences 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 characteristics are not easy to predict

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 at these initial stages of process development.


149

3.5. REFERENCES

Abès R, and Teillaud J-L (2010). Impact of glycosylation on effector functions of therapeutic IgG.


Andersen DC, and Goochee CF (1994). The effect of cell-culture conditions on the oligosaccharide

structures of secreted glycoproteins. Current Opinion in Biotechnology 5 (5):546-549.

Anthony RM, Nimmerjahn F, Ashline DJ, Reinhold VN, Paulson JC, and Ravetch JV (2008). Recapitulation

of IVIG anti-inflammatory activity with a recombinant IgG Fc. Science 320 (5874):373-376.

Anthony R, and Ravetch J (2010). A novel role for the IgG Fc glycan: the anti-inflammatory activity of

sialylated IgG Fcs. Journal of Clinical Immunology 30 (0):9-14.

Arnold JN, Wormald MR, Sim RB, Rudd PM, and Dwek RA (2007). The impact of glycosylation on the

biological function and structure of human immunoglobulins. Annual Review of Immunology 25

(1):21-50.

Bai Y, Wu C, Zhao J, Liu Y-H, Ding W, and Ling WLW (2011). Role of iron and sodium citrate in animal

protein-free CHO cell culture medium on cell growth and monoclonal antibody production.


Beck A, Bussat M-C, Zorn N, Robillard V, Klinguer-Hamour C, Chenu S, Goetsch L, Corvaïa N, Van

Dorsselaer A, and Haeuw J-F (2005). Characterization by liquid chromatography combined with

mass spectrometry of monoclonal anti-IGF-1 receptor antibodies produced in CHO and NS0 cells.

Journal of Chromatography B 819 (2):203-218.

Boyd PN, Lines AC, and Patel AK (1995). The effect of the removal of sialic acid, galactose and total

carbohydrate on the functional activity of Campath-1H. Molecular Immunology 32 (17-18):1311-

1318.

Brunner D, Frank J, Appl H, Schöffl H, Pfaller W, and Gstraunthaler G (2010). Serum-free cell culture: the

serum-free media interactive online database. ALTEX 27 (1):53-62.

Cabrera G, Cremata JA, Valdés R, García R, González Y, Montesino R, Gómez H, and González M (2005).

Influence of culture conditions on the N-glycosylation of a monoclonal antibody specific for

recombinant hepatitis B surface antigen. Biotechnology and Applied Biochemistry 41:67-76.

Carter PJ (2006). Potent antibody therapeutics by design. Nature Reviews Immunology 6 (5):343-357.

Chotigeat W, Watanapokasin Y, Mahler S, and Gray PP (1994). Role of environmental conditions on the

expression levels, glycoform pattern and levels of sialyltransferase for hFSH produced by

recombinant CHO cells. Cytotechnology 15 (1):217-221.


150

Costa AR, Rodrigues ME, Henriques M, Melton DW, Cunnah P, Oliveira R, and Azeredo J (2012).

Evaluation of the OSCAR™ system for the production of monoclonal antibodies by CHO-K1 cells.

International Journal of Pharmaceutics 430:42-46.

Crowell CK, Grampp GE, Rogers GN, Miller J, and Scheinman RI (2007). Amino acid and manganese

supplementation modulates the glycosylation state of erythropoietin in a CHO culture system.


Davies J, Jiang LY, Pan L-Z, LaBarre MU, Andersen D, and Reff M (2001). Expression of GnTIII in

recombinant anti-CD20 CHO production cell line: expression of antibodies with altered glycoforms

leads to an increase in ADCC through higher affinity for FcyRIII. Biotechnology & Bioengineering

74:288-294.

Doyle A, and Griffiths JB (1998). Cell and Tissue Culture: Laboratory Procedures in Biotechnology.

Chichester, England: John Wiley & Sons.

Elliott S, Lorenzini T, Asher S, Aoki K, Brankow D, Buck L, Busse L, Chang D, Fuller J, Grant J, Hernday

N, Hokum M, Hu S, Knudten A, Levin N, Komorowski R, Martin F, Navarro R, Osslund T, Rogers G,

Rogers N, Trail G, and Egrie J (2003). Enhancement of therapeutic protein in vivo activities through

glycoengineering. Nature Biotechnology 21 (4):414-421.

Fernandes D (2005). Demonstrating comparability of antibody glycosylation during biomanufacturing.

European Biopharmaceutical Review:106-110.

Ferrara C, Brünker P, Suter T, Moser S, Püntener U, and Umaña P (2006). Modulation of therapeutic

antibody effector functions by glycosylation engineering: Influence of Golgi enzyme localization

domain and co-expression of heterologous beta1, 4-N-acetylglucosaminyltransferase III and Golgi

alpha-mannosidase II. Biotechnology and Bioengineering 93 (5):851-861.

Flynn GC, Chen X, Liu YD, Shah B, and Zhang Z (2010). Naturally occurring glycan forms of human

immunoglobulins G1 and G2. Molecular Immunology 47 (11–12):2074-2082.

Geigert J (2004). The Challenge of CMC Regulatory Compliance for Biopharmaceuticals. New York:

Springer.

Goetze AM, Liu YD, Arroll T, Chu L, and Flynn GC (2012). Rates and impact of human antibody glycation

in vivo. Glycobiology 22 (2):221-234.



Gramer MJ, and Goochee CF (1994). Glycosidase activities of the 293 and NS0 cell lines, and of an

antibody-producing hybridoma cell line. Biotechnology and Bioengineering 43 (5):423-428.

Griffiths JB (1987). Serum and growth factors in cell culture media--an introductory review. Developments

in Biological Standardization 66:155-160.


151

Gu X, Xie L, Harmon BJ, and Wang DIC (1997). Influence of Primatone RL supplementation on sialylation

of recombinant human interferon-γ produced by Chinese hamster ovary cell culture using serum-free

media. Biotechnology and Bioengineering 56 (4):353-360.

Haase H, and Beyersmann D (1999). Uptake and intracellular distribution of labile and total Zn(II) in C6

rat glioma cells investigated with fluorescent probes and atomic absorption. BioMetals 12 (3):247-

254.

Hashimoto K, Goto S, Kawano S, Aoki-Kinoshita KF, Ueda N, Hamajima M, Kawasaki T, and Kanehisa M

(2006). KEGG as a glycome informatics resource. Glycobiology 16 (5):63R-70R.

Ho E, and Ames BN (2002). Low intracellular zinc induces oxidative DNA damage, disrupts p53, NFκB,

and AP1 DNA binding, and affects DNA repair in a rat glioma cell line. Proceedings of the National

Academy of Sciences 99 (26):16770-16775.



Hossler P, Khattak SF, and Li ZJ (2009). Optimal and consistent protein glycosylation in mammalian cell

culture. Glycobiology 19 (9):936-949.

Jacobia S, Kenerson R, Tescione L, Gruber D, Jayme D, Munroe D, and Gorfien S (2006). Trace element

optimization enhances performance and reproducibility of serum-free medium. In Animal Cell

Technology: Basic & Applied Aspects, edited by S Iijima and K-I Nishijima: Springer Netherlands.



Jefferis R, Lund J, Mizutani H, Nakagawa H, Kawazoe Y, Arata Y, and Takahashi N (1990). A comparative

study of the N-linked oligosaccharide structures of human IgG subclass proteins. The Biochemical

Journal 268 (3):529-537.

Jenkins N, Parekh RB, and James DC (1996). Getting the glycosylation right: implications for the

biotechnology industry. Nature Biotechnology 14 (8):975-981.

Jones D, Kroos N, Anema R, Van Montfort B, Vooys A, van der Kraats S, van der Helm E, Smits S,

Schouten J, Brouwer K, Lagerwerf F, Van Berkel P, Opstelten D-J, Logtenberg T, and Bout A (2003).

High-level expression of recombinant IgG in the human cell line PER.C6. Biotechnology Progress 19

(1):163-168.





Kanda Y, Yamane-Ohnuki N, Sakai N, Yamano K, Nakano R, Inoue M, Misaka H, Iida S, Wakitani M,

Konno Y, Yano K, Shitara K, Hosoi S, and Satoh M (2006). Comparison of cell lines for stable


152

production of fucose-negative antibodies with enhanced ADCC. Biotechnology and Bioengineering 94

(4):680-688.

Kaneko Y, Nimmerjahn F, and Ravetch JV (2006). Anti-inflammatory activity of immunoglobulin G resulting

from Fc sialylation. Science 313 (5787):670-673.

Kovář J (1989). Various cell lines grow in protein-free hybridoma medium. In Vitro Cellular &

Developmental Biology - Plant 25 (5):395-396.

Krapp S, Mimura Y, Jefferis R, Huber R, and Sondermann P (2003). Structural analysis of human IgG-Fc

glycoforms reveals a correlation between glycosylation and structural integrity. Journal of Molecular

Biology 325 (5):979-989.




(3-4):143-151.

Kunkel JP, Jan DC, Butler M, and Jamieson JC (2000). Comparisons of the glycosylation of a monoclonal

antibody produced under nominally identical cell culture conditions in two different bioreactors.


LeFloch F, Tessier B, Chenuet S, Guillaume J-M, Cans P, Goergen J-L, and Marc A (2006). Related effects

of cell adaptation to serum-free conditions on murine EPO production and glycosylation by CHO

cells. Cytotechnology 52 (1):39-53.

Lifely MR, Hale C, Boyce S, Keen MJ, and Phillips J (1995). Glycosylation and biological activity of

CAMPATH-1H expressed in different cell lines and grown under different culture conditions.

Glycobiology 5 (8):813-822.

Liu CH, and Chang T-Y (2006). Rational development of serum-free medium for Chinese hamster ovary

cells. Process Biochemistry 41 (11):2314-2319.

Lowe JB, and Marth JD (2003). A genetic approach to mammalian glycan function. Annual Review of


Lund J, Takahashi N, Nakagawa H, Goodall M, Bentley T, Hindley SA, Tyler R, and Jefferis R. 1993.

Control of IgG/Fc glycosylation: a comparison of oligosaccharides from chimeric human/mouse and

mouse subclass immunoglobulin Gs. Molecular Immunology 30 (8):741-748.

Magdelaine-Beuzelin C, Kaas Q, Wehbi V, Ohresser M, Jefferis R, Lefranc M-P, and Watier H (2007).

Structure–function relationships of the variable domains of monoclonal antibodies approved for

cancer treatment. Critical Reviews in Oncology/Hematology 64 (3):210-225.

Masuda K, Yamaguchi Y, Kato K, Takahashi N, Shimada I, and Arata Y (2000). Pairing of oligosaccharides

in the Fc region of immunoglobulin G. FEBS letters 473 (3):349-357.


153

Mimura Y, Ghirlando R, Sondermann P, Lund J, and Jefferis R (2001). The molecular specificity of IgG-Fc

interactions with Fc gamma receptors. Advances in Experimental Medicine and Biology 495:49-53.

Mimura Y, Church S, Ghirlando R, Ashton PR, Dong S, Goodall M, Lund J, and Jefferis R (2000). The

influence of glycosylation on the thermal stability and effector function expression of human IgG1-Fc:

properties of a series of truncated glycoforms. Molecular Immunology 37 (12–13):697-706.

Müthing J, Kemminer SE, Conradt HS, Sagi D, Nimtz M, Kärst U, and Peter-Katalinić J (2003). Effects of

buffering conditions and culture pH on production rates and glycosylation of clinical phase I anti-

melanoma mouse IgG3 monoclonal antibody R24. Biotechnology and Bioengineering 83 (3):321-

334.

Nimmerjahn F, and Ravetch JV (2008). Fcγ receptors as regulators of immune responses. Nature Reviews

Immunology 8 (1):34-47.

Ozturk S, Kasebo G, Mahaworasilpa T, and Coster HG (2003). Adaptation of cell lines to serum-free

culture medium. Hybrid Hybridomics 22 (4):267-272.



Ozturk SS, and Palsson BO (1991). Physiological changes during the adaptation of hybridoma cells to low

serum and serum-free media. Biotechnology and Bioengineering 37 (1):35-46.

Parekh RB, Dwek RA, Sutton BJ, Fernandes DL, Leung A, Stanworth D, Rademacher TD, Mizuochi T,

Taniguchi T, Matsuta K, Takeuchi F, Nagano Y, Miyamoto T, and Kobata A (1985). Association of

rheumatoid arthritis and primary osteoarthritis with changes in the glycosylation pattern of total

serum IgG. Nature 316 (6027):452-457.

Patel TP, Parekh RB, Moellering BJ, and Prior CP (1992). Different culture methods lead to differences in

glycosylation of a murine IgG monoclonal antibody. Biochemical Journal 285:839-845.

Primack J, Flynn GC, and Pan H (2011). A high-throughput microchip-based glycan screening assay for

antibody cell culture samples. Electrophoresis 32 (10):1129-1132.

Pučić M, Knežević A, Vidič J, Adamczyk B, Novokmet M, Polašek O, Gornik O, Šupraha-Goreta S, Wormald

MR, Redžić I, Campbell H, Wright A, Hastie ND, Wilson JF, Rudan I, Wuhrer M, Rudd PM, Josić D,

and Lauc G (2011). High throughput isolation and glycosylation analysis of IgG–variability and

heritability of the IgG glycome in three isolated human populations. Molecular & Cellular Proteomics

10 (10).

Raju TS (2003). Glycosylation variations with expression systems and their impact on biological activity of

therapeutic immunoglobulins. BioProcess International 1 (4):44-53.

Raju TS, Briggs JB, Borge SM, and Jones AJS (2000). Species-specific variation in glycosylation of IgG:

evidence for the species-specific sialylation and branch-specific galactosylation and importance for

engineering recombinant glycoprotein therapeutics. Glycobiology 10 (5):477-486.


154

Reichert JM, Rosensweig CJ, Faden LB, and Dewitz MC (2005). Monoclonal antibody successes in the

clinic. Nature Biotechnology 23 (9):1073-1078.

Reichert JM, and Valge-Archer VE (2007). Development trends for monoclonal antibody cancer

therapeutics. Nature Reviews Drug Discovery 6 (5):349-356.

Rodrigues ME, Costa AR, Henriques M, Azeredo J, and Oliveira R (2009). Technological progresses in

monoclonal antibody production systems. Biotechnology Progress 26 (2):332-351.

Routier F, Davies M, Bergemann K, and Hounsell E (1997). The glycosylation pattern of a humanized IgGl

antibody (D1.3) expressed in CHO cells. Glycoconjugate Journal 14 (2):201-207.

Routier FH, Hounsell EF, Rudd PM, Takahashi N, Bond A, Hay FC, Alavi A, Axford JS, and Jefferis R

(1998). Quantitation of the oligosaccharides of human serum IgG from patients with rheumatoid

arthritis: a critical evaluation of different methods. Journal of Immunological Methods 213 (2):113-

130.

Saba JA, Kunkel JP, Jan DCH, Ens WE, Standing KG, Butler M, Jamieson JC, and Perreault H (2002). A

Study of immunoglobulin G glycosylation in monoclonal and polyclonal species by electrospray and

matrix-assisted laser desorption/ionization mass spectrometry. Analytical Biochemistry 305 (1):16-

31.



(7):1524-1534.







47:113-124.

Sethuraman N, and Stadheim TA (2006). Challenges in therapeutic glycoprotein production. Current









155


278 (5):3466-3473.

Sibéril S, de Romeuf C, Bihoreau N, Fernandez N, Meterreau J-L, Regenman A, Nony E, Gaucher C,

Glacet A, Jorieux S, Klein P, Hogarth MP, Fridman W-H, Bourel D, Béliard R, and Teillaud J-L (2006).

Selection of a human anti-RhD monoclonal antibody for therapeutic use: Impact of IgG glycosylation

on activating and inhibitory FcγR functions. Clinical Immunology 118 (2–3):170-179.

Takahashi N, and Tomiya N (1992). Analysis of N-linked oligosaccharides: application of glycoamidase A.

In Handbook of Endoglycosidases and Glycoamidases edited by N Takahashi and T Muramatsu.

Boca Raton, FL: CRC Press.

Trummer E, Fauland K, Seidinger S, Schriebl K, Lattenmayer C, Kunert R, Vorauer-Uhl K, Weik R, Borth N,

Katinger H, and Müller D (2006). Process parameter shifting: Part II. Biphasic cultivation - A tool for

enhancing the volumetric productivity of batch processes using Epo-Fc expressing CHO cells.


Tsuchiya N, Endo T, Matsuta K, Yoshinoya S, Aikawa T, Kosuge E, Takeuchi F, Miyamoto T, and Kobata A

(1989). Effects of galactose depletion from oligosaccharide chains on immunological activities of

human IgG. The Journal of Rheumatology 16 (3):285-290.

Wong CFD, Wong TKK, Goh TL, Heng KC, and Yap GSM (2005). Impact of dynamic online fed-batch

strategies on metabolism, productivity and N-glycosylation quality in CHO cell cultures. Biotechnology


Yoon SK, Choi SL, Song JY, and Lee GM (2005). Effect of culture pH on erythropoietin production by

Chinese hamster ovary cells grown in suspension at 32.5 and 37.0°C. Biotechnology and


Yoon SK, Kim SH, and Lee GM (2003). Effect of low culture temperature on specific productivity and

transcription level of anti-4-1BB antibody in recombinant Chinese hamster ovary cells. Biotechnology

Progress 19 (4):1383-1386.

Zhang J, and Robinson D (2005). Development of animal-free, protein-free and chemically-defined media

for NS0 cell culture. Cytotechnology 48 (1):59-74.

Zhang J, Robinson D, and Salmon P (2006). A novel function for selenium in biological system: Selenite

as a highly effective iron carrier for Chinese hamster ovary cell growth and monoclonal antibody

production. Biotechnology and Bioengineering 95 (6):1188-1197.


Impact of microcarrier culture on product

quality

CC

THE PRESENT CHAPTER WAS ADAPTED FROM THE FOLLOWING RESEARCH PAPERS

ANA RITA COSTA, Joanne Withers, Maria Elisa Rodrigues, Niaobh McLaughlin,

Mariana Henriques, Rosário Oliveira, Pauline Rudd, Joana Azeredo (2012). The

impact of microcarrier culture optimization on the glycosylation profile of a

monoclonal antibody. Submitted to Journal of Industrial Microbiology and

Biotechnology.

Maria Elisa Rodrigues, ANA RITA COSTA, Pedro Fernandes, Mariana Henriques, Philip

Cunnah, David Melton, Joana Azeredo, Rosário Oliveira (2012). Evaluation of

macroporous and microporous carriers for CHO-K1 cell growth and monoclonal

antibody production. Submitted to Biotechnology Progress.

159

icrocarriers are widely used for the large-scale culture of attachment-dependent

cells, providing increased cell density and product yield. However, the specific culture conditions

used in these processes can affect the quality of the product, which is closely related to its

glycosylation pattern. The lack of studies in the area reinforces the need to better understand the

effects of these microcarrier cultures on product glycosylation. Consequently, this chapter centers

on the evaluation of the glycosylation profile of a monoclonal antibody (mAb) produced by

Chinese hamster ovary (CHO-K1) cells grown in Cytodex 3, comparing different culture conditions

(initial culture volume and cell concentration, rocking methodology and speed, and culture

vessel). The glycosylation profile of the microcarrier cultures is also compared to that obtained of

typical adherent cultures.

It was found that microcarriers resulted in a glycosylation profile with different

characteristics from T-flask cultures, with a general increase in galactosylation and decrease in

fucosylation levels, both with a potentially positive impact on mAb efficacy. The levels of

sialylation also diverged, but without a general tendency. This study then showed that the specific

culture conditions used in microcarrier culture influenced the mAb glycan profile, and that each

functional element (galactose, core fucose, sialic acid) was independently affected by these

conditions. In particular, great reductions of fucosylation were obtained with a lower initial volume

of culture, and notable decreases in sialylation and glycoform heterogeneity were observed for

shake flask culture, with these last modifications being associated with improved cell densities

achieved in these culture vessels.

Keywords: Monoclonal antibody; Chinese hamster ovary cells; Microcarrier; Cytodex 3;

Glycosylation

MM

CHAPTER 4 │ IMPACT OF MICROCARRIER CULTURE ON PRODUCT QUALITY

161

4.1. INTRODUCTION

The currently preferred system for the large-scale production of therapeutics is the

suspended culture of mammalian cells. These systems provide surface-to-volume ratios higher

that the common anchorage-dependent cultures and consequently result in improved cell

densities and product yields (Ozturk and Hu 2006). Despite this, there are still continuous efforts

to develop anchorage-dependent systems that can compete with the suspended culture, either

due to some limitations of suspended systems, the impossibility of growing certain cell types in

suspension, or the inevitable use of adherent cells to obtain specific products. Among the diverse

technologies developed to the effect, microcarrier systems are the most successful and well-

known (Chu and Robinson 2001; Ziao et al. 2002; Kong et al. 1999), with large application in the

production of viral vaccines (Berry et al. 1999; Wu and Huang 2002; Mendonça et al. 1993; Wu,

Huang, and Liu 2002; Butler et al. 2000) and recombinant proteins (Hu et al. 2000; Wang et al.

2002; Blüml 2007; Wang and Ouyang 1999; Cosgrove et al. 1995). These systems improve the

surface area available for cell adhesion and consequently result in improved cell density and

product yield (Hirtenstein et al. 1980; Blüml 2007; Van Wezel 1967; Schürch, Cryz, and Lang

1992; Rudolph et al. 2008).

Generally, microcarriers can be structurally divided into micro- and macroporous, according

to the size of the porous. In the microporous carriers (e.g. Cytodex) cells are able to grow only on

the external surface, while in macroporous carriers (e.g. CultiSpher, Cytoline, and Cytopore) cells

can also colonize the inner spaces (Ozturk and Hu 2006; Almgren et al. 1991; Spearman et al.

2005; Yokomizo et al. 2004). The characteristics of the microcarriers are important for the

success of the culture (Ozturk and Hu 2006; Butler 1996), but other factors must be considered,

such as the scale and type of culture vessel (Wu, Hsieh, and Liau 1998; Wang et al. 2002), and

the culture environment (Ng, Berry, and Butler 1996; Nam et al. 2008). Furthermore, it is

important to recognize that the specific conditions of the culture can influence not only cell and

product yields, but also the product quality, for example through effects on protein glycosylation.

Indeed, it has become evident that glycosylation plays critical roles in protein effectiveness in

vivo (Jenkins and Curling 1994), and the ability to perform this post-translational modification is a

major reason for the current choice of mammalian cells as hosts for recombinant therapeutic

production (Nam et al. 2008). It is therefore essential to ensure that glycosylation of recombinant


162

proteins in microcarrier culture is consistent with the desired product (Spearman et al. 2005).

However, to date, there are limited studies examining the effects of changing between adherent,

suspension, and microcarrier cultures on glycosylation of the recombinant product (Nam et al.

2008). Furthermore, the few studies performed typically compare microcarrier culture to

suspension culture (Watson et al. 1994; Hooker et al. 2007; Wang et al. 2002), and not to the

normal adherent cell culture conditions, and have shown results that seem to be specific to the

type of cell, product, and microcarrier used. For example, increased sialylation was found in

Cytodex culture for recombinant human tissue kallikrein production (Watson et al. 1994) and in

Cytoline culture of Chinese hamster ovary (CHO) cells in a fluidized bed bioreactor for interferon-γ

production (Hooker et al. 2007), while no differences were found in human recombinant

erythropoietin (EPO) produced by CHO cells in fluidized bed bioreactor cultures with Cytoline

(Wang et al. 2002).

In an effort to elucidate the effects of microcarrier culture on protein glycosylation, the

present work evaluated the glycan profile of a monoclonal antibody (mAb) produced by CHO-K1

cells grown in microcarriers. Different culture conditions were assessed to determine their

influence on mAb glycosylation, particularly on the galactosylation, fucosylation, and sialylation

levels, and the potential impact on the biological effectiveness of the mAb. Additionally, the

glycosylation profiles obtained in microcarrier cultures were compared to those of adherent

culture in common culture vessels (T-flasks).


163


4.2.1. CELL LINE AND CELL CULTURE

MAb-producing CHO-K1 cells, previously obtained (Costa et al. 2012) through the

transfection of a CHO-K1 cell line (CCL-61, American Type Culture Collection, Spain) for the

production of a recombinant human mAb (CAB051, Biotecnol SA, Portugal) was used. These

cells were grown in Dulbecco’s Modified Eagles Medium (DMEM, Sigma-Aldrich, United Kingdom)

supplemented with 10 % fetal bovine serum (FBS, Sigma-Aldrich) and 1x hypoxanthine

aminopterin thymidine (HAT, Sigma-Aldrich), at 37 ºC and 5 % carbon dioxide (CO2), in 75 cm2

culture flasks (T-flasks, Frilabo, Portugal). For microcarrier culture, cells were detached after

reaching a 70-80 % confluence and the concentration of the cell suspension adjusted to the

values needed for each assay (see section 4.2.3.).

4.2.2. MICROCARRIER PREPARATION

The microporous Cytodex 3 carriers (Sigma-Aldrich) were prepared according to the

manufacturer instructions. Briefly, the dry microcarriers were swollen and hydrated in calcium-

and magnesium-free phosphate buffered saline (Ca2+/Mg2+-free PBS, 137 mM sodium chloride

(NaCl), 2.7 mM potassium chloride (KCl), 10 mM sodium phosphate dibasic (Na2HPO4), and 2

mM potassium dihydrogen phosphate (KH2PO4), pH 7.4) (Sigma-Aldrich) for at least 3 h at room

temperature, sterilized (121 ºC, 15 min, 15 psi), and stored at 4 ºC. Prior to use, the sterilized

microcarriers were washed twice with culture medium and once with Ca2+/Mg2+-free PBS.

4.2.3. MICROCARRIER CULTURE

Microcarrier cultures were initially performed in 50 mL vented conical tubes (Inopat,

Portugal) and latter in 250 mL shake flasks (Sigma-Aldrich). In vented conical tubes, 5 mL


164

cultures of mAb-producing CHO-K1 cells were performed with 3 g/L of Cytodex 3, and different

conditions were assessed as described in Table 4.1.

TABLE 4.1. Conditions tested for the culture of the monoclonal antibody-producing CHO-K1 cells in Cytodex 3 (CY)

microcarriers

ASSAY INOCULUM CONCENTRATION

(cells/ml)

ROCKING MECHANISM ROCKING SPEED

(rpm)

VOLUME AT

INOCULATION

CULTURE VESSEL

CY1 2x105 Pulse followed by continuous 60 Total Vented conical tube

CY2 2x105 Continuous 60 Total Vented conical tube



CY5 2x105 Pulse followed by continuous 60 Half Vented conical tube



CYS 4x105 Pulse followed by continuous 60 Total Shake flask

All cultures were incubated at 37 ºC and 5 % CO2, with a fixed angle of inclination of 30º.

The rocking mechanism of ‘pulse followed by continuous’ refers to the use of a pulse rocking (60

or 40 rpm for 1 min at each 30 min) during the first six hours of culture (considered as the

critical phase for initial cell adhesion) followed by continuous rocking for the remaining time of

culture; the ‘continuous rocking’ mechanism uses a continuous rocking throughout the whole

culture. The assays using half the volume of medium at inoculation were completed with the

remaining volume after the first six hours of culture. For all assays, the medium was renewed

daily by replacing half of the culture volume with fresh growth medium.

In shake flasks, cultures of 20 mL were performed as described for the vented conical

tubes, with the exception of the 30 º inclination that has no application in these vessels.

4.2.4. CULTURE MONITORING

Cell growth was monitored in T-flask and microcarrier cultures. For this, cells were released

from the microcarriers by enzymatic digestion with trypsin (Sigma-Aldrich) and enumerated in a

haemocytometer with trypan blue staining (Sigma-Aldrich) for cell concentration and viability


165

assessment. In microcarrier cultures, the cell concentration in the sample was standardized with

the number of microcarriers counted microscopically in the sample (Nmicrocarriers in sample), and the

concentration of cells per microcarrier (Ccell/microcarrier) determined according to Equation 4.1.

Ccell/microcarrier=6(Ncell 4⁄ )×1047×F×VsampleNmicrocarriers in sample

EQUATION 4.1. Determination of the cell concentration per microcarrier (Ccell/microcarrier). Ncells - total number of cells counted

in the haemocytometer; F - sample dilution factor; Vsample - volume of sample in mL; and Nmicrocarriers in sample - total number of

microcarriers counted in the sample.

The cell concentration in the culture (Ccell/mL) was then determined according to Equation 4.2,

using the concentration of microcarriers (Cmicrocarriers/mL) obtained from Equation 4.3.

Ccell/mL= Ccell/microcarrier×Cmicrocarriers/mL

EQUATION 4.2. Determination of the cell concentration in the culture (Ccell/mL) in cells/mL. Ccell/microcarrier – cell concentration

per microcarrier; and Cmicrocarriers/mL- concentration of microcarriers in the culture in number of microcarriers per mL.

Cmicrocarriers/mL=Cmicrocarriers in g/mL×Nmicrocarriers g dry weight⁄

EQUATION 4.3. Determination of the concentration of microcarriers in the culture in number of microcarriers per mL

(Cmicrocarriers/mL); Cmicrocarriers in g/mL – concentration of microcarriers used in the culture in grams of dry weight per mL; and

Nmicrocarriers/g dry weight - approximate number of microcarriers per gram of dry weight (data provided by the manufacturer).

4.2.5. ANTIBODY QUANTIFICATION BY ENZYME-LINKED IMMUNOSORBENT ASSAY

Cell productivity was assessed by enzyme-linked immunosorbent assay (ELISA), following an

optimized procedure described in the confidential and proprietary standard operating procedure

(SOP) 2008-01 ANL from Biotecnol SA. Briefly, 96 well plates (CoStar, United States of America)

were coated with a capture antibody (Sigma-Aldrich) overnight. After a 45 min blocking at room

temperature with 1 % bovine serum albumin in phosphate buffered saline (BSA-PBS, 11.5 µM

BSA in 137mM NaCl, 2.68 mM KCl, 1.76 mM KH2PO4, and 12.72 mM Na2HPO4, pH 7.4) (Sigma-

Aldrich), sample dilutions in 0.2 % BSA-PBS (3.03 mM BSA in PBS), a standard of known

concentration and a quality control (Biotecnol SA), were added to the plates and incubated for 2


166

h at 37 ºC. The detection antibody was added and the plates incubated for additional 2 h at room

temperature. A 3,3’,5,5’-Tetramethylbenzidine (TMB, Sigma-Aldrich) substrate solution was

added and allowed to react for 10 min at room temperature. After stopping the reaction with

Phosphoric acid 75 % (Phosphoric acid 85 % in water, Frilabo), the absorbance was read at 450

nm. MAb concentration (CmAb) was determined using Equation 4.4 and the productivity (qmAb)

calculated according to Equation 4.5.

C� = �� − 1�

�� × c

EQUATION 4.4. Determination of monoclonal antibody concentration (CmAb) in µg/mL. Abs450 - absorbance read at 450

nm; a - estimated response at zero concentration; b - slope factor; c - mid-range concentration; and d - estimated

response at infinite concentration (parameters determined using the R software, version 2.6.2, from the R

Foundation for Statistical Computing).

q� (pg/cell/day)= �� Ccell/mL⁄t

× 10#

EQUATION 4.5. Determination of monoclonal antibody productivity (qmAb) in pg/cell/day. CmAb - mAb concentration in

µg/mL; Ccell/mL - cell concentration in cells/mL; and t - time of production in days.


The glycosylation profile of the mAb obtained from the T-flask and microcarrier cultures at

different conditions was assessed as described below. To note that in the following procedures,

all water used was of type 1 ultra-pure, resistivity of 18 MΩ cm, particle-free (> 0.22 µm), and

total organic content of less than ten parts per billion.

ANTIBODY PURIFICATION

Prior to immunoglobulin G (IgG) purification, samples were concentrated to volumes of 200

µL using 15 mL spin concentrators of 5 KDa molecular weight cutoff (Agilent Technologies,

Ireland). Then, the IgG from the concentrated samples was purified with Protein A spin plates

(Thermo Scientific, Ireland) according to manufacturer instructions, with slight modifications.


167

Briefly, the wells of a spin plate were equilibrated with two washes of 200 µL of 1x PBS (pH 7.2,

Merck). Samples were diluted two-fold in 1x PBS and added (200 µL) to the plate for a period of

incubation of 30 min with moderate agitation. The resin was washed three times with 500 µL of

PBS, and the purified IgG eluted three times with 300 µL of 0.5 M acetic acid (pH 2.5, AnalaR,

UK) and neutralized with 20 µL of 1 M ammonium bicarbonate (pH 7-8.5, Sigma-Aldrich). The

neutralized elutions were dried overnight in a vacuum centrifuge.











(Sigma-Aldrich) and 20 mM sodium bicarbonate (pH 7, Merck, UK), vortexed and mixed for 10

min after each wash, and dried in a vacuum centrifuge.


of 2 µL of Peptide N-glycosidase F (PNGase F, Prozyme, USA) and 48 µL of 20 mM sodium

bicarbonate (pH 7), and a further 12-16 h at 37 ºC after adding extra 50 µL 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 sonication for each

wash. The supernatants obtained were collected and dried in a vacuum centrifuge. The dried

samples were redissolved in 20 µL of formic acid (1 %, Sigma-Aldrich), incubated for 40 min at

room temperature and dried in a vacuum centrifuge.




168




tips (PhyNexus, USA). Briefly, samples were diluted in 95 µL water and 900 µL acetonitrile. The

PhyNexus tips were prepared by washing with ten 500 µL uptakes of 95 % acetonitrile (v/v in

water), ten 500 µL uptakes of 20 % acetonitrile (v/v in water), and another ten 500 µL uptakes of

95 % acetonitrile. Samples were loaded into the PhyNexus tips through ten 1 mL in-out cycles,

followed by washing of the tips with ten 1 mL uptakes of 95 % acetonitrile. The glycans were then

eluted with five uptakes of 200 µL of 20 % acetonitrile, and the elutions collected and dried in a

vacuum centrifuge.



was ressuspended in 20 µL water and 80 µL acetonitrile. NP-HPLC was performed using a
















169


IgG1 profile as a guide.

The relative percentage of N-linked glycans was calculated for agalactosylated (G0),

monogalactosylated (G1), digalactosylated (G2), trigalactosylated (G3), galactosylated

(G1+G2+G3), core fucosylated (F), core fucosylated agalactosylated (FG0), core fucosylated

monogalactosylated (FG1), core fucosylated digalactosylated (FG2), core fucosylated

trigalactosylated (FG3), monosialylated (S1), disialylated (S2), trisialylated (S3), sialylated

(S1+S2+S3), monosialylated monogalactosylated (S1G1), monosialylated digalactosylated

(S1G2), disialylated digalactosylated (S2G2), disialylated trigalactosylated (S2G3), trisialylated

trigalactosylated (S3G3), high mannose, and complex glycan structures. The relative percentages

were determined by summing the percentage of area of the peaks pertaining to each of the

abovementioned structures.


170


Microcarriers have emerged as one of the most promising technologies for the culture of

adherent cell types at scales adequate for biopharmaceutical production. To establish a

successful microcarrier culture, several parameters are typically manipulated to optimize the final

cell and product yields. However, these changes in the culture parameters may also affect the

quality of the product, for example, by inducing modifications on the glycosylation profile. Since

an appropriate glycosylation of therapeutic proteins such as mAbs is important for their biological

activity and clinical efficacy (Spearman et al. 2005), it is essential to ensure that glycosylation in

microcarrier cultures is consistent with the desired product. In this study, mAb-producing CHO-

K1 cells were cultured in microporous Cytodex 3 carriers, under different culture conditions to

evaluate their impact on the glycosylation profile of the mAb. Additionally, the glycosylation

obtained in microcarrier cultures was compared to that of normal adherent culture conditions in

T-flasks.

The glucose unit (GU) values and the tentative structure assignments of the peaks identified

by normal phase HPLC, as well as the relative peak area (%) obtained in each assay, are shown

in Table 4.2.

Glycans found in all conditions are mainly complex biantennary structures with a high

degree of heterogeneity, containing different terminal sugars, including sialic acid, galactose, N-

acetylglucosamine and core fucose. However, differences can be found between the microcarrier

cultures and the typical adherent culture in T-flask in the prevalence of certain glycans,

particularly of the most typical IgG1 structures: FA2, FA2G1 (6 and 3 arms), and FA2G2. Indeed,

FA2 and FA2G1 (6 arm) show a general tendency to decrease, while FA2G1 (3 arm) and FA2G2

increase in microcarrier cultures. The reduction of the prevalence of these main structures in

microcarrier culture is obviously accompanied by an increase in the remaining structures,

particularly of core fucosylated disialylated glycans such as FA2G2S2 and FA3G3S2, and of A1

and A1G1.


171

TABLE 4.2. Data of mean glucose unit (GU) value, tentative structure assignment, and relative (%) area of the peaks

of the monoclonal antibody produced by CHO-K1 cells under different culture conditions: T-flask; CYi – Cytodex 3 in

vented conical tubes; CYS – Cytodex 3 in shake flask

PEAK MEAN GU 1 TENTATIVE

ASSIGNMENT 2

% AREA

T-FLASK CY1 CY2 CY3 CY4 CY5 CY6 CY7 CYS

1 4.9 ± 0.01 A1 2.09 3.33 3.54 3.03 3.76 10.54 4.48 1.77 6.10

2 5.33 ± 0.01 FA1 / A2 2.01 1.75 1.59 2.54 3.03 3.40 1.56 1.74 3.32

3 5.66 ± 0.01 A1G1 - 5.69 5.76 5.15 5.31 14.16 5.70 - 8.60

4 5.80 ± 0.01 FA2 40.99 22.01 16.59 20.59 21.24 16.93 16.87 30.92 27.57

5 6.18 ± 0.01 M5 / A2G1 1.50 1.93 1.68 2.44 2.39 3.13 1.77 2.45 2.82

6 6.31 ± 0.04 A2G1[6] - 1.61 1.19 1.97 2.55 - 1.36 1.29 4.86

7 6.42 ± 0.01 A2G1[3] - 2.34 2.49 1.82 1.88 6.22 2.77 - -

8 6.64 ± 0.01 FA2G1[6] 23.52 20.74 15.60 16.12 17.10 13.20 15.62 25.59 23.46

9 6.75 ± 0.01 FA2G1[3] 9.11 15.18 10.53 11.45 12.98 7.33 11.77 13.23 9.85

10 7.15 ± 0.02 A2G2 / M6 1.65 2.59 2.15 1.94 2.37 3.25 2.72 1.81 2.59

11 7.59 ± 0.01 FA2G2 8.40 14.65 11.74 9.86 11.60 7.76 12.61 13.43 9.27

12 7.97 ± 0.01 FA2G1S1 2.73 2.16 2.18 1.91 1.83 2.86 2.54 1.72 1.56

13 8.46 ± 0.01 A2G2S1 0.91 1.29 2.12 2.03 1.58 2.44 1.98 1.09 -

14 8.70 ± 0.02 FA2G2S1[3] - - - - - 0.85 1.08 - -

15 8.88 ± 0.01 FA2G2S1[6] 0.98 1.26 2.38 2.19 1.71 0.78 1.68 1.06 -

16 9.17 ± 0.02 A2G2S2[3] - - - - - - - - -

17 9.36 A3G3S1 - - - - - - - - -

18 9.58 ± 0.01 A2G2S2[6] 1.99 0.73 3.70 2.31 1.88 1.95 2.64 0.94 -

19 10.03 ± 0.01 FA2G2S2 0.95 1.25 6.43 5.66 3.17 2.84 4.94 1.54 -

20 10.22 ± 0.00 A3G3S2 - - 1.18 1.11 0.54 - 1.04 - -

21 10.43 ± 0.01 FA3G3S2[3] 1.14 1.49 4.56 5.85 3.73 2.35 5.08 1.40 -

22 10.56 ± 0.01 FA3G3S2[6] - - 2.05 - - - - - -

23 10.68 FA3G3S2[3,6] 0.92 - - - - - - - -

24 10.93 ± 0.04 A3G3S3 0.84 - 2.54 2.01 1.34 - 1.79 - -

25 11.12 ± 0.01 A3G3S3[3] - - - - - - - - -

26 11.43 FA3G3S3 0.27 - - - - - - - -

Main structures 82.02 72.58 54.46 58.02 62.92 45.22 56.87 83.17 70.15

Other structures 17.98 27.42 45.54 41.96 37.07 54.77 43.13 16.81 29.85



The changes on the mAb glycosylation profile observed between the T-flask and the

microcarrier cultures may result from the creation of a different microenvironment for the cells in

the microcarriers (Spearman et al. 2005). This may impact glycosylation directly or through

variations caused on cell growth characteristics and mAb productivity, such as the ones observed


172

in this study (Table 4.3). Effectively, it is clear from Table 4.3 that microcarriers provide variable

cell concentrations but mAb productivities are always enhanced when compared to T-flasks.

Furthermore, it is known that CHO cells produce enzymes such as sialidase, beta-galactosidase,

and fucosidase that can accumulate extracellularly in culture and potentially lead to extracellular

modifications of the glycans (Gramer and Goochee 1993). By affecting cell characteristics,

microcarrier culture may influence the production of such enzymes and potentially interfere with

glycosylation.

TABLE 4.3. Data of average cell concentration and productivity of monoclonal antibody-producing CHO-K1 cells

cultured in different conditions: T-flask; CYi – Cytodex 3 in vented conical tubes; CYS – Cytodex 3 in shake flask

ASSAY AVERAGE CELL CONCENTRATION

(x105 cells/mL)

qmAb

(pg/cell/day)

T-flask 8.51 ± 0.06 0.21 ± 0.05

CY1 6.86 ± 1.06 2.57 ± 1.38

CY2 16.02 ± 1.98 0.84 ± 0.35

CY3 3.73 ± 0.63 3.89 ± 1.17

CY4 6.21 ± 0.62 2.16 ± 0.94

CY5 7.63 ± 0.76 1.75 ± 0.66

CY6 18.51 ± 1.08 1.26 ± 0.99

CY7 10.25 ± 0.73 1.51 ± 0.30

CYS 104.1 ± 6.00 1.86 ± 0.58

Furthermore, microcarrier cultures were performed under rocking conditions, which were

not present in the typical T-flask cultures, and which may also be a factor leading to the

differences found between the glycosylation profile of the mAb obtained in these two types of

cultures.

Several studies have focused on the effect of specific elements of the glycan profile on the

effector function of glycosylated proteins. Known as functional elements, these include galactose,

core fucose, and sialic acid, and have been shown to influence antibody-dependent cellular

cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC). In fact, higher levels of

galactosylation (Abès and Teillaud 2010; Serrato et al. 2007; Jefferis and Lund 1997;

Hodoniczky, Zheng, and James 2005), reduced core fucosylation (Sibéril et al. 2006; Shields et


173

al. 2002; Shinkawa et al. 2003) or reduced sialylation (Scallon et al. 2007; Kaneko, Nimmerjahn,

and Ravetch 2006; Anthony and Ravetch 2010; Anthony et al. 2008) are suggested to enhance

the clinical efficacy of mAbs that exert their therapeutic effect by ADCC and CDC mediated killing.

This is particularly relevant for core fucosylation, whose absence can improve ADCC activity by

up to 100-fold (Shields et al. 2002).

The relative percentages of the main functional elements of the mAb produced under the

different culture conditions evaluated in this study are shown in Table 4.4.

TABLE 4.4. Galactose, core fucose and sialic acid composition of the monoclonal antibody produced by CHO-K1 cells

during culture in different conditions: T-flask; CYi – Cytodex 3 in vented conical tubes; CYS – Cytodex 3 in shake

flask

GLYCAN STRUCTURES % AREA

T-FLASK CY1 CY2 CY3 CY4 CY5 CY6 CY7 CYS

Total number of peaks 17 17 20 19 19 17 20 15 11

Agalactosylated (G0) 46.67 29.35 23.64 28.35 30.41 34.06 25.16 36.56 39.70

Galactosylated (G1+G2+G3) 53.34 70.65 76.37 71.63 69.58 65.93 74.85 63.42 60.31

Monogalactosylated (G1) 36.11 48.69 38.59 39.64 42.85 45.34 40.65 43.06 49.74

Digalactosylated (G2) 14.06 20.48 27.45 23.02 21.13 18.25 26.29 18.97 10.57

Trigalactosylated (G3) 3.17 1.49 10.33 8.97 5.61 2.35 7.91 1.40 0.00

Core fucosylated (F) 89.01 78.74 72.06 73.63 73.36 54.90 72.19 88.89 71.71

Core fucosylated agalactosylated (FG0) 40.99 22.01 16.59 20.59 21.24 16.93 16.87 30.92 27.57

Core fucosylated monogalactosylated (FG1) 35.36 38.08 28.31 29.48 31.91 23.39 29.93 40.54 34.87

Core fucosylated digalactosylated (FG2) 10.33 17.16 20.55 17.71 16.48 12.23 20.31 16.03 9.27

Core fucosylated trigalactosylated (FG3) 2.33 1.49 6.61 5.85 3.73 2.35 5.08 1.40 0.00

Sialylated (S1+S2+S3) 10.73 8.18 27.14 23.07 15.78 14.07 22.77 7.75 1.56

Monosialylated (S1) 4.62 4.71 6.68 6.13 5.12 6.93 7.28 3.87 1.56

Disialylated (S2) 5.00 3.47 17.92 14.93 9.32 7.14 13.70 3.88 0.00

Trisialylated (S3) 1.11 0.00 2.54 2.01 1.34 0.00 1.79 0.00 0.00

Monosialylated monogalactosylated (S1G1) 2.73 2.16 2.18 1.91 1.83 2.86 2.54 1.72 1.56

Monosialylated digalactosylated (S1G2) 1.89 2.55 4.50 4.22 3.29 4.07 4.74 2.15 0.00

Monosialylated trigalactosylated (S1G3) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Disialylated digalactosylated (S2G2) 2.94 1.98 10.13 7.97 5.05 4.79 7.58 2.48 0.00

Disialylated trigalactosylated (S2G3) 2.06 1.49 7.79 6.96 4.27 2.35 6.12 1.40 0.00

Trisialylated trigalactosylated (S3G3) 1.11 0.00 2.54 2.01 1.34 0.00 1.79 0.00 0.00

Complex glycans 98.43 97.74 98.09 97.81 97.62 96.81 97.76 97.87 97.30

High mannose 1.58 2.26 1.92 2.19 2.38 3.19 2.25 2.13 2.71


174

In a first assessment it is possible to detect a general increase of galactosylation (mainly

digalacto (G2) and trigalacto (G3) structures) and decrease of fucosylation in microcarrier

cultures compared to T-flasks, both having a potential positive effect on mAb activity. On the

other hand, although also diverging from the T-flask cultures, the levels of sialylation in the

microcarrier assays are more variable, without showing a common tendency, which indicates a

possible higher susceptibility of this element to the culture conditions here evaluated. As said

above, it has been shown that CHO cells produce an extracellular sialidase capable of modifying

the sialic acid content of glycoproteins (Gramer and Goochee 1993), and that the

production/accumulation of this enzyme may be influenced by culture conditions, therefore

affecting the degree of sialylation. Additionally, it should be noted that the reported degree of

sialylation in human IgG has been below 20 % (Pučić et al. 2011; Arnold et al. 2007; Serrato et

al. 2007), but some of the microcarrier cultures resulted in superior levels (e.g. CY2), mainly due

to a higher prevalence of disialylated (S2) structures, both digalacto (S2G2) and trigalacto

(S2G3)), with potential negative effects on mAb activity by reducing ADCC.

Hypergalactosylation has been associated with low cell densities in previous studies (Kumpel

et al. 1994), but there was no correlation found in this work between the higher levels of mAb

galactosylation (Table 4.4) and the cell densities (Table 4.3) achieved in microcarrier cultures.

For core fucose, it is interesting to note that in addition to the global decrease of fucosylation,

there are also variations in the ratio of the agalacto:monogalacto:digalacto:trigalacto

(FG0:FG1:FG2:FG3) structures, with a general switch from the FG0 prevalence in T-flask cultures

to a FG1 prominence in microcarriers. In both cases, the divergences observed between

microcarrier and T-flask cultures may have been caused by the use of rocking in the former.

Furthermore, it should be noted that the potentially better levels (in terms of mAb activity) of

each of the functional elements are not obtained in the same conditions. For example, in terms

of galactosylation, mAb with improved characteristics (higher galactose) was produced in CY2;

while for fucosylation, the best quality mAb (less core fucose) was obtained in CY5; and for

sialylation, the lower and best levels were obtained in CY1, CY7, and CYS.

It has been suggested that increases in productivity, such as the ones observed in the

microcarrier cultures, may potentially lead to decreased effectiveness of recombinant proteins

(Nam et al. 2008). However, in this work, the higher productivities of microcarrier cultures (Table

4.3) do not seem to correlate with a lower quality of the mAb.


175

In a more specific analysis, the different conditions evaluated for the microcarrier cultures

were considered.

Rocking speed influenced the mAb glycosylation profile, but this effect was highly dependent

on the rocking methodology. For example, when using pulse/continuous rocking methodology,

60 rpm enabled the production of mAbs with lower levels of sialylation (reduced S2 structures)

compared to 40 rpm (CY1 versus CY3), which are potentially more effective through ADCC.

However, using a continuous rocking (CY2 versus CY4), these observations are reversed, and 40

rpm (CY4) appears as the advantageous rocking speed. There seems to be no correlation

between these results and the cell concentrations and productivities achieved in these assays

(Table 4.3). Nevertheless, the combination of rocking speed and mechanism should be

considered for mAb quality.

It is mentioned in the literature that the success of the microcarrier culture is dependent on

the initial colonization of the microcarriers with the cells (Butler 1996; Voigt and Zintl 1999;

Blüml 2007), so the culture conditions during the first hours of culture should be optimized to

improve this initial cell adhesion. In this work, two parameters that have been related to the

efficiency of initial cell adhesion were assessed, namely the rocking methodology used for the

first six hours of culture and the volume of the culture at inoculation.

Regarding the rocking methodology, it was found that this culture parameter affected the

glycosylation profile, but diverged with the concentration of cell inoculum used. Indeed, when

using 2x105 cells/mL (CY1 versus CY2), continuous rocking (CY2) resulted in increased levels of

sialylation, mainly due to a higher presence of S2 structures, which potentially results in a less

efficient mAb due to a decreased ADCC activity. However, when using a cell inoculum of 4x105

cells/mL (CY6 versus CY7), the continuous rocking (CY7) led to a mAb with lower galactosylation

and sialylation, and higher fucosylation, whose biological impact is difficult to predict due to the

opposite effects of these functional elements, although a decline of mAb effectiveness would be

expected due to the greater impact of fucosylation on mAb functionality (Shields et al. 2002).

Nevertheless, the combination of rocking methodology at the beginning of the culture and

concentration of cell inoculum seems to be important for mAb quality.

Concerning the culture volume, the use of total (5 mL) or half (2.5 mL) volume was tested

(CY1 versus CY5), with half (CY5) leading to a higher sialylation (mainly due to increased S2) and

to a great reduction of the levels of fucosylation. Due to the known strong impact of fucosylation

on IgG activity through ADCC (Shields et al. 2002), it is expected that the positive effects


176

associated with its reduction in these conditions will surpass the negative effects of an increased

sialylation, and result in a mAb with improved activity. The use of different volumes at inoculation

may affect the microenvironment of the cells and therefore influence glycosylation. Furthermore,

although it has been suggested that the use of half volume during the initial hours of culture may

improve cell adhesion and consequently the final cell and product yields, such effect was not

observed in this study.

In addition to the functional elements already discussed, 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 related to reduced ADCC and CDC (Kanda et al.

2007), as well as to a rapid IgG clearance from serum (Abès and Teillaud 2010). In this study,

the levels of high mannose are, in general, slightly increased in microcarrier cultures (particularly

for CY5).

The influence of the culture vessel on the glycosylation profile of the mAb produced by CHO-

K1 cells in microcarriers was also considered in this work, by comparing vented conical tubes

with shake flasks. The culture vessel was highly influential particularly in the sialylation levels of

the mAb, which greatly decreased to practically absent sialylation in shake flasks, predicting a

positive outcome on the biological activity of the mAb. Additionally, shake flasks seem to provide

a less heterogeneity of glycoforms, as demonstrated by the lowest number of peaks obtained (11

peaks, Table 4.4). Since cell concentrations are highly improved in shake flasks when compared

to the vented conical tube cultures, it is possible that these improved levels, or the factors that

led to them (e.g. better oxygenation and mass transfer), have caused these modifications on

sialylation and heterogeneity.

Some additional cultures were performed with a different type of microcarrier, specifically

the macroporous CultiSpher-S, for effects of comparison. The global influence of CultiSpher-S

cultures on glycosylation, in comparison with T-flask cultures, was similar to that previously

observed with Cytodex 3, causing a general reduction on the percentage of main structures, an

increase of galactosylation, a slightly reduction of fucosylation, and a variable effect on sialylation.

Furthermore, two culture variables tested with CultiSpher-S (initial culture volume and culture

vessel) also demonstrated effects on glycosylation similar to those of Cytodex 3 cultures. Of note

is the strong impact of the culture vessel on the glycosylation profile obtained with both

microcarrier cultures, with shake flasks confirming to be the best option by decreasing the

heterogeneity of glycoforms and leading to remarkable decreases of mAb sialylation with


177

CultiSpher-S. The cell densities achieved in these microcarriers were also significantly improved

in shake flasks, reinforcing the possibility of this being the cause for the positive changes

observed in the glycosylation profile of the mAb.


178

4.4. CONCLUSION

An understanding of the role of culture conditions on glycosylation is becoming increasingly

important to ensure a consistent quality of the product. In this study, the effect of microcarrier

culture and different culture conditions on the glycosylation pattern of a mAb produced by CHO-

K1 cells was evaluated. Higher levels of galactosylation (main di- and trigalacto structures) and

lower degrees of fucosylation were found in mAb produced by cells grown in microcarriers when

compared to the typical T-flask culture, both potentially improving the mAb effectiveness. On the

other hand, sialylation was found to be highly variable without a specific tendency. Furthermore,

contrarily to what has been suggested, the increased productivities obtained in microcarrier

cultures did not correlate with a lower quality of the mAb. On the other hand, the fact that

microcarrier cultures are performed under rocking conditions, in opposition to T-flask cultures,

may be a cause for the divergences found between them in the glycosylation profile of the mAb

produced.

All the culture conditions assessed in microcarrier culture led to modifications on the mAb

glycan profile, with each of the functional elements (galactose, core fucose, sialic acid) being

divergently affected by these conditions. In particular, the choice of the culture vessel appears to

have a strong influence in mAb sialylation, with great and desirable reductions achieved in shake

flask cultures. This is potentially related to highly improved cell densities achieved in these

vessels, which may have also been the cause for a reduction of the glycoform heterogeneity of

the mAb.


179

4.5. REFERENCES

Abès R, and Teillaud J-L (2010). Impact of glycosylation on effector functions of therapeutic IgG.


Almgren J, Nilsson C, Peterson AC, and Nilsson K (1991). Cultispher-macroporous gelatine microcarrier-

new applications. In Production of biologicals from animal cells in culture, edited by R. E. Spier, J. B.

Griffiths and B. Meignier. Oxford: Butterworth-Heinemann.

Anthony RM, Nimmerjahn F, Ashline DJ, Reinhold VN, Paulson JC, and Ravetch JV (2008). Recapitulation

of IVIG anti-inflammatory activity with a recombinant IgG Fc. Science 320 (5874):373-376.

Anthony R, and Ravetch J (2010). A novel role for the IgG Fc glycan: the anti-inflammatory activity of

sialylated IgG Fcs. Journal of Clinical Immunology 30 (0):9-14.

Arnold JN, Wormald MR, Sim RB, Rudd PM, and Dwek RA (2007). The impact of glycosylation on the

biological function and structure of human immunoglobulins. Annual Review of Immunology 25

(1):21-50.

Berry JM, Barnabé N, Coombs KM, and Butler M (1999). Production of reovirus type-1 and type-3 from

Vero cells grown on solid and macroporous microcarriers. Biotechnology and Bioengineering 62

(1):12-19.

Blüml G (2007). Microcarrier cell culture technology. In Animal Cell Biotechnology: Methods and

Protocols, Second Edition, edited by R Pörtner. Totowa, NJ: Humana Press Inc.

Butler M, Burgener A, Patrick M, Berry M, Moffatt D, Huzel N, Barnabé N, and Coombs K (2000).

Application of a serum-free medium for the growth of Vero cells and the production of reovirus.


Butler M (1996). Chapter 10 - Modes of culture for high cell densities. In Animal cell culture and

technology, edited by M Butler. Routledge, UK: Taylor and Francis.



Cosgrove L, Lovrecz GO, Verkuylen A, Cavaleri L, Black LA, Bentley JD, Howlett GJ, Gray PP, Ward CW,

and McKern NM (1995). Purification and properties of insulin receptor ectodomain from large-scale

mammalian cell culture. Protein Expression and Purification 6 (6):789-798.

Costa AR, Rodrigues ME, Henriques M, Melton D, Cunnah P, Oliveira R, and Azeredo J (2012). Evaluation

of the OSCAR™ system for the production of monoclonal antibodies by CHO-K1 cells. International

Journal of Pharmaceutics 430:42-46.


180



Hirtenstein M, Clark J, Lindgren G, and Vretblad P (1980). Microcarriers for animal cell culture: a brief

review of theory and practice. Developments in Biological Standardization 46:109-116.



Hooker AD, Goldman MH, Markham NH, James DC, Ison AP, Bull AT, Strange PG, Salmon I, Baines AJ,

and Jenkins N (2007). Monitoring recombinant human interferon-gamma N-glycosylation during

perfused fluidized-bed and stirred-tank batch culture of CHO cells. Biotechnology and Bioengineering

48 (6):596-607.

Hu X, Xiao C, Huang Z, Guo Z, Zhang Z, and Li Z (2000). Pilot production of u-PA with porous microcarrier

cell culture. Cytotechnology 33 (1-3):13-19.



Jenkins N, and Curling EM (1994). Glycosylation of recombinant proteins: problems and prospects.

Enzyme and Microbial Technology 16 (5):354-364.





Kaneko Y, Nimmerjahn F, and Ravetch JV (2006). Anti-inflammatory activity of immunoglobulin G resulting

from Fc sialylation. Science 313 (5787):670-673.

Kong D, Chen M, Gentz R, and Zhang J (1999). Cell growth and protein formation on various

microcarriers. Cytotechnology 29 (2):151-158.




(3-4):143-151.

Mendonça RZ, Ioshimoto LM, Mendonça RM, De-Franco M, Valentini EJ, Beçak W, Raw I, and Pereira CA

(1993). Preparation of human rabies vaccine in VERO cell culture using a microcarrier system.

Brazilian Journal of Medical and Biological Research 26 (12):1305-1317.

Nam JH, Zhang F, Ermonval M, Linhardt RJ, and Sharfstein ST (2008). The effects of culture conditions

on the glycosylation of secreted human placental alkaline phosphatase produced in Chinese hamster

ovary cells. Biotechnology and Bioengineering 100 (6):1178-1192.


181

Ng YC, Berry JM, and Butler M (1996). Optimization of physical parameters for cell attachment and

growth on macroporous microcarriers. Biotechnology and Bioengineering 50 (6):627-635.



Pučić M, Knežević A, Vidič J, Adamczyk B, Novokmet M, Polašek O, Gornik O, Šupraha-Goreta S, Wormald

MR, Redžić I, Campbell H, Wright A, Hastie ND, Wilson JF, Rudan I, Wuhrer M, Rudd PM, Josić D,

and Lauc G (2011). High throughput isolation and glycosylation analysis of IgG–variability and

heritability of the IgG glycome in three isolated human populations. Molecular & Cellular Proteomics

10 (10).

Rudolph G, Lindner P, Gierse A, Bluma A, Martinez G, Hitzmann B, and Scheper T (2008). Online

monitoring of microcarrier based fibroblast cultivations with in situ microscopy. Biotechnology and




(7):1524-1534.

Schürch U, Cryz SJ, and Lang AB (1992). Scale-up and optimization of culture conditions of a human

heterohybridoma producing serotype-specific antibodies to Pseudomonas aeruginosa. Applied

Microbiology and Biotechnology 37 (4):446-450.




47:113-124.








278 (5):3466-3473.

Sibéril S, de Romeuf C, Bihoreau N, Fernandez N, Meterreau J-L, Regenman A, Nony E, Gaucher C,

Glacet A, Jorieux S, Klein P, Hogarth MP, Fridman W-H, Bourel D, Béliard R, and Teillaud J-L (2006).

Selection of a human anti-RhD monoclonal antibody for therapeutic use: impact of IgG glycosylation

on activating and inhibitory FcγR functions. Clinical Immunology 118 (2–3):170-179.


182

Spearman M, Rodriguez J, Huzel N, and Butler M (2005). Production and glycosylation of recombinant β-

interferon in suspension and cytopore microcarrier cultures of CHO cells. Biotechnology Progress 21

(1):31-39.

Van Wezel AL (1967). Growth of cell-strains and primary cells on micro-carriers in homogeneous culture.

Nature 216 (5110):64-65.

Voigt A, and Zintl F (1999). Hybridoma cell growth and anti-neuroblastoma monoclonal antibody

production in spinner flasks using a protein-free medium with microcarriers. Journal of Biotechnology

68 (2-3):213-226

Wang MD, Yang M, Huzel N, and Butler M (2002). Erythropoietin production from CHO cells grown by

continuous culture in a fluidized-bed bioreactor. Biotechnology and Bioengineering 77 (2):194-203.

Wang Y, and Ouyang F (1999). Bead-to-bead transfer of Vero cells in microcarrier culture. Bioprocess and

Biosystems Engineering 21 (3):211-213.

Watson E, Shah B, Leiderman L, Hsu Y-R, Karkare S, Lu HS, and Lin F-K (1994). Comparison of N-linked

oligosaccharides of recombinant human tissue kallikrein produced by Chinese hamster ovary cells

on microcarrier beads and in serum-free suspension culture. Biotechnology Progress 10 (1):39-44.

Wu S-C, Hsieh W-C, and Liau M-Y (1998). Comparisons of microcarrier cell culture processes in one

hundred mini-liter spinner flask and fifteen-liter bioreactor cultures. Bioprocess Engineering 19

(6):431-434.

Wu S-C, and Huang GY-L (2002). Stationary and microcarrier cell culture processes for propagating

japanese encephalitis virus. Biotechnology Progress 18 (1):124-128.

Wu S-C, Huang GY-L, and Liu J-H (2002). Production of retrovirus and adenovirus vectors for gene

therapy: a comparative study using microcarrier and stationary cell culture. Biotechnology Progress

18 (3):617-622.

Yokomizo AY, Antoniazzi MM, Galdino PL, Azambuja N, Jorge SAC, and Pereira PA (2004). Rabies virus

production in high vero cell density cultures on macroporous microcarriers. Biotechnology and


Ziao C, Huang Z, Li W, Hu X, Qu W, Gao L, and Liu G (2002). High density and scale-up cultivation of

recombinant CHO cell line and hybridomas with porous microcarrier cytopore. In New Developments

and New Applications in Animal Cell Technology, edited by O-W Merten, P Perrin and B Griffiths:

Lavoisier.


General conclusions and considerations

for future work

CC

185

he work compiled in the present thesis is a contribute to the path into

biopharmaceutical processes with improved control of the quality of the product. Specifically, it

provides novel information about the influence of different processes and parameters involved in

the development stages of production on the glycosylation profile of the product.

The present chapter closes the work by providing a summary of the main conclusions of the

work performed, and proposing ideas to be addressed in future works in the field.

TT

CHAPTER 5 │ GENERAL CONCLUSIONS AND CONSIDERATIONS FOR FUTURE WORK

187

5.1. GENERAL CONCLUSIONS

In its still relatively short existence, biopharmaceutical industry has already revolutionized

modern medicine by applying recombinant technology in therapeutic protein production for the

treatment of several conditions that were previously thought incurable. Simultaneously, the industry

has faced many technological challenges in the pursuit of improved high-yielding processes capable

of fulfilling the market needs, while controlling and reducing the costs of said processes. One of the

most recent and important challenges is related to the control and maintenance of the product’s

quality during process development and at the final stages of production. The quality of the product

is majorly related to its glycosylation profile, a characteristic that has a strong impact on several

protein properties, such as serum half-life, in vivo activity, and immunogenicity. Unfortunately,

glycosylation is highly variable and can be affected by many parameters of process development.

So, to ensure the production of a consistent high-quality product it is essential to gain a deeper

understanding about the influence of these parameters on glycosylation. Such knowledge will

ultimately allow not only the preservation of the efficacy and safety characteristics of the product

during production, but also the manipulation of the culture parameters towards the enhancement of

these characteristics. The work presented in this thesis intended to be a contribute to this purpose,

by evaluating the glycosylation profile of a monoclonal antibody (mAb) produced by Chinese

hamster ovary (CHO-K1) cells during different stages of process development.

To complete this major goal, a mAb-producing CHO-K1 cell line was initially established using

the novel OSCARTM expression system as an alternative to traditional methods that typically require

long development times. OSCARTM showed several advantages over the most traditional systems:

first, transfection was simple and considerably faster, taking about two-three months compared to

the minimum six required by the common methods; second, the growth characteristics of the CHO-

K1 cells were maintained after transfection, as opposed to the common growth impair caused by

this type of processes; and third, the levels of mAb productivity initially obtained were high (10

pg/cell/day), without the need to perform the traditional process of amplification. This productivity

showed an abrupt decay in the first weeks in culture (1 pg/cell/day), but remained stable thereafter

for at least two years, indicating that if the initial decrease is controlled, the OSCARTM system may

provide producer cell lines with very stable expressions over time without the need to use the costly

toxic compounds required by the traditional technologies. Additionally, this work demonstrated that


188

a critical step for the successful transfection using the OSCARTM system is the choice of the most

adequate minigene. In fact, only one of the three minigenes assayed, pDMW128, worked with the

CHO-K1 cells tested, showing that the choice of minigene is cell line and/or product type-specific,

and its selection represents an additional workload of this methodology. In this context, it would be

useful to create a database to collect all data of studies performed with the OSCARTM expression

system involving different cells and products, so that in the future the selection of the best minigene

for each application will require less trial-and-error experimentation.

During the development of the mAb-producing CHO-K1 cell line, several producing cell clones

were obtained, which had to be subjected to a selection of the highest and most stable producers.

The methodology used for the first selection proved to be highly influent on the outcome of the

process. Although the typically used method of comparison of absorbance values of the culture

supernatants is the simplest and fastest approach, which is advantageous considering the high

number of clones that usually need to be screened, it does not certify that the highest cell

producers are selected. Indeed, the use of a more labor-intensive but accurate approach based on

the analysis of productivity clearly demonstrated that there is no guarantee that the highest-

producing clones are selected when using the first methodology. This has a huge impact on the

product yields that will be achievable in the final process of production, so this study recommends

the assessment of productivity as the approach of choice for the initial selection. Still in this stage of

cell-line development, it was found that the highest-producer cell clones express mAbs with similar

glycosylation profiles.

The glycosylation profile was also assessed during the adaptation of mAb-producing CHO-K1

cells to serum-free medium, a process currently required by regulatory agencies for the production

of therapeutics to avoid safety problems related to the presence of animal-derived components.

Since serum provides many essential elements for cell proliferation and viability, its removal from

the culture is typically challenging and time-consuming, as shown in the present study. The work

performed also highlights the importance of selecting adequate media supplements to support cells

in this process of adaptation, and of avoiding aggressive procedures like centrifugation and the use

of enzymes, as well as the need to give the time needed for cells to reacquire a healthy growth at

each stage. Furthermore, this work clearly demonstrated that the process of adaptation strongly

influences mAb glycosylation. During the different steps involved in this process, the number of

glycoforms (heterogeneity) and the levels of certain functional elements of glycosylation, specifically

galactose, core fucose, and sialic acids, varied considerably, being possible to separate the effects


189

into two stages. At the middle steps of adaptation, an increase of galactosylation and a decrease of

fucosylation were observed, potentially having positive outcomes on mAb efficacy by improving the

antibody-dependent cellular cytotoxicity (ADCC) and the complement-dependent cytotoxicity (CDC).

Sialylation, on the other hand, increased to higher levels than those normally reported, potentially

reducing mAb efficacy. A higher level of heterogeneity of glycoforms was also found for the middle

stages of adaptation, going against the intended homogeneity of glycosylation that improves efficacy

by reducing the presence of less efficient glycoforms. These results were related to the lower cell

concentrations and cell viabilities observed in this phase. Indeed, at the final steps of the process of

adaptation, where cells have reacquired a cell growth similar to the initial values, the glycosylation

profile became less heterogeneous and the levels of core fucose became similar to those initially

observed. However, galactosylation and sialylation of the serum-free cultures were decreased in

comparison to the initial serum-supplemented conditions, having divergent potential impacts on the

biological activity of the mAb (negative and positive, respectively). A graphical summary of the main

conclusions on this subject can be seen on Figure 5.1.

FIGURE 5.1. Summary of the modifications, and potential biological effects, detected in the glycosylation profile of the

monoclonal antibody produced by CHO-K1 cells during adaptation to growth in serum-free conditions. The

divergences found in the glycosylation profile were divided into two stages: during adaptation, and at the final stages

of adaptation and thereafter. Probable causes for the differences found are mentioned on the right.

Higher galactosylation

Lower fucosylation

Higher sialylation

Higher heterogeneity

Similar highmannose

Improved CDC

Improved ADCC

Reduced ADCC

Reduced efficacy due to the presence of

unefficient glycoforms

CELLS GROWING IN SERUM-SUPPLEMENTED MEDIUM

CELLS DURING ADAPTATION TO SERUM-FREE MEDIUM

Lower sialylation

Lower galactosylation

Similar fucosylation

Similar heterogeneity

Similar highmannose

Improved ADCC

Reduced CDC

CELLS AFTER ADAPTATION TO SERUM-FREE MEDIUM

+

-

+

-

Cell density and viability decrease

Growth and production rate

decrease

Cells shift from adherent to

suspended growth

Recovery of initial cell growth rate

Cells grow in suspension

Recovery of initial cell density and viability


190

The work reported in this thesis has also shown that glycosylation is influenced by cell

culture in microcarriers. Microcarriers are a successful, and increasingly used, alternative

technology for the culture of anchorage-dependent cells with high cell and product yields. In this

study, the glycosylation of the mAb produced by CHO-K1 cells growing adherently in Cytodex 3

microcarriers was different from the profile obtained in normal T-flask cultures. The changes

observed were attributed to divergences between the two types of culture in terms of mAb

productivity (higher in microcarriers), to the use of rocking in the microcarrier cultures, and to the

possible generation of a specific microenvironment in the microcarriers that may affect pH

and/or potentiate the accumulation of extracellular enzymes (e.g. sialidases, fucosidases,

galactosidases) that alter the glycan content of the mAb. To specify the differences found, the

mAb produced in microcarrier cultures had an increase on galactosylation and a decrease of

fucosylation when compared to the initial T-flask culture, with both modifications potentially

leading to a more efficient mAb. Sialylation, for its turn, was found to be highly variable without a

general tendency, demonstrating to be more sensitive to specific variations in the culture

conditions. Indeed, this study assessed the impact of several parameters commonly included in

the optimization of microcarrier culture, such as cell concentration at inoculum, initial culture

volume, rocking mechanism and rocking speed, as well as culture vessel. All these culture

parameters led to modifications on the mAb glycosylation profile, affecting the functional

elements of galactose, core fucose, and sialic acid divergently. In particular, fucosylation

appeared to be more affected by the volume of the culture at inoculation; galactosylation by the

combination of rocking mechanism and speed; and sialylation by the volume of the culture at

inoculation, the combination of rocking mechanism, speed and cell concentration, and especially

by the culture vessel. Indeed, it was found that the choice of the culture vessel is highly influent

on the sialylation of the mAb, with shake flasks resulting on its reduction to almost absent levels,

which potentially improves mAb activity. Furthermore, the use of shake flasks also led to a more

desired homogeneous glycosylation. These positive changes were related to the highly improved

cell densities obtained in these vessels, which may have resulted from better oxygenation and

mass transfer. A graphical summary of the main conclusions regarding glycosylation in

microcarrier culture can be seen in Figure 5.2.


191

FIGURE 5.2. Summary of the modifications, and potential biological effects, detected in the glycosylation profile of a

monoclonal antibody produced by CHO-K1 cells during microcarrier culture. Probable causes for the differences

found in comparison to the normal adherent culture in T-flasks are mentioned on the left.

In summary, the information resulting from this thesis contributes to a better understanding

of the effects of different technologies on product quality (glycosylation) in biopharmaceutical

processes, and highlights the need to extend quality monitoring and control to the early stages of

process development and not just limited to the final production process. The results obtained

open the doors to other investigations, and will hopefully contribute to a better control of product

quality in the near future, and perhaps to the improvement of the therapeutic efficacy of the

product through the simple manipulation of the culture conditions. A higher-quality product will

ultimately result in increased therapeutic efficacy, which will reduce the dose and frequency of

administration required, as well as the side effects, and consequently improve the patient

compliance with the treatment.

MICROCARRIER CULTURE

IN GENERAL SPECIFIC CULTURE CONDITIONS

T-FLASK CULTURE

Higher galactosylation

Lower fucosylation

Improved CDC

Improved ADCC+

Variable sialylation

Using half volume at inoculation

Using continuous rocking of 60 rpm

Using continuous rocking of 60 rpm

Using half volume at inoculation

Using shake flasks (it also reducesheterogeneity) – associated with highlyimproved cell densities

Using continuous rocking of 40 rpm,

cell inoculum of 4x105 cells/mL

Using pulse followed by continuous rocking

of 60 rpm, cell inoculum of 2x105 cells/mL

Reduced ADCC

Improved ADCC

Variable cell density

Increase ofprodutivity

Microenvironment

onmicrocarriers -

pH, extracellular

accumulation of

enzymes such as

sialidases,

galactosidases, and

fucosidases

Rockingparticularly

particularly


192

5.2. CONSIDERATIONS FOR FUTURE WORK

Although the present work is a step forward into the understanding of how glycosylation

varies with culture parameters, it is clear that this field still has many challenges to tackle.

First, it is important to understand that glycosylation is no longer seen merely as a form to

evaluate the quality of the product, but it is now regarded as a means to improve the therapeutic

function of the product. To this purpose, it is essential to develop a better understanding on the

relationship between glycans and the physiological functions of the protein. This should consider

three aspects: (i) the different components of the glycosylation pathway should not be addressed

individually, but instead should be analyzed for the results of their interplay; (ii) the optimization

of a glycan structure for one activity may compromise another property equally important, so the

global effect should be evaluated; and (iii) the optimal glycoform profile must be addressed in a

case-by-case basis, considering that the biological functions to optimize depend on the type of

product and therapeutic application intended. The structure-function relationships can only be

assessed by direct experimentation, creating another challenge related to the development of

animal models more adequate than those currently employed, since the effects observed in these

do not correctly simulate what happens in humans.

Second, knowing the optimal glycoform profile to target for during process development and

production, the control of culture conditions in order to meet this profile would present a simple

and economical way to obtain a more effective product. This goal will require more fundamental

studies on the dependence of glycosylation on different process conditions, such as those

performed in the work described herein, but the wide range of culture conditions imply that major

works will be necessary to reach the level of knowledge required for such control. Some of the

conditions to assess include: temperature, pH, culture media, culture mode (adherent,

suspension, microcarriers), type of bioreactor, mode of operation (batch, fed-batch, continuous),

rocking, as well as transfection, scale-up and adaptation processes.

Third, cells have a major impact on the glycosylation profile of the product, as proven by the

current reliance on mammalian cells for therapeutic protein production for their capacity to

produce glycoproteins similar to those of human. The differences on glycosylation among cells

are related to their specific sets of gene/protein expression levels, and availability and spatial

localization of glycoenzymes and nucleotide-sugar precursors. Consequently, it would be


193

important to identify the critical cell characteristics for glycosylation and to have the capacity to

implement their evaluation during early cell clone selection, in addition to the already assessed

properties of cell growth and productivity, so that the best clone for both product yields and

quality can be chosen.

Fourth, the understanding and control of glycosylation will, to a great extent, depend on

progresses made in the analytical methodologies. Currently, a complete structural analysis of

glycosylation can only be accomplished by combining different methodologies, therefore being

time-consuming and complex. So, future developments on the area will have to focus on the

following: (i) shortening total analysis time, to enable the use of the analytical data to regulate

process development or production, which will require the simplification or elimination of the

current protocols for sample preparation/purification; (ii) capacity to analyze a large number of

samples; (iii) improve sensitivities to allow the analysis of low-abundance samples; (iv) simplify

and automate the analysis to make the methodology more robust and accessible to non-

specialist laboratories with minimal user expertise; and (v) allow the accurate and complete

structural analysis of complex glycans and heterogeneous samples.

Fifth, the engineering of glycosylation offers great potential for the production of therapeutic

proteins with improved efficacy. Current researches are exploring different techniques to either:

(i) modify the biosynthetic pathway of the host cells, for example to avoid potential immunogenic

structures, humanize glycosylation, or increase the presence of desired glycans, an approach

that may enable the use of non-mammalian expression systems that have some advantages from

an economical and product yield perspective; or (ii) directly alter the protein glycosylation, which

may ultimately allow the rational design of desired glycoprotein structures.

To conclude, major progresses are expected in the next few years concerning glycosylation

understanding, control, and analysis. The relevance of glycosylation for product quality, the

diversity of paths that remain to explore, and the certainty that many challenges will still emerge,

make this a particularly appealing field of research.

“It always

seems

impossible

until it’s

done.”

Nelson Mandela

ana rita martins costa - core.ac.uk · trago uma caixinha mágica, à qual recorro com frequência,...

Documents