Methods in molecular biology vol 1603 heterologous protein production in CHO cells methods and protocols

The next generation of improvements is expected to be made via genetic engineering of the host CHO cell itself to increase or decrease the expression of endogenous genes depending on the

Heterologous Protein

Production

in CHO Cells

Paula Meleady Editor

Methods and Protocols

Methods in

Molecular Biology 1603

Me t h o d s i n Mo l e c u l a r Bi o l o g y

Series Editor

John M Walker School of Life and Medical Sciences University of Hertfordshire Hatfield, Hertfordshire, AL10 9AB, UK

For further volumes:

http://www.springer.com/series/7651

Heterologous Protein Production in CHO Cells

Methods and Protocols

Edited by

Paula Meleady

National Institute for Cellular Biotechnology, Dublin City University, Dublin, Ireland

ISSN 1064-3745 ISSN 1940-6029 (electronic)

Methods in Molecular Biology

ISBN 978-1-4939-6971-5 ISBN 978-1-4939-6972-2 (eBook)

DOI 10.1007/978-1-4939-6972-2

Library of Congress Control Number: 2017935545

© Springer Science+Business Media LLC 2017

This work is subject to copyright All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction

on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed.

The use of general descriptive names, registered names, trademarks, service marks, etc in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use.

The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to

be true and accurate at the date of publication Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations Cover Illustration: The front cover image, kindly provided by Alan Costello (National Institute for Cellular Biotechnology, Dublin City University), shows Chinese hamster ovary (CHO) cells with inducible green fluorescent protein (GFP) expression (from Chapter 6).

Printed on acid-free paper

This Humana Press imprint is published by Springer Nature

The registered company is Springer Science+Business Media LLC

The registered company address is: 233 Spring Street, New York, NY 10013, U.S.A.

Editor

Paula Meleady

National Institute for Cellular Biotechnology

Dublin City University

Dublin, Ireland

Since their introduction into the market over 20 years ago, biotherapeutics have tuted a large and growing percentage of the total pharmaceutical market, as well as approxi-mately 25% of the R&D pipeline in industry These biotherapeutics are having a huge global impact on the treatment of challenging and previously untreatable chronic disease Currently biopharmaceuticals generate global revenues of $163 billion, making up about 20% of the pharma market, and predicted to grow to over $320 billion by 2020 The num-ber of approved products in Europe and the USA has steadily increased to 2016 in 2014,

consti-of which 37 have “blockbuster” status, i.e., sales over $1 billion per year, with monoclonal antibodies (Mabs) representing the most lucrative single product class [1] Most signifi-cantly, nearly 50% of these biopharmaceutical products are produced in a single production host, i.e., Chinese hamster ovary (CHO) cells Improving the efficiency of production of these biologics will be critical in controlling costs to healthcare systems as more of these drugs come to market

There has been considerable success in developing high-producing CHO cell culture processes using approaches such as optimization of media formulation, improvements in expression vector design, and also improvements in the design of bioreactors The next generation of improvements is expected to be made via genetic engineering of the host (CHO) cell itself to increase or decrease the expression of endogenous genes depending on the desired outcome, in order to improve the efficiency of the production of therapeutic protein product In order to enhance the production capabilities and efficiency of the host cell line, an increased understanding of cellular physiology of CHO cells is of critical impor-tance There are substantial research efforts in progress focusing on the ‘omic analysis and systems biology of CHO cells to understand CHO cell physiology The publication of the draft CHO-K1 genome in 2011 represented a major milestone in CHO systems biology This information has been supplemented further with the publication of draft genomes for Chinese hamster and the CHO-S, CHO DG44 and CHO DXB11 cell lines Availability of the genome sequence will facilitate the interpretation and analysis of transcriptomic and proteomic data to assess the physiological state of the cells under different growth and pro-duction systems Combining all levels of regulation through systems biology models will unveil the underlying complexity inherent in CHO cell biology and will ultimately enhance and accelerate CHO productive capabilities in the coming decades

This book includes reviews and protocols for genetic manipulation of CHO cells for recombinant protein production, including “difficult-to-express” therapeutics A method is also included on the use of the recently described genome editing tool, CRISPR/Cas9, and how this can be applied to CHO cells The book also includes a review and protocols for characterization of CHO cells using ‘omic approaches and how these methods can be used

to improve efficiency of recombinant protein production during cell line development Analytical methods for characterization of recombinant protein product, such as glycosyl-ation and host cell protein analysis, are also described in this book

Preface

I am deeply grateful to all authors for giving up their valuable time and for contributing

to the book I would also like to thank the series editor, Prof John Walker, for help and guidance during the process of getting the book to publication

Reference

1 Walsh G (2014) Biopharmaceutical benchmarks 2014 Nat Biotechnol 32(10):992–1000

Preface

Contents

Preface v Contributors ix

1 Strategies and Considerations for Improving Expression of “Difficult

to Express” Proteins in CHO Cells 1

Christina S Alves and Terrence M Dobrowsky

2 Glycoengineering of CHO Cells to Improve Product Quality 25

Qiong Wang, Bojiao Yin, Cheng-Yu Chung, and Michael J Betenbaugh

3 Large-Scale Transient Transfection of Chinese Hamster Ovary Cells

in Suspension 45

Yashas Rajendra, Sowmya Balasubramanian, and David L Hacker

4 Cloning of Single-Chain Antibody Variants by Overlap- Extension PCR

for Evaluation of Antibody Expression in Transient Gene Expression 57

Patrick Mayrhofer and Renate Kunert

5 Anti-Apoptosis Engineering for Improved Protein Production

from CHO Cells 71

Eric Baek, Soo Min Noh, and Gyun Min Lee

6 Conditional Knockdown of Endogenous MicroRNAs in CHO Cells

Using TET-ON-SanDI Sponge Vectors 87

Alan Costello, Nga Lao, Martin Clynes, and Niall Barron

7 Application of CRISPR/Cas9 Genome Editing to Improve

Recombinant Protein Production in CHO Cells 101

Lise Marie Grav, Karen Julie la Cour Karottki, Jae Seong Lee,

and Helene Faustrup Kildegaard

8 Improved CHO Cell Line Stability and Recombinant Protein Expression

During Long-Term Culture 119

Zeynep Betts and Alan J Dickson

9 Selection of High-Producing Clones Using FACS for CHO

Cell Line Development 143

Clair Gallagher and Paul S Kelly

10 The ‘Omics Revolution in CHO Biology: Roadmap to Improved

CHO Productivity 153

Hussain Dahodwala and Susan T Sharfstein

11 A Bioinformatics Pipeline for the Identification of CHO Cell Differential

Gene Expression from RNA-Seq Data 169

Craig Monger, Krishna Motheramgari, John McSharry, Niall Barron,

and Colin Clarke

12 Filter-Aided Sample Preparation (FASP) for Improved Proteome

Analysis of Recombinant Chinese Hamster Ovary Cells 187

Orla Coleman, Michael Henry, Martin Clynes, and Paula Meleady

13 Phosphopeptide Enrichment and LC-MS/MS Analysis to Study the

Phosphoproteome of Recombinant Chinese Hamster Ovary Cells 195

Michael Henry, Orla Coleman, Prashant, Martin Clynes,

and Paula Meleady

14 Engineer Medium and Feed for Modulating N-Glycosylation

of Recombinant Protein Production in CHO Cell Culture 209

Yuzhou Fan, Helene Faustrup Kildegaard, and Mikael Rørdam Andersen

15 Glycosylation Analysis of Therapeutic Glycoproteins Produced

in CHO Cells 227

Sara Carillo, Stefan Mittermayr, Amy Farrell, Simone Albrecht,

and Jonathan Bones

16 Characterization of Host Cell Proteins (HCPs) in CHO Cell Bioprocesses 243

Catherine E.M Hogwood, Lesley M Chiverton, and C Mark Smales

Index 251

Contents

Dublin, Ireland

of Denmark, Kgs Lyngby, Denmark

eric bAek • Department of Biological Sciences, KAIST, Daejeon, Republic of Korea

Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland

Dublin, Ireland

Johns Hopkins University, Baltimore, MD, USA

University, Izmit, Kocaeli, Turkey

Dublin, Ireland

Dublin, Ireland

University of Kent, Canterbury, Kent, UK

cheng-yu chung • Department of Chemical and Biomolecular Engineering,

Johns Hopkins University, Baltimore, MD, USA

Gaithersburg, MD, USA; SUNY Polytechnic Institute, Albany, NY, USA

AlAn J dickSon • Faculty of Life Sciences, The University of Manchester, Manchester, UK

Lyngby, Denmark; The Novo Nordisk Foundation Center for Biosustainability, Technical University of Denmark, Hørsholm, Denmark

Dublin, Ireland

Dublin, Ireland

Contributors

liSe mArie grAv • The Novo Nordisk Foundation Center for Biosustainability, Technical

University of Denmark, Hørsholm, Denmark

dAvid l hAcker • Laboratory of Cellular Biotechnology (LBTC), École Polytechnique

Fédérale de Lausanne (EPFL), Lausanne, Switzerland; Protein Expression Core Facility (PECF), École Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland

Dublin, Ireland

University of Kent, Canterbury, Kent, UK

pAul S kelly • National Institute for Cellular Biotechnology, Dublin City University,

Dublin, Ireland

Biosustainability, Technical University of Denmark, Lyngby, Denmark

Biosustainability, Technical University of Denmark, Lyngby, Denmark

University of Natural Resources and Life Sciences-Vienna, Vienna, Austria

ngA lAo • National Institute for Cellular Biotechnology, Dublin City University, Dublin,

Ireland

gyun min lee • Department of Biological Sciences, KAIST, Daejeon, Republic of Korea

JAe Seong lee • The Novo Nordisk Foundation Center for Biosustainability, Technical

University of Denmark, Lyngby, Denmark

University of Natural Resources and Life Sciences-Vienna, Vienna, Austria

Dublin, Ireland

Dublin, Ireland

(NIBRT), Dublin, Ireland

Dublin, Ireland; National Institute for Cellular Biotechnology, Dublin City University, Dublin, Ireland

(NIBRT), Dublin, Ireland; National Institute for Cellular Biotechnology, Dublin City University, Dublin, Ireland

Soo min noh • Department of Biological Sciences, KAIST, Daejeon, Republic of Korea

Ireland

de Lausanne (EPFL), Lausanne, Switzerland; Biotechnology Discovery Research, Lilly Research Laboratories, Eli Lilly and Company, Lilly Corporate Center, Indianapolis, IN, USA

SuSAn t ShArFStein • SUNY Polytechnic Institute, Albany, NY, USA

Kent, Canterbury, Kent, UK

Qiong wAng • Department of Chemical and Biomolecular Engineering, Johns Hopkins

University, Baltimore, MD, USA

University, Baltimore, MD, USA

Contributors

Paula Meleady (ed.), Heterologous Protein Production in CHO Cells: Methods and Protocols, Methods in Molecular Biology,

vol 1603, DOI 10.1007/978-1-4939-6972-2_1, © Springer Science+Business Media LLC 2017

Chapter 1

Strategies and Considerations for Improving Expression

of “Difficult to Express” Proteins in CHO Cells

Christina S Alves and Terrence M Dobrowsky

Abstract

Despite substantial advances in the field of mammalian expression, there are still proteins that are characterized

as difficult to express Determining the expression bottleneck requires troubleshooting techniques specific for the given molecule and host The complex array of intracellular processes involved in protein expres- sion includes transcription, protein folding, post-translation processing, and secretion Challenges in any

of these steps could result in low protein expression, while the inherent properties of the molecule itself may limit its production via mechanisms such as cytotoxicity or inherent instability Strategies to identify the rate-limiting step and subsequently improve expression and production are discussed here.

Key words Productivity, Difficult to express, Vector design, Cell engineering, Process optimization

1 Introduction

CHO cells have been utilized extensively for recombinant protein expression; however, not all proteins are expressed at high levels in this host system There are many reasons why a protein may be

“difficult to express” and require an alternative strategy to standard platform workflows for CHO cell production Although there are clearly monoclonal antibodies (mAb) that can be challenging to express at industry standard levels of 5 g/L or more, productivity improvements for non-mAb therapeutic proteins have lagged behind [1 2] It is more difficult to define productivity levels that constitute low expression for non-mAb products and a molecule may be difficult to express not only because of intracellular chal-lenges but also due to its biophysical properties Determining the expression bottleneck requires troubleshooting techniques specific for the given molecule and has historically focused on transcrip-tion, protein folding, post-translational processing, and secretion [3–6] Alternatively, challenges in producing unique or difficult to express proteins may have solutions in the bioprocessing space such as operating parameters or media and feed optimization

Often, business, regulatory, or biological limitations may require the introduction of additional process steps or modifications to reach yield demands with an existing cell line Previous bioprocess strategies for improved protein production include chemically induced specific productivity increases, affecting secreted protein stability or toxicity in culture, continuous removal of protein from culture, and general increases to culture biomass The diversity that exists among the different CHO host lineages can also be lev-eraged to improve expression of problematic proteins and antibod-ies [7 8]

Engineering CHO cells and their production bioprocess to better express difficult proteins requires a two-step approach by which you first determine what is the rate-limiting function and then develop a strategy to alleviate it This chapter will outline vari-ous strategies that can be used to determine the expression bottle-neck and consequently to improve protein expression

2 Strategy and Methods

In order to design a comprehensive set of experiments to increase the production of a difficult to express protein, a fundamental understanding of the biophysical and biochemical properties of the protein is essential The process by which a DNA sequence is con-verted to a fully folded protein product is complex with steps that include transcription, translation, post-translational modification, protein folding, and ultimately secretion (Fig 1) Any of these indi-vidual steps could limit protein expression and may be attributed to either a poorly designed molecule or suboptimal DNA coding sequence Inherent properties of the molecule can also result in the protein being prone to degradation, aggregation, and other unfa-vorable inter-protein interactions that can lead to cytotoxicity Due

to these factors, knowledge of the biology of the protein will aid in narrowing the scope and focus of troubleshooting efforts

Transfection of a gene into a cell is followed by the integration of that gene into the host cell’s genome This has historically been a random event in standard cell line development protocols Efficient expression of the transgene is highly dependent on both the num-ber of gene copies integrated into the genome as well as the sites of integration The latter is greatly influenced by positional effect variation, which is affected by the local permissiveness of the site as well as the proximity and interaction with local and distal enhancers Although it has been shown that high copy numbers of transgenes

do not always correlate with high cellular productivity [9 10], the number of integrated transgenes is an important parameter to measure as it has been shown to affect expression in some cell lines [4 11] Several methods can be used to measure transgene copy

2.1 Determining

the Bottleneck

2.2 Integration

of the Gene of Interest

Christina S Alves and Terrence M Dobrowsky

number including analysis by southern blot [4], qPCR [11], and digital droplet PCR [12] In the situations where gene copy num-ber does not correlate with product expression it is possibly a result

of the transgene being integrated in a suboptimal location in the genome Random genome insertion could result in the transgene benefiting from a location in a highly transcribed genetic region of euchromatin (referred to as a “hot spot”) or possibly suffering from the effects of epigenetic gene silencing By using a targeted integration approach whereby the gene of interest is inserted into

a predetermined loci in the genome, one can circumvent such issues by integrating the gene into a known hot spot where a region

of euchromatin and high gene-expression have already been lished Determination of the desired hot spot is often the most challenging part of developing a targeted integration system One approach is to utilize a screen of high expressing cell lines to deter-mine whether expression is driven by a single copy of the trans-gene The location of the single integration site may be in an area that naturally drives high expression and which can be utilized for other genes of interest Elaborate systems to determine permissive loci for integration have been used such as transfecting CHO cells with a plasmid containing a FRT-tag to specifically screen for single integration loci with high transcriptional activity [13] In this work,

Post-Protein Folding

Fig 1 Summary of protein synthesis RNA is transcribed in the nucleus and then transported to the cytoplasm

and translated by the ribosomes The proteins become bound to the rough ER, where they undergo folding and processing before moving to the golgi Soluble proteins undergo post-translational modifications and are sub-sequently processed through the secretory pathway

Optimizing ‘Difficult to Express’ Proteins

fluorescence in situ hybridization (FISH) was used to locate the integration site of the FRT sequence or the antibody genes in the chromosomes Given the substantial advances in the field of CHO

‘omics, it is now feasible to use next generation sequencing (NGS)

to determine hot spots for integration Advancements in this nology have increased the speed and throughput of whole genome (DNA-seq) and transcriptome (RNA-Seq) sequencing such that it

tech-is now feasible to screen clones for the location of genes that have

a high level of expression A more refined method is targeted sequencing where the genome is fragmented, incubated with probes specific for the transgene, and then enriched via a wash step This enables sequencing of just the genes with high expres-sion to elucidate their location in the euchromatin Although it has not yet been demonstrated, it may be possible to screen early in the cell line development process and identify clones that display a pre-defined ‘omics profile that is predictive of productivity using RNA-seq [14]

Once a desired site has been elucidated, several methods exist

to insert a gene of interest into a specific location These methods which include site-specific recombinases, integrases, or transposases for the integration of the expression cassettes are summarized in Table 1 Integrases and transposases allow for multiple integrations with higher copy numbers at various recognition sites within the genome [15] Phage integrases such as PhiC31 integrase rely on unmodified, native acceptor (attP) and donor (attB) sites, but the quantity of these sites in the CHO genome may be limiting On the other hand, site-specific recombinases have a higher specificity

of integration into a predetermined single site [16] Flp nases have been used in combination with Flp recognition target

recombi-Table 1

Methods for targeted integration

Integrases and

transposases Multiple integrations with higher copy numbers using native

donor and receptor sites

Integrate randomly, limited number of sites in CHO genome

TALENs Easy to design for knock in/out,

target DNA sequences using proteins

High frequency of insertion- deletion mutations, expensive, and time consuming to develop

[ 21 ]

CRISPR/Cas9

system Target-specific DNA sequences using RNA, inexpensive, and

able to screen many sites quickly

Can have off target effects,

IP landscape is undefined [22, 23]Christina S Alves and Terrence M Dobrowsky

sites (FRT) for targeted integration of transgenes into mammalian cells with a high specificity of integration and low off target effects [17] This is accomplished either by using Flp-in or Flp recombinase- mediated cassette exchange (RMCE) strategies RMCE uses a set

of hetero-specific FRT sites to direct a gene of interest to a termined and tagged locus that has been characterized to yield high protein expression [18] A binary RMCE expression system has been used to co-express multiple proteins with different com-binations of expression levels [17] The Cre/loxP system for site- specific DNA recombination has also been used as a tool for transgene integration in CHO cells [19] Recent work has demon-strated the ability to insert multiple transgenes into a targeted site

prede-of the CHO cell genome using Cre recombinase-incorporating integrase-defective retroviral vectors [20] More recently, site- specific gene insertion in CHO cells has been performed using transcription activator-like effector nucleases (TALENs) [21] or CRISPR/Cas9 RNA-guided nucleases [22] Precise insertion of a gene expression cassette at a defined loci in CHO cells has been accomplished using the CRISPR/Cas9 system following a simple drug-selection methodology that resulted in homogeneous trans-gene expression [23]

Traditionally, transcription has been considered the dominant tor in controlling protein production

fac-In a particular study to elucidate the mechanisms and processes- limiting gene expression in CHO cells, transcription appeared to

be the primary limitation for low- and medium-producing cell lines, whereas in high-producing cell lines post-translational limita-tions tended to dominate [6] Within the process of transcription the rate-limiting step is likely to be initiation Due to the highly condensed nature of DNA into chromatin structures, transcription complexes often have trouble accessing certain regions of DNA Chromatin remodeling, the rearrangement of chromatin structure by various remodeling complexes, is therefore required for activation of transcription Additionally, chromatin as well as other proteins involved in transcriptional control can be altered by methylation, acetylation, phosphorylation, and other modifica-tions to affect whether a gene is active or inactive [24] The syner-gistic effect of these modifications, known as the “histone code,” adds complexity in the form of epigenetic regulation of genes Specific methods to affect these features are detailed here

Unstable protein expression has been observed in CHO cells where mRNA decreases despite constant transgene copy numbers [11], which suggests that either the mRNA is degrading or that the promoter is being silenced The expression level of mRNA tran-script for the gene of interest can be determined by quantitative real-time reverse transcription-PCR (qRT-PCR) There have been mixed results on the correlation between productivity of a cell line

2.3 Transcription

Optimizing ‘Difficult to Express’ Proteins

and mRNA levels in CHO cells Some studies showing that high mRNA levels and gene copy numbers in methotrexate amplified cells correspond to high specific productivities [4], whereas others have seen no correlation between mRNA levels and expression [25] Despite these conflicting reports, it may still be valuable to assess mRNA levels of the protein of interest to ensure that the sequence of interest is being adequately transcribed Additionally, in the case of antibodies, using qPCR to determine mRNA levels can

be a useful diagnostic tool for determining the ratio of heavy to light chain which may be important to ensure assembly of the mAb Studies have indicated that it is advantageous to have an excess of light chain in relation to the heavy chain for optimal antibody pro-duction [26, 27] The ratio of heavy to light chain can be influ-enced by optimizing the quantities of DNA transfected or by the vector design If a two-plasmid system is utilized, where the heavy and light chains are located on different vectors, the mixture of DNA used at the point of transfection can be used to modulate the ratio Alternatively, one can use a single- plasmid system that employs IRES-mediated bi- or tri-cistronic vectors that enable control of heavy to light chain expression at different ratios

DNA methylation has been reported to repress gene expression, whereas hypomethylation of DNA in the promoter region can ele-vate gene transcription activity [28, 29] Enzymatic methylation of cytosine at carbon 5 is well known as a fundamental epigenetic mechanism that results in gene silencing [30] DNA methylation often occurs at CpG dinucleotides sites within promoter regions which subsequently renders the promoter transcriptionally inactive Bisulfite treatment of DNA can be used to differentiate between methylated and unmethylated CpG sites In this method, sodium bisulfite converts cytosine residues to uracil residues via deamination

at C4, while 5-methylcytosine remains unaffected [31] Subsequent amplification of the region by PCR allows for further analysis via DNA sequencing [31] or microarray analysis [32] Methylation-specific real-time qPCR is a highly sensitive measurement of pro-moter methylation and has been utilized to correlate hCMV-IE methylation with unstable protein expression in recombinant CHO cell lines [28] Chemical compounds exist that can affect the degree

of DNA methylation, specifically a class of molecules known as DNA methyltransferase inhibitors (iDNMTs) These compounds, which include azacytidine, RG-108, and hydralazine, have been tested in CHO cells for their capacity to increase cellular productivity in tran-sient gene expression systems with some success [33]

Acetylation of histones typically plays a role in transcriptional control

of active genes [34] Histone acetyltransferases (HATs) and histone deacetylases (HDACs) control the enhancement of transcription by modifying histone acetylation The most commonly used mechanism

2.3.1 Methylation

2.3.2 Acetylation

Christina S Alves and Terrence M Dobrowsky

to control acetylation in CHO cell cultures is the use of HDAC inhibitors to prevent deacetylation Several studies have demonstrated that sodium butyrate [35, 36] and/or valproic acid [37] can be used

to enhance mRNA transcription and increase specific productivities However, these compounds can also have adverse negative effects on cell growth due to cytotoxicity and induction of apoptosis [38] Their appropriateness must be evaluated to determine the optimal concentration for enhanced productivity, but they are commonly used over short production durations

The design of vectors to promote active transcription by creating a favorable chromatin environment around the transgene has been extensively reviewed [39] The available methods either alter the epi-genetic environment of the DNA surrounding the transgene or pre-vent the surrounding environment from affecting transcription of the gene of interest [39] A list of vector design elements for CHO cells

is shown in Table 2 In order to enhance gene transcription and reduce transgene expression dependence on the surrounding chro-matin, strong cellular enhancers such as the Locus Control Region (LCR) have been utilized [40] The LCR is a cis- acting DNA ele-ment that controls the expression of human β-globin locus genes Unfortunately, these enhancers only function in certain cell lines and cannot be used as general regulatory elements in all mammalian cells

2.3.3 Vector Design

Elements

Table 2

Vector design elements to enhance transcription

Locus control

HGH Can lead to stable high expression of transgene in a copy number dependent

manner Limited usefulness in mammalian cell lines

[ 40 ]

modestly increase transgene expression in CHO, but may not be universally effective

[ 41 ]

Matrix Attachment

Regions (MARs) Chicken lysozyme MAR, human

β-globin MAR, X MAR

Bind to the nuclear matrix and affect the arrangement of chromatin into loops

Have shown some positive effects on transgene expression in CHO

Large increases in gene expression were observed but elements are typically large (~16 kb), promoter dependent

copy number-dependent stable expression

[ 45 , 46 ] Optimizing ‘Difficult to Express’ Proteins

Regulatory elements that block interactions between the enhancer and promoter while not directly affecting their individual activity are referred to as insulators Insulators such as the chicken β-globin 5′ hypersensitive site 4 (cHS4) have been used to control the effects of the surrounding chromatin environment on the transgene [41].Matrix-attachment regions (MARs) are DNA elements that bind to the nuclear matrix and are believed to influence gene expression by affecting the arrangement of chromatin into loops MARs, such as chicken lysozyme MAR, human β-globin MAR, and X MAR, can associate with euchromatin and act as boundary

or insulator elements, and hence create an independent chromatin structure from the surroundings [42] Although the specific mech-anisms by which MARs function in the cell are not entirely under-stood, they have been effective in enhancing the expression of target proteins in mammalian cell cultures MARs can be integrated into expression vectors that may increase the percentage of high- producer cells in a population to reduce the number of clones that need to be screened Protocols are available that describe how to incorporate MARs into vectors that can then be transfected into CHO cells for increased transgene expression [43]

Other elements that have been shown to protect a transgene from silencing and convey higher transgene expression are ubiquitous chro-matin opening elements (UCOEs), which are derived from the pro-moters of housekeeping genes that are typically transcriptionally active [44] Some well-characterized UCOE pairs include HNRPA2B1 and

CBX3 or TBP and PSNB1, which are DNA regions that contain a pair

of divergent gene promoters that are transcriptionally active in all cells

of an organism Large UCOEs of up to 16 kb have been used to erate high-level and stable transgene expression for cells in extended culture by increasing the efficiency of the CMV promoter Because UCOEs directly affect transcriptional regulation that is dependent on the promoter and its activity, these elements have variable effects on expression of a target protein in CHO cells and need to be tested for specific host and vector strategies [41]

gen-Antirepressor or STAR (stabilizing and antirepressor) are DNA elements that block chromatin-associated repressors and have been used to flank transgenes in mammalian expression vectors These elements affect the spread of methylation and histone deacetylation from the adjacent chromatin environment into the transgene region They can enhance protein expression as well as overcome genetic instability caused by positional effects, epigenetic silencing,

or loss of gene copy number [45] The positive effects of STAR elements are most pronounced when high selection stringency is used to develop stable clones in CHO cells [46]

The process of translation consists of initiation, elongation, nation, and recycling The initiation of mRNA translation is an essential precursory step that influences cell growth and protein

termi-2.4 Translation

Christina S Alves and Terrence M Dobrowsky

synthesis via the coordination of numerous initiation factors [47] The secondary structure of the mRNA can affect translational effi-ciency Formation of a closed loop structure consisting of mRNA,

a number of eukaryotic initiation factors (eIFs), and ribosomal proteins can potentially increase global translation efficiency by promoting re-initiation of translation High-producing cell lines have been shown to maintain appropriate levels of these translation initiation factors [48] Use of cell engineering approaches to main-tain the levels of these initiation factors may allow for generation of new host cell lines with high growth and recombinant protein pro-ductivity Another target to improve translation is the global meta-bolic sensor and processing protein mammalian target of rapamycin (mTOR) The treatment of CHO cell cultures with adenosine results in growth arrest but also increases productivity The ade-nosine contributes to high ATP levels which increase mTOR activ-ity, inhibiting the key translation initiation repressor 4E-BP1 [49] mTOR has also been shown to influence ribosomal protein synthe-sis, translation initiation, and translation elongation in addition to other cellular functions Its overexpression in CHO cells has resulted in increased specific antibody productivity [50] making it

an attractive engineering target for difficult to express proteins.Because it is often the case that human proteins are being express-ing in CHO cells and synonymous codons are used with different frequencies in different organisms (known as codon bias) [51], it is important to ensure that the transgene sequence is optimized By optimizing the DNA sequences for expression in CHO cells, one can ensure that certain preferred codons are translated more accu-rately and/or efficiently Poorly optimized sequences can adversely affect protein translation, and subsequently protein expression, by preventing the host from efficiently translating the rare codons Codon optimization has been used to increase protein expression

in multiple studies [52, 53] and there are several websites and vices that will perform codon optimization for expression in CHO cells of a given amino acid sequence A list of codon usage for CHO cells is shown in Table 3

ser-Another important consideration is the translation initiation sequence located upstream of the start codon (AUG) The efficient consensus sequence GCCACC(AUG)G, known as the Kozak sequence [54], yields high fidelity and efficiency of initiation and is typically used at the start of the coding sequence

Splice sites are located between an exon and an intron The splice site upstream of an intron is referred to as the donor splice site (5′–3′ direction), while the one downstream of an intron is the acceptor splice site (3′–5′ direction) The acceptor splice site corresponds to the end of an intron (AG) and the donor splice site corresponds to the beginning of an intron (GT) Splice sites can

2.4.1 Codon Optimization

2.4.2 Splice Sites

Optimizing ‘Difficult to Express’ Proteins

also unintentionally exist in a coding sequence As possible tor and donor splice sites, every AG and GT in a DNA sequence needs to be evaluated as either a real splice site or a pseudo splice site to ensure that the sequence is not compromised during transla-tion In addition to the sequences immediately adjacent to the

accep-Table 3

Codon usage in Chinese hamster genes

Amino

acid Codon Relative frequency Amino acid Codon Relative frequency Amino acid Codon Relative frequency

The amino acid abbreviation is shown adjacent to the codon and the relative frequency in identified genes of the

Chinese hamster (Cricetulus griseus) The source of these data is http://www.kazusa.or.jp/codon/ These records were

a snapshot of usage as of March 2016 A total of 331 genes and 153,527 codons contributed to this data set.

Christina S Alves and Terrence M Dobrowsky

splice event, distal sequences also contribute to the probability of splicing Several programs that are summarized in Table 4 exist online to help evaluate a sequence and the probability that donor and acceptor splice sites are present

Proper protein folding is essential for adequate expression of a molecule The ER is responsible for ensuring that proteins are properly processed and folded and as such there are specific quality control systems to aid in the efficiency of folding and eliminate misfolded proteins When a protein is misfolded in the ER it is proteolytically destroyed via the ER-associated degradation (ERAD) pathway Similarly, the unfolded protein response (UPR),

a signal cascade that protects cells from aggregated protein by restoring ER function, can be triggered by intracellular accumula-tion of misfolded protein Several chaperones and cofactors are involved in the process of protein folding and assembly and can be modulated to enhance protein expression

Molecular chaperones are proteins that assist the folding and bly of intracellular proteins which may be good targets for cellular engineering to improve protein expression Heat shock proteins (HSPs) function as molecular chaperones and are primarily respon-sible for protein folding, assembly, translocation, and degradation under cellular stress Chaperones also prevent newly synthesized polypeptide chains from aggregating into defective proteins BiP is

assem-a HSP70 moleculassem-ar chassem-aperone thassem-at binds newly synthesized teins as they are translocated into the ER, and preserves them in a state suitable for subsequent folding Protein disulfide isomerase (PDI) is an enzyme in the ER that catalyzes the formation and breakage of disulfide bonds to assist in protein folding [55] Cyclophilin B (CypB) interacts with other proteins in the ER including BIP and PDI to form chaperone complexes that facilitate protein folding Some work has been done on expressing molecular

pro-2.5 Protein Folding

and Processing

2.5.1 Chaperones

Table 4

Programs available to identify potential splice sites in a DNA sequence

GENIO splice site and

exon predictor http://biogenio.com/splice/

SpliceView http://l25.itba.mi.cnr.it/~webgene/wwwspliceview.html [ 111 ]

Optimizing ‘Difficult to Express’ Proteins

by addition of chemical chaperones at the start of stationary phase increased cell-specific production and eliminated protein aggrega-tion [57] The disparity in these findings suggests that the engineer-ing of molecular chaperones for increased protein expression may

be product and cell line dependent

Chemical chaperones are a group of compounds that can improve the folding capacity of the ER, facilitate protein folding in the ER, and enhance the secretion of protein Chaperones can be added to cell culture to potentially improve expression of a difficult to express protein, especially if misfolding or aggregation is occurring intracel-lularly This approach provides a simpler alternative to overexpression

of molecular chaperones given their uncertain effect on productivity PBA (Sodium 4-phenylbutyrate) has been used to promote the secre-tion of a mutated protein C from CHO cells by utilizing an uncon-ventional GRASP55-dependent pathway that restores normal intracellular trafficking through the ER and golgi [58] Treatment of mammalian cells with PBA has been shown to suppress ER stress by chemically enhancing the ER capacity to cope with the expression of misfolded protein, preventing intracellular aggregates by facilitating protein degradation [59] Osmotically active chaperones such as DMSO, glycerol, and proline have been used to increase specific pro-ductivity, but also can have a negative impact on cell growth Additionally, DMSO and proline can reduce protein aggregate for-mation in culture supernatants by an undefined mechanism [60] To counteract the negative effect on cell growth and viability, the addi-tion of DMSO [61] and glycerol [62] in a two-staged approach has been utilized to increase the specific productivity of CHO cell lines Similarly, a combination of PBA and glycerol has been added at the start of stationary phase alongside expression of the molecular chap-erone CypB to maximize cell specific production and eliminate pro-tein aggregation [57] Analogous to the overexpression of molecular chaperones, the effect of these chemicals may be cell line and protein specific and their suppression of cell growth requires that their con-centration and dosing strategy be carefully considered

In addition to enabling or improving chaperone protein function, aggregation can be prevented through bioprocess modifications Altering the cellular redox potential by supplementing media with the antioxidant glutathione can reduce aggregation [63], while media optimization of components such as cysteine or glycerol can reduce aggregation [64, 65] Additionally, a reduction in tempera-ture has been shown to reduce aggregation and positively affect

2.5.2 Bioprocess

Modifications

Christina S Alves and Terrence M Dobrowsky

Translocation of a nascent protein from ribosomes through the cytosol into the endoplasmic reticulum is mediated by its signal peptide and is an essential stage in protein secretion The efficient secretion of recombinant proteins from CHO cells is strongly dependent on the signal peptide used, which makes identifying the optimal signal sequence for each target protein an important step

in maximizing the efficiency of protein secretion [71] In malian cells, a signal peptide that ranges from 5 to 30 amino acids

mam-at the N-terminal end of nascent proteins is recognized by the nal recognition particle (SRP) in the cytosol as the protein is being synthesized on the ribosome The SRP then transfers the complex consisting of the SRP and ribosome-nascent chain to a receptor on the endoplasmic reticulum (ER) membrane, where it is eventually translocated to the lumen of the ER and the signal peptide is cleaved by a signal peptide peptidase [72] The translocation of proteins into the ER lumen is considered a bottleneck of the secre-tory pathway and has motivated further investigation into enhanc-ing the capacity of signal peptides for recombinant protein expression Several studies have shown positive effects with native signal peptides, natural signal peptides derived from human albu-min and human azurocidin, as well as optimized signal sequences [71, 73, 74] indicating the importance of carefully evaluating sig-nal sequences for the expression of a given protein There are sev-eral resources online (Table 5) that assess the probability that a peptide is a suitable signal sequence as well as how efficiently the sequence will be cleaved from the protein [75, 76]

sig-Secretion of antibodies has been affected by improper cleavage

of the light chain from the signal peptide due to a dysfunctional SRP

2.6 Secretion

Optimizing ‘Difficult to Express’ Proteins

complex, which results in its precipitation in an insoluble cellular fraction [3] Western blotting of intracellular fractions can be used to determine whether light chain is precipitating in the cells This can

be achieved by using standard lysis techniques to create protein extracts, followed by blotting with light chain-specific antibodies If this inadequate cleavage of the light chain is affecting expression, proper processing and secretion can be restored by over-expressing SRP proteins such as the signal recognition protein, SRP14 [3].Russell bodies are intracellular aggregates of immunoglobulins stored in the endoplasmic reticulum that can form during protein biosynthesis The formation of Russell bodies depends on the phys-iochemical properties of the protein coded by the variable regions of the heavy and light chains as well as extrinsic factors such as stressful cell culture conditions [77] Immunofluorescent microscopy can be used to determine whether Russell bodies are forming intracellularly for a given protein Cells must be fixed, permeabilized, and stained with fluorescently conjugated antibodies to the IgG of interest fol-lowed by fluorescent imaging and quantification [78] A frequency

of Russell body phenotype can be calculated by determining the number of Russell bodies observed and normalizing to the overall number of cells in the image This value can be compared to an alternative cell line that produces an easy to express molecule to determine whether intracellular aggregation is resulting in reduced protein expression Recent studies suggest that there are IgG anti-body sequences with intrinsically high condensation/aggregation propensities that are more prone to form Russell bodies in the ER lumen [78] This implies that if a protein is not being expressed due

to the formation of these intracellular aggregates the sequence may need to be altered to enable better expression

Another mechanism that leads to insufficient protein production is the inherent toxicity of the protein being expressed Limiting cel-lular exposure to high concentrations of the protein or adapting cell lines specifically to be resistant to the toxic protein can improve growth and subsequently yield Toxicity of the protein of interest may diminish a cell’s ability to recover after transfection and selec-tion Ultimately, this toxicity will result in the unintentional selec-tion of low-producing cells from the population A dose-response

2.6.1 Russell Bodies

2.7 Protein Toxicity

Table 5 Websites that offer free signal peptide prediction algorithms

SignalP 4.1 Server ( http://www.cbs.dtu.dk/services/SignalP/ ) PrediSi: Prediction of Signal peptides ( http://www.predisi.de/ )

Signal-BLAST Signal Peptide Prediction (http://sigpep.services.came sbg.ac.at/signalblast.html )

Christina S Alves and Terrence M Dobrowsky

study with purified protein and host cells is often the most direct determination of cytotoxicity, wherein the toxicity of the purified protein and buffer to the naive culture is assessed If the quantity

of purified protein is limiting, similar results can be confirmed by a less ideal but easy-to-execute experiment utilizing spent media from a transient transfection Clarified culture supernatant from a transiently transfected culture will likely contain a toxic level of protein but sufficient unprocessed metabolites for continued growth Performing a dose-response study using this supernatant incrementally blended with fresh media may yield similar results to dosing purified protein However, toxicity will be confounded with other supernatant components such as metabolic waste prod-ucts and the highest concentration available for testing will be limited

Protein toxicity can be mitigated by multiple methods One approach, likely the most extreme, is to alter the protein itself to reduce its cytotoxicity Modifying the protein to produce a more stable, less toxic form while retaining its intended biological func-tion can be difficult but possible with extensive knowledge of the structure function relationship [79] Introducing stabilizing agents that can be degraded later can be effective as long as a more com-plicated downstream purification process is acceptable N-terminal tags such as a small ubiquitin-like modifier (SUMO) can be used to create “dormant fusion” proteins with decreased toxicity that are capable of being cleaved downstream [80] An alternative option is

to modify the promoter in the transfected vector rather than the protein of interest When the optimum environment for cell growth varies significantly from that of protein production, an inducible expression system may be appropriate An inducible expression system can enable high cell densities to be achieved prior to protein production and subsequently alleviate the effects

of toxicity or degradation [81] Inducible promoter systems are commercially available [82] for direct implementation In general,

it can be difficult to ensure that transgene expression is entirely inhibited prior to the addition of the inducing agent [83] Therefore, most industrially relevant systems utilize promotor-transactivator combinations In these systems, the activity of a con-stitutively expressed transactivator is controlled via supplementation

of some complementary ligand [84] The tetracycline (Tet) ible system, often referred to as Tet-on, allows for the expression of the protein of interest in the presence of tetracycline Alternatively, protein expression could be repressed in the presence of tetracy-cline, the Tet-off system, and activated by complete media exchange The Tet-on system is often preferred for recombinant protein production as it is relatively straightforward to supplement culture with tetracycline while removing it would require signifi-cant liquid handling at scale [85, 86] Other applications for induc-ible systems include increasing specific productivity by arresting

induc-Optimizing ‘Difficult to Express’ Proteins

cellular proliferation [87] In this case, protein production is not induced directly, but rather through the subsequent reactions to decreased cell growth such as increased mitochondrial mass and activity The selection of stable, clonal cell lines is more involved with inducible systems and may increase development time

Cell cultures can often be adapted to new growth ments to suit productivity needs Adapting cultures from serum containing to serum free, chemically defined medium for example

environ-or from adherent to suspension growth environments are place with commercially available materials and protocols [88, 89]

common-If the protein of interest is determined to be toxic, one option for improving production is adapting host cells Naive host cells cul-tured in the presence of low concentrations of the protein can decrease sensitivity of the cell line over time The resulting host culture can then be transfected and ideally recover with the ability

to survive higher productivity than before adaptation [8]

Concerns over protein toxicity or degradation may also be mitigated by reducing contact of the protein with producing cul-ture This can be performed through chemical supplementation, wherein the toxic protein is competitively inhibited from interact-ing with the cells through an antagonist [8] Alternatively, this can

be achieved through the use of perfusion growth systems where culture supernatant is removed continuously via filtration and cell mass is retained Perfusion growth systems are also commercially available and enable high cell densities and increased volumetric productivity [90, 91] However, these can be technically challeng-ing to implement even at smaller scales If small protein require-ments are desired for non-industrial purposes, complete media exchange through centrifugation is a simple alternative [92]

In addition to a protein’s ability to limit cell culture growth, the cell culture may limit the availability or stability of protein in the supernatant Adhesion to lipid head groups or lipoproteins in the cell membrane can accelerate protein degradation and reabsorp-tion into the producing cells themselves One example of such pro-tein loss is the generation of recombinant Factor VIII (rFVIII) [93] Here, 90% of the secreted rFVIII in serum-free conditions were determined to be bound to cells This effect was limited by supplementing culture with a complimentary protein, von Willibrand Factor (vWF) capable of competitive inhibition [93] Co-expressing vWF, rather than supplementation, was also found

to have a stabilizing effect [94] Co-expression of a complex tein antagonist is not typically the most efficient mitigation strat-egy and can complicate cell line selection Chemical inhibition may

pro-be possible and substantially easier to implement In this current example, rFVIII can be prevented from binding to host cell culture

by supplementation of o-phospho-l-serine (OPLS) [93, 95] Implementation of these methods will require specific knowledge

pH control can alter specific productivity of erythropoietin (EPO) producing culture at one temperature differently than another [98] Because these process parameter sensitivities will often be cell specific they are difficult to predict without in-depth experience with the cell line of interest.

Defining a well-understood operating space for all process parameters using Design of Experiment (DOE) methodologies is recommended and may reveal configurations that allow for suffi-cient protein expression [99, 100] Alternatively, shifts in process parameters during production to temporarily increase productivity can be useful [101] While this often results in decreased overall cell mass, that in turn may be compensated for by maintaining high process temperature during the growth phase of the culture A net increase in productivity with lower temperatures may be the result

of either increased specific productivity or decreased protein radation at reduced temperatures Growth arrest and subsequent increases in specific productivity can also be accomplished through hyperosmolality or chemical treatment with agents such as sodium butyrate or DMSO (as discussed earlier) [101, 102] Addition of a small molecule inhibitor of cyclin-dependent kinases (CDK) 4/6 mid-way through production can mediate G0/G1 growth arrest without impacting G2/M phase This resulted in sustained cell mass and increased specific productivity without negatively impact-ing product quality [103] While improvements in specific produc-tivity may be obtained through process condition shifts or media supplementation, the medium formulation itself may be optimized for a more consistent increase to protein production [104, 105] Also, temporary increases to specific productivity will likely increase product titer before cell death becomes a concern, it usually comes

deg-at the expense of product quality such as increased heterogeneity, decreased biological activity, altered acidic isoforms, and inconsis-tent sialylated species [69] Modifications to your bioprocess can reduce protein loss by limiting these mechanisms Adjusting your medium formulation, operating parameters, or harvest procedure can significantly increase your product yield [106]

3 Summary

Determining the source of low protein production can be lenging given the large number of potential causes Successful troubleshooting of low productivity requires a fundamental understanding of the protein as well as a methodical approach to investigating the inhibiting factors In the end, there are three fun-damental approaches to maximizing a difficult to express protein: host cell engineering, improved vector design, and the optimiza-tion of the cell culture process This chapter has covered many, but not all of the currently known methods to improve protein expres-sion in CHO cells A single solution may result in improved expres-sion or alternatively, a protein may require a synergistic approach where multiple strategies are combined to ultimately increase pro-ductivity In either case, the solution may be both cell line and protein specific, thus requiring a well-designed set of experiments

chal-to discover and relieve the bottleneck

References

1 Ohya T, Hayashi T, Kiyama E et al (2008)

Improved production of recombinant human

antithrombin III in Chinese hamster ovary cells

by ATF4 overexpression Biotechnol Bioeng

100:317–324 doi: 10.1002/bit.21758

2 Novo JB, Morganti L, Moro AM et al (2012)

Generation of a Chinese hamster ovary cell line

producing recombinant human

glucocerebrosi-dase J Biomed Biotechnol 2012:875383

doi: 10.1155/2012/875383

3 Le Fourn V, Girod P-A, Buceta M et al (2014)

CHO cell engineering to prevent polypeptide

aggregation and improve therapeutic protein

secretion Metab Eng 21:91–102

doi: 10.1016/j.ymben.2012.12.003

4 Jiang Z, Huang Y, Sharfstein ST (2006)

Regulation of recombinant monoclonal

anti-body production in chinese hamster ovary

cells: a comparative study of gene copy

num-ber, mRNA level, and protein expression

Biotechnol Prog 22:313–318

5 Nishimiya D, Mano T, Miyadai K et al

(2013) Overexpression of CHOP alone and

in combination with chaperones is effective

in improving antibody production in

mam-malian cells Appl Microbiol Biotechnol

97:2531–2539 doi: 10.1007/s00253-012-

4365-9

6 Mead EJ, Chiverton LM, Smales CM, von

der Haar T (2009) Identification of the

limi-tations on recombinant gene expression in

CHO cell lines with varying luciferase

pro-duction rates Biotechnol Bioeng 102:1593–

1602 doi: 10.1002/bit.22201

7 Hu Z, Guo D, Yip SSM et al (2013) Chinese hamster ovary K1 host cell enables stable cell line development for antibody molecules which are difficult to express in DUXB11- derived dihydrofolate reductase deficient host cell Biotechnol Prog 29:980–985 doi: 10.1002/ btpr.1730

8 Alves C, Gilbert A, Dalvi S et al (2015) Integration of cell line and process develop- ment to overcome the challenge of a difficult

to express protein Biotechnol Prog:1–11 doi: 10.1002/btpr.2091

9 Lattenmayer C, Trummer E, Schriebl K et al (2007) Characterisation of recombinant CHO cell lines by investigation of protein productivities and genetic parameters

J Biotechnol 128:716–725 doi: 10.1016/j jbiotec.2006.12.016

10 Reisinger H, Steinfellner W, Stern B et al (2008) The absence of effect of gene copy number and mRNA level on the amount of mAb secretion from mammalian cells Appl Microbiol Biotechnol 81:701–710 doi: 10.1007/s00253-008-1701-1

11 Chusainow J, Yang YS, Yeo JHM et al (2009)

A study of monoclonal antibody-producing CHO cell lines: what makes a stable high pro- ducer? Biotechnol Bioeng 102:1182–1196 doi: 10.1002/bit.22158

12 Hindson BJ, Ness KD, Masquelier D a., et al (2011) High-throughput droplet digital PCR system for absolute quantitation of DNA copy number Anal Chem 83:8604–8610 doi: 10.1021/ac202028g

Christina S Alves and Terrence M Dobrowsky

13 Huang Y, Li Y, Wang YG et al (2007) An

effi-cient and targeted gene integration system for

high-level antibody expression J Immunol

Methods 322:28–39 doi: 10.1016/j.jim.2007

01.022

14 Wright C, Estes SD (2014) Pharmaceutical

next-generation bioprocess: an industry

per-spective of how the ‘ omics era will affect

future biotherapeutic development Pharm

Bioprocess 2:371–375

15 Büssow K (2015) Stable mammalian

pro-ducer cell lines for structural biology Curr

Opin Struct Biol 32:81–90 doi: 10.1016/j.

sbi.2015.03.002

16 Turan S, Galla M, Ernst E et al (2011)

Recombinase-mediated cassette exchange

(RMCE): traditional concepts and current

chal-lenges J Mol Biol 407:193–221 doi: 10.1016/j.

jmb.2011.01.004

17 Baser B, Spehr J, Büssow K, van den Heuvel

J (2015) A method for specifically targeting

two independent genomic integration sites

for co-expression of genes in CHO cells

Methods 95:3–12 doi: 10.1016/j.ymeth

2015.11.022

18 Turan S, Zehe C, Kuehle J et al (2013)

Recombinase-mediated cassette exchange

(RMCE) - a rapidly-expanding toolbox for

targeted genomic modifications Gene 515:1–

27 doi: 10.1016/j.gene.2012.11.016

19 Inao T, Kawabe Y, Yamashiro T et al (2015)

Improved transgene integration into the

Chinese hamster ovary cell genome using the

Cre-loxP system J Biosci Bioeng 120:99–

106 doi: 10.1016/j.jbiosc.2014.11.019

20 Kawabe Y, Shimomura T, Huang S et al

(2016) Targeted transgene insertion into the

CHO cell genome using Cre recombinase-

incorporating integrase-defective retroviral

vectors Biotechnol Bioeng:1600–1610

doi: 10.1002/bit.25923

21 Sakuma T, Takenaga M, Kawabe Y et al

(2015) Homologous recombination-

independent large gene cassette knock-in in

CHO cells using TALEN and MMEJ-directed

donor plasmids Int J Mol Sci 16:23849–

23866 doi: 10.3390/ijms161023849

22 Bachu R, Bergareche I, Chasin L a (2015)

CRISPR-Cas targeted plasmid integration

into mammalian cells via non-homologous

end joining Biotechnol Bioeng 112:2154–

2162 doi: 10.1002/bit.25629

23 Lee JS, Kallehauge TB, Pedersen LE,

Kildegaard HF (2015) Site-specific integration

in CHO cells mediated by CRISPR/Cas9 and

homology-directed DNA repair pathway Sci

Rep 5:8572 doi: 10.1038/srep08572

24 Peterson CL, Laniel M (2004) Histones and

histone modifications Curr Biol 14:546–551

25 Reinhart D, Sommeregger W, Debreczeny M

et al (2013) Characterization of recombinant IgA producing CHO cell lines by qPCR BMC Proc 7:P114 doi: 10.1186/1753-6561-7-S6- P114

26 Schlatter S, Stansfield SH, Dinnis DM et al (2005) On the Optimal Ratio of Heavy to Light Chain Genes for Efficient Recombinant Antibody Production by CHO Cells Biotechnol Prog 21:122–133

27 Ho SCL, Koh EYC, van Beers M et al (2013) Control of IgG LC:HC ratio in stably trans- fected CHO cells and study of the impact on expression, aggregation, glycosylation and con- formational stability J Biotechnol 165:157–

166 doi: 10.1016/j.jbiotec.2013.03.019

28 Osterlehner A, Simmeth S, Göpfert U (2011) Promoter methylation and transgene copy numbers predict unstable protein production

in recombinant Chinese hamster ovary cell lines Biotechnol Bioeng 108:2670–2681 doi: 10.1002/bit.23216

29 Wippermann A, Rupp O, Brinkrolf K et al (2015) The DNA methylation landscape of Chinese hamster ovary (CHO) DP-12 cells

J Biotechnol 199:38–46 doi: 10.1016/j jbiotec.2015.02.014

30 De Carvalho DD, You JS, Jones PA (2011) DNA methylation and cellular reprogram- ming Trends Cell Biol 20:609–617 doi: 10.1016/j.tcb.2010.08.003.DNA

31 Clark SJ, Harrison J, Paul CL et al (1994) High sensitivity mapping of methylated cyto- sines Nucleic Acids Res 22:2990–2997

32 Adorján P, Distler J, Lipscher E et al (2002) Tumour class prediction and discovery by microarray-based DNA methylation analysis Nucleic Acids Res 30:e21 doi: 10.1016/ S0959-8049(01)80570-0

33 Backliwal G, Hildinger M, Kuettel I et al (2008) Valproic acid: a viable alternative to sodium butyrate for enhancing protein expression in mammalian cell cultures Biotechnol Bioeng 101:182–189 doi: 10.1002/bit.21882

34 Kouzarides T (2000) Acetylation: a tory modification to rival phosphorylation? EMBO J 19:1176–1179 doi: 10.1093/ emboj/19.6.1176

35 Jiang Z, Sharfstein ST (2008) Sodium ate stimulates monoclonal antibody over- expression in CHO cells by improving gene accessibility Biotechnol Bioeng 100:189–

butyr-194 doi: 10.1002/bit.21726

36 Kyoung MIN (2007) Correlation between enhancing effect of sodium butyrate on spe- cific productivity and mRNA transcription level in recombinant chinese hamster ovary cells producing antibody J Microbiol Biotechnol 17:1036–1040

Optimizing ‘Difficult to Express’ Proteins

37 Yang WC, Lu J, Nguyen NB et al (2014)

Addition of valproic acid to CHO cell fed- batch

cultures improves monoclonal antibody titers

Mol Biotechnol 56:421–428 doi: 10.1007/

s12033-013-9725-x

38 Kim NS, Lee GM (2002) Inhibition of

sodium butyrate-induced apoptosis in

recom-binant Chinese hamster ovary cells by

consti-tutively expressing antisense RNA of

caspase-3 Biotechnol Bioeng 78:217–228

doi: 10.1002/bit.10191

39 Barnes LM, Dickson AJ (2006) Mammalian

cell factories for efficient and stable protein

expression Curr Opin Biotechnol 17:381–

386 doi: 10.1016/j.copbio.2006.06.005

40 Kwaks THJ, Otte AP (2006) Employing

epi-genetics to augment the expression of

thera-peutic proteins in mammalian cells Trends

Biotechnol 24:137–142 doi: 10.1016/j.

tibtech.2006.01.007

41 Maksimenko O, Gasanov NB, Georgiev P,

Lines TPC (2015) Regulatory elements in

vec-tors for efficient generation of cell Lines

produc-ing target proteins Acta Naturae 7:15–26

42 Harraghy N, Gaussin A, Mermod N (2008)

Sustained transgene expression using MAR

elements Curr Gene Ther 8:353–366

doi: 10.2174/156652308786071032

43 Harraghy N, Buceta M, Regamey A et al

(2012) Using matrix attachment regions to

improve recombinant protein production

Methods Mol Biol 801:93–110

doi: 10.1007/978-1-61779-352-3_7

44 Betts Z, Croxford AS, Dickson AJ (2015)

Evaluating the interaction between UCOE

and DHFR-linked amplification and stability

of recombinant protein expression Biotechnol

Prog 31:1014–1025 doi: 10.1002/btpr.2083

45 Kwaks THJ, Barnett P, Hemrika W et al

(2003) Identification of anti-repressor

ele-ments that confer high and stable protein

pro-duction in mammalian cells Nat Biotechnol

21:553–558 doi: 10.1038/nbt814

46 Otte AP, Kwaks THJ, Van Blokland RJM et al

(2007) Various expression-augmenting DNA

elements benefit from STAR-select, a novel

high stringency selection system for protein

expression Biotechnol Prog 23:801–807

doi: 10.1021/bp070107r

47 Mead EJJ, Smales CMM (2011) mRNA

translation and recombinant gene expression

from mammalian cell expression systems

Compr Biotechnol 1:403–409 doi: 10.1016/

B978-0-08-088504-9.00043-X

48 Mead EJ, Masterton RJ, Feary M et al (2015)

Biological insights into the expression of

translation initiation factors from

recombi-nant CHOK1SV cell lines and their ship to enhanced productivity Biochem

relation-J 472:261–273 doi: 10.1042/BJ20150928

49 Chong WPK, Sim LC, Wong KTK, Yap MGS (2009) Enhanced IFN c production in adenosine- treated CHO cells: a mechanistic study Biotechnol Prog 25:866–873 doi: 10.1021/bp.118

50 Dreesen IA, Fussenegger M (2011) Ectopic expression of human mTOR increases viabil- ity, robustness, cell size, proliferation, and antibody production of chinese hamster ovary cells Biotechnol Bioeng 108:853–866 doi: 10.1002/bit.22990

51 Gustafsson C, Govindarajan S, Minshull

J (2004) Codon bias and heterologous tein expression Trends Biotechnol 22:346–

pro-353 doi: 10.1016/j.tibtech.2004.04.006

52 Welch M, Villalobos a., Gustafsson C, Minshull J (2009) You’re one in a googol: optimizing genes for protein expression J R Soc Interface 6:S467–S476 doi: 10.1098/ rsif.2008.0520.focus

53 Chung BK-S, Yusufi FNK, Mariati, et al (2013) Enhanced expression of codon opti- mized interferon gamma in CHO cells

J Biotechnol 167:326–333 doi: 10.1016/j jbiotec.2013.07.011

54 Kozak M (1991) Structural features in otic mRNAs that modulate the initiation of translation J Biol Chem 266:19867–19870

55 Nishimiya D (2014) Proteins improving recombinant antibody production in mamma- lian cells Appl Microbiol Biotechnol 98:1031–

1042 doi: 10.1007/s00253-013-5427-3

56 Pybus LP, Dean G, West NR et al (2013) Model-directed engineering of “difficult-to- express” monoclonal antibody production by Chinese hamster ovary cells Biotechnol Bioeng xxx:1–14 doi: 10.1002/bit.25116

57 Johari YB, Estes SD, Alves CS et al (2015) Integrated cell and process engineering for improved transient production of a

“difficult- to- express” fusion protein by CHO cells Biotechnol Bioeng 112:2527–

2542 doi: 10.1002/bit.25687

58 Chollet ME, Skarpen E, Iversen N et al (2015) The chemical chaperone sodium 4-phenylbutyrate improves the secretion of the protein CA267T mutant in CHO-K1 cells trough the GRASP55 pathway Cell Biosci 5:57 doi: 10.1186/s13578-015-0048-4

59 De Almeida SF, Picarote G, Fleming JV et al (2007) Chemical chaperones reduce endoplas- mic reticulum stress and prevent mutant HFE aggregate formation J Biol Chem 282:27905–

27912 doi: 10.1074/jbc.M702672200 Christina S Alves and Terrence M Dobrowsky

60 Hwang S-J, Jeon C-J, Cho SM et al (2011)

Effect of chemical chaperone addition on

pro-duction and aggregation of recombinant flag-

tagged COMP-angiopoietin 1 in Chinese

hamster ovary cells Biotechnol Prog 27:587–

591 doi: 10.1002/btpr.579

61 Liu C-H, Chen L-H (2007) Promotion of

recombinant macrophage colony stimulating

factor production by dimethyl sulfoxide

addi-tion in Chinese hamster ovary cells J Biosci

Bioeng 103:45–49 doi: 10.1263/jbb.103.45

62 Liu C-H, Chen L-H (2007) Enhanced

recom-binant M-CSF production in CHO cells by

glycerol addition: model and validation

Cytotechnology 54:89–96 doi: 10.1007/s10616-

007-9078-z

63 Chakravarthi S, Jessop CE, Bulleid NJ (2006)

The role of glutathione in disulphide bond

for-mation and endoplasmic-reticulum- generated

oxidative stress EMBO Rep 7:271–275

doi: 10.1038/sj.embor.7400645

64 Jing Y, Borys M, Nayak S et al (2012)

Identification of cell culture conditions to

con-trol protein aggregation of IgG fusion proteins

expressed in Chinese hamster ovary cells

Process Biochem 47:69–75 doi: 10.1016/j.

procbio.2011.10.009

65 Rezaei M, Zarkesh-Esfahani SH, Gharagozloo

M (2013) The effect of different media

com-position and temperatures on the production

of recombinant human growth hormone by

CHO cells Res Pharm Sci 8:211–217

66 Vergara M, Berrios J, Martínez I et al (2015)

Endoplasmic reticulum-associated rht-PA

pro-cessing in CHO cells: Influence of mild

hypo-thermia and specific growth rates in batch and

Chemostat cultures PLoS One 10:e0144224

doi: 10.1371/journal.pone.0144224

67 Walsh G, Jefferis R (2006) Post-translational

modifications in the context of therapeutic

proteins Nat Biotechnol 24:1241–1252

doi: 10.1038/nbt1252

68 De Leon GM, Wlaschin KF, Nissom PM et al

(2007) Comparative transcriptional analysis

of mouse hybridoma and recombinant

Chinese hamster ovary cells undergoing

butyrate treatment J Biosci Bioeng 103:82–

91 doi: 10.1263/jbb.103.82

69 Sung YH, Song YJ, Lim SW et al (2004)

Effect of sodium butyrate on the production,

heterogeneity and biological activity of human

thrombopoietin by recombinant Chinese

hamster ovary cells J Biotechnol 112:323–

335 doi: 10.1016/j.jbiotec.2004.05.003

70 Kantardjieff A, Jacob NM, Yee JC et al (2010)

Transcriptome and proteome analysis of Chinese

hamster ovary cells under low temperature and

butyrate treatment J Biotechnol 145:143–159 doi: 10.1016/j.jbiotec.2009.09.008

71 Knappskog S, Ravneberg H, Gjerdrum C et al (2007) The level of synthesis and secretion of

Gaussia princeps luciferase in transfected

CHO cells is heavily dependent on the choice

of signal peptide J Biotechnol 128:705–715 doi: 10.1016/j.jbiotec.2006.11.026

72 Hegde RS, Bernstein HD (2006) The ing complexity of signal sequences Trends Biochem Sci 31:563–571 doi: 10.1016/j tibs.2006.08.004

73 Kober L, Zehe C, Bode J (2013) Optimized signal peptides for the development of high expressing CHO cell lines Biotechnol Bioeng 110:1164–1173 doi: 10.1002/bit.24776

74 Haryadi R, Ho S, Kok YJ et al (2015) Optimization of heavy chain and light chain signal peptides for high level expression of therapeutic antibodies in CHO cells PLoS One 10:e0116878 doi: 10.1371/journal pone.0116878

75 Petersen TN, Brunak S, von Heijne G, Nielsen

H (2011) SignalP 4.0: discriminating signal tides from transmembrane regions Nat Methods 8:785–786 doi: 10.1038/nmeth.1701

76 Frank K, Sippl MJ (2008) High-performance signal peptide prediction based on sequence alignment techniques Bioinformatics 24:2172–

2176 doi: 10.1093/bioinformatics/btn422

77 Stoops J, Byrd S, Hasegawa H (2012) Russell body inducing threshold depends on the vari- able domain sequences of individual human IgG clones and the cellular protein homeosta- sis Biochim Biophys Acta 1823:1643–1657 doi: 10.1016/j.bbamcr.2012.06.015

78 Hasegawa H, Woods CE, Kinderman F et al (2014) Russell body phenotype is preferen- tially induced by IgG mAb clones with high intrinsic condensation propensity: relations between the biosynthetic events in the ER and solution behaviors in vitro MAbs 6:1518–1532 doi: 10.4161/mabs.36242

79 Marotta NP, Lin YH, Lewis YE et al (2015) O-GlcNAc modification blocks the aggrega- tion and toxicity of the protein α-synuclein associated with Parkinson’s disease Nat Chem 7:913–920 doi: 10.1038/nchem.2361

80 Peroutka RJ, Elshourbagy N, Piech T, Butt TR (2008) Enhanced protein expression in mam- malian cells using engineered SUMO fusions: secreted phospholipase A2 Protein Sci 17:1586–

1595 doi: 10.1110/ps.035576.108.of

81 Kaufmann H, Fussenegger M (2003) Metabolic engineering of mammalian cells for higher protein yield New Compr Biochem 38:457–469

Optimizing ‘Difficult to Express’ Proteins

83 Gossen M, Bujard H (1992) Tight control of

gene expression in mammalian cells by

tetracycline- responsive promoters Proc Natl

Acad Sci U S A 89:5547–5551 doi: 10.1073/

pnas.89.12.5547

84 Gossen M, Freundlieb S, Bender G et al

(1995) Transcriptional activation by

tetracy-clines in mammalian cells Science 268:1766–

1769 doi: 10.1126/science.7792603

85 Zhang Y, Katakura Y, Ohashi H, Shirahata S

(1997) Efficient and inducible production of

human interleukin 6 in Chinese hamster ovary

cells using a novel expression system

Cyto-technology 25:53–60

86 Fussenegger M, Bailey JE, Hauser H, Mueller

PP (1999) Genetic optimization of

recombi-nant glycoprotein production by mammalian

cells Trends Biotechnol 17:35–42

doi: 10.1016/S0167-7799(98)01248-7

87 Bi JX, Shuttleworth J, Al-Rubeai M (2004)

Uncoupling of cell growth and proliferation

results in enhancement of productivity in

p21CIP1-arrested CHO cells Biotechnol

Bioeng 85:741–749 doi: 10.1002/bit.20025

88 Sinacore MS, Drapeau D, Adamson SR

(2000) Adaptation of mammalian cells to

growth in serum-free media Mol Biotechnol

15:249–257 doi: 10.1385/MB:15:3:249

89 Costa AR, Rodrigues ME, Henriques M et al

(2011) Strategies for adaptation of mAb-

producing CHO cells to serum-free medium

BMC Proc 5(Suppl 8):P112 doi: 10.1186/

1753-6561-5-S8-P112

90 Langer ES (2011) Trends in perfusion

biore-actors Bioprocess Int 9:18–22

91 Li L, Shi M, Song Y et al (2009) A single-use,

scalable perfusion bioreactor system

Bioprocess Int:46–54

92 Kompala DS, Ozturk SS (2006) Optimization

of high cell density perfusion bioreactors Cell

culture technologies for pharmaceutical and

cell-based therapies CRC Press, Taylor and

Francis Group, Boca Raton

93 Kolind MP, Nørby PL, Berchtold MW,

Johnsen LB (2011) Optimisation of the

fac-tor VIII yield in mammalian cell cultures by

reducing the membrane bound fraction

J Biotechnol 151:357–362 doi: 10.1016/j.

jbiotec.2010.12.019

94 Kaufman RJ, Wasley LC, Davies MV et al

(1989) Effect of von Willebrand factor

coex-pression on the synthesis and secretion of factor VIII in Chinese hamster ovary cells Mol Cell Biol 9:1233–1242 doi: 10.1128/ MCB.9.3.1233

95 Spiegel PC, Kaiser SM, Simon JA, Stoddard

BL (2004) Disruption of protein-membrane binding and identification of small-molecule inhibitors of coagulation factor VIII Chem Biol 11:1413–1422 doi: 10.1016/j.chembiol 2004.08.006

96 Trummer E, Fauland K, Seidinger S et al (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 bioreac- tors Biotechnol Bioeng 94:1033–1044 doi: 10.1002/bit

97 Zhu MM, Goyal A, Rank DL et al (2005) Effects of elevated pCO 2 and osmolality on growth of CHO cells and production of antibody- fusion protein B1: a case study Biotechnol Prog 21:70–77

98 Yoon SK, Choi SL, Song JY, Lee GM (2005) Effect of culture pH on erythropoietin pro- duction by Chinese hamster ovary cells grown

in suspension at 32.5 and 37.0 degrees

C Biotechnol Bioeng 89:345–356 doi: 10.1002/bit.20353

99 Mandenius CF, Brundin A (2008) Bioprocess optimization using design-of-experiments methodology Biotechnol Prog 24:1191–

1203 doi: 10.1002/btpr.67

100 Jiang Z, Droms K, Geng Z et al (2012) Fed- batch cell culture process optimization Bioprocess Int 10:40–45

101 Sunley K, Butler M (2010) Strategies for the enhancement of recombinant protein produc- tion from mammalian cells by growth arrest Biotechnol Adv 28:385–394 doi: 10.1016/j biotechadv.2010.02.003

102 Min Lee G, Koo J (2010) Osmolarity effects, Chinese hamster ovary cell culture In: Encyclopedia of Industrial Biotechnology John Wiley & Sons, Inc., pp 1–8

103 Du Z, Treiber D, McCarter JD et al (2015) Use

of a small molecule cell cycle inhibitor to control cell growth and improve specific productivity and product quality of recombinant proteins in CHO cell cultures Biotechnol Bioeng 112:141–155 doi: 10.1002/bit.25332

104 Ganne V, Mignot G (1991) Application of tistical design of experiments to the optimiza- tion of factor VIII expression by CHO cells Cytotechnology 6:233–240 doi: 10.1007/ BF00624762

105 Torkashvand F, Vaziri B, Maleknia S et al (2015) Designed amino acid feed in improvement of Christina S Alves and Terrence M Dobrowsky

23 production and quality targets of a therapeutic

monoclonal antibody PLoS One 10:e0140597

doi: 10.1371/journal.pone.0140597

106 Ellert A, Vikström C (2014) Design of

experi-ments with small-scale bioreactor systems for

efficient bioprocess development and

optimiza-tion Bioprocess Int 12(5):10–13

107 Pertea M, Lin X, Salzberg S (2001)

GeneSplicer: a new computational method

for splice site prediction Nucleic Acids Res

29:1185–1190 doi: 10.1093/nar/29.5.1185

108 Brunak S, Engelbrecht J, Knudsen S (1991)

Prediction of human mRNA donor and

accep-tor sites from the DNA sequence J Mol Biol

220:49–65 doi: 0022-2836(91)90380-O [pii]

109 Solovyev VV, Salamov AA, Lawrence CB (1994) Predicting internal exons by oligonu- cleotide composition and discriminant analy- sis of spliceable open reading frames Nucleic Acids Res 22:5156–5163 doi: 10.1093/ nar/22.24.5156

110 Reese MG, Eeckman FH, Kulp D, Haussler D (1997) Improved splice site detection in genie J Comput Biol 4:311–323 doi: 10.1089/ cmb.1997.4.311

111 Rogozin IB, Milanesi L (1997) Analysis of donor splice sites in different eukaryotic organisms J Mol Evol 45:50–59 doi: 10.1007/PL00006200

Optimizing ‘Difficult to Express’ Proteins

Paula Meleady (ed.), Heterologous Protein Production in CHO Cells: Methods and Protocols, Methods in Molecular Biology,

vol 1603, DOI 10.1007/978-1-4939-6972-2_2, © Springer Science+Business Media LLC 2017

Chapter 2

Glycoengineering of CHO Cells to Improve Product Quality

Qiong Wang, Bojiao Yin, Cheng-Yu Chung, and Michael J Betenbaugh

Abstract

Chinese hamster ovary (CHO) cells represent the predominant platform in biopharmaceutical industry for the production of recombinant biotherapeutic proteins, especially glycoproteins These glycoproteins include oligosaccharide or glycan attachments that represent one of the principal components dictating product quality Especially important are the N-glycan attachments present on many recombinant glyco- proteins of commercial interest Furthermore, altering the glycan composition can be used to modulate the production quality of a recombinant biotherapeutic from CHO and other mammalian hosts This review first describes the glycosylation network in mammalian cells and compares the glycosylation pat- terns between CHO and human cells Next genetic strategies used in CHO cells to modulate the sialylation patterns through overexpression of sialyltransfereases and other glycosyltransferases are summarized In addition, other approaches to alter sialylation including manipulation of sialic acid biosynthetic pathways and inhibition of sialidases are described Finally, this review also covers other strategies such as the glyco- sylation site insertion and manipulation of glycan heterogeneity to produce desired glycoforms for diverse biotechnology applications.

Key words Chinese hamster ovary (CHO), N-linked glycosylation, Glycoengineering, Sialylation,

Glycosylation site insertion, Heterogeneity

1 Introduction

Therapeutic glycoproteins represent a rapidly growing segment of the biopharmaceutical industry with total sales of many tens of bil-lion dollars annually [1] These products include several protein classes such as enzymes, hormones, cytokines, growth factors, clot-ting factors, as well as monoclonal antibodies and Ig-Fc-Fusion proteins [2–4] The increasing demand of biotherapeutics for the treatments of diseases, such as cancer, immune disorders, infec-tious diseases, genetic disorders, and ailments such as Alzheimer’s and Parkinson’s, are the main drivers for the development of gly-coprotein therapeutics [1]

Glycosylation is a critical posttranslational modification found

on most of these biotherapeutics What is unique about ation compared to other posttranslational processing events is the

structural variety and functional diversity present, in which the cosylation can vary widely even for a single protein and also from organism to organism Glycosylation characteristics can play a major role in modulating a protein’s stability, folding, targeting/traffick-ing, immunogenicity, biological activity, and especially circulatory half-life [5] Oligosaccharides are attached cotranslationally through glycosidic linkages on specific asparagine (N-linked) or serine/thre-onine (O-linked) residues While N-glycans are the most common modification on biotherapeutics including monoclonal antibodies and will be the focus of the current review, several therapeutic gly-coproteins such as erythropoietin (EPO) and etanercept (Enbrel) also include O-glycan modifications [6] N-glycans are linked to the Asn of the Asn-X-Ser/Thr consensus sequence in which X denotes any amino acid except proline [7] A consensus sequence for O-linked glycosylation has yet to be identified [5] Given its non-template-driven nature, heterogeneity of glycosylation arises both from variations in glycosylation site occupancy and in the diversity

gly-of final glycan structures attached to glycoproteins emerging from the cellular secretory compartments As a result of the stochastic nature of interactions between enzymes and oligosaccharide sub-strates and the variety of enzymes that can act on any one glycan substrate, a wide range of different glycans are generated for most proteins as these polypeptides traverse through the endoplasmic reticulum (ER) and various Golgi compartments [8 9]

In particular, the N-linked glycosylation pathway in mammalian cells involves a highly complex and interconnected reaction network catalyzed by glycosidases and glycosyltransferases contained within different compartments of the ER and Golgi apparatus, depicted in the schematic of Fig 1 The biosynthesis of mammalian N-glycans initiates at the cytoplasmic face of the ER membrane with the transfer

of GlcNAc-P from UDP-GlcNAc to the dolichol phosphate (Dol-P) lipid carrier to generate dolichol pyrophosphate N-acetylglucosamine (Dol-P-P-GlcNAc) [10] Then 14 sugars are sequentially added to Dol-P-P-GlcNAc to form an oligosaccharide precursor (Glc3Man9

GlcNAc2) [10] Next, oligosaccharyltransferase (OST) selects Ser/Thr sequons in a nascent polypeptide and proceeds with an en bloc transfer of Glc3Man9GlcNAc2 to the side chain amide of aspara-gine and releasing Dol-P-P in the process [11] The glucose residues

Asn-X-on the precursor are sequentially trimmed by ER α-glucosidase I and

II to form monoglucosylated glycan, which is a key intermediate in the ER lectin chaperones calnexin/calreticulin-associated glycopro-tein folding control cycle [12] Once correctly folded, the precursor

is trimmed by ER α-mannosidase I to yield Man8GlcNAc2-protein before exiting ER After translocation into the cis-Golgi, the Man8GlcNAc2 glycoform is further trimmed by Golgi α-mannosidases

I to give Man5GlcNAc2, a key intermediate along the pathway to form hybrid and complex N-glycans and sometimes found as a final glycan product

Qiong Wang et al.

Fig 1 Schematic of N-glycosylation biosynthesis pathway in CHO cell

CHO Glycoengineering

As shown in Fig.1, biosynthesis of hybrid and complex N-glycans begins in the medial-Golgi by the action of an

N-acetylglucosaminyltransferase (GnT-1 or Mgat1), which adds a

GlcNAc to Man5GlcNAc2 [10] Then the majority of N-glycans are trimmed by Golgi α-mannosidase II removing two mannoses from GlcNAcMan5GlcNAc2 to yield GlcNAcMan3GlcNAc2 Hybrid N-glycans result when a structure such as GlcNAcMan3GlcNAc2 either undergoes no further extension or trimming to remove exposed mannose residues resulting in structures with one

or two terminal Man residue In addition, sometimes another GlcNac can be added to the innermost Man group by the enzyme

β1,4-N-acetylglucosaminyltransferase III (GnT-III or Mgat3) in

the medial Golgi, resulting in bisecting GlcNAc structures that can also alter the capacity for other downstream enzymes to act on the glycan structure

Next, the enzyme β-1,2-N-acetylglucosaminyltransferase II

(GnT-II or Mgat2) adds a GlcNAc to the GlcNAcMan3GlcNAc2

structure to generate the glycan product GlcNAc2Man3GlcNAc2, which is the precursor for all multiantennary complex N-glycans Tri-and tetra-antennary branches can be achieved by adding GlcNAc at

α(1,3)-mannose site by N-acetylglucosaminyltransferase IV (GnT-IV

or Mgat 4) and at α(1,6)-mannose site by N-acetylglucosaminyltransferase

V (GnT-V or Mgat 5) Additional modifications of complex and hybrid N-glycans can occur in the trans-Golgi and include the addi-tion of core α(1,6)-fucose to the GlcNAc adjacent to Asn at the N-glycan sites by α-(1,6)-fucosyltransferase and the branch elongation

by the addition of a β-linked galactose residue to GlcNAc by syltransferase to produce Galβ1-4GlcNAc, or poly-acetyl-lactosamine (poly- LacNAc) sequences Finally, these terminal Gal residues can serve as acceptors for several sialyltransferases, leading to even more complex structures [10]

galacto-Chinese hamster ovary (CHO) cells are widely used for duction of many commercial and clinical biopharmaceuticals due

pro-to their capacity pro-to produce glycoforms that are, with exceptions, accepted by the human immune system [2 13] Alternative mam-malian cell lines also used in the production of biopharmaceuticals include baby hamster kidney (BHK21), murine myeloma and hybridoma cell lines (NS0 and Sp2/0), and, to a lesser extent, human host cell lines such as human embryonic kidney (HEK293) and human retinal cells (PER C6) [2 3 14]

Two nonhuman glycans—terminal Galα1,3-Gal linkages (alpha-Gal) and N-glycolylneuraminic acid (Neu5Gc) residues—exist in nonhuman mammalian cells and could elicit adverse immunogenic reactions in humans [2 15] Mouse cells have an α1,3-galactosyltransferase enzyme that produces glycans contain-ing the alpha-Gal linkage [16] The second potential immunogenic epitope Neu5Gc is common in all non-primate mammalian cells [2] due to the presence of the enzyme, N-acetylneuraminic acid

Qiong Wang et al.

hydroxylase, which coverts CMP-Neu5Ac to CMP-Neu5Gc in all mammals other than old-world primates [17] Furthermore, the presence of a circulating polyclonal anti-Neu5Gc antibody response has been detected in humans [2 15] In contrast to the alpha-Gal epitope, Neu5Gc can even be taken up from the media as a metab-olite by all mammalian cells, including human cells, and then meta-bolically incorporated onto cell surface glycoconjugates, While all mammalian cells have the potential for immunogenic epitopes, mouse myeloma cells (NS0 and Sp2/0) tend to express higher levels of both of these epitopes compared to hamster (CHO and BHK), making recombinant products from murine cells a higher likelihood for being immunogenic in humans This potential immunogenicity can be especially concerning when the therapeu-tic glycoproteins are administered repeatedly in large doses for chronic diseases [17–19]

Even without these two nonhuman immunogenic epitopes, the glycosylation patterns of proteins expressed in CHO and human cell lines are likely to differ [20] CHO cells typically do

not express N-acetylglucosaminyltransferase III (GnT-III) and

thus lack bisecting GlcNAc residues in their glycoforms, which can impact antibody efficacy [21] Human cells contain GnT III and can produce glycans with bisecting GlcNAc, while antibodies pro-duced in mouse myeloma cells also contain a fraction of glycans with bisecting GlcNAc residues [22]

The glycosylation of biotherapeutics has been identified as a critical quality attribute [23] because each biotherapeutic requires defining glycosylation characteristics to maintain consistent quality parameters such as solubility, thermal stability, protease resistance [24], aggregation [2 3], serum half-life [25], immunogenicity [5], and efficacy [26] Thus, in order to tailor the glycosylation structures produced by CHO cells, a number of researchers have undertaken metabolic glycoengineering strategies to alter the final glycan profiles and distribution in CHO In this review, we will document recombinant protein N-linked glycoengineering studies

in CHO cells and evaluate the impact on the N-glycosylation terns attached to proteins used across the biotechnology industry Given the diversity of structures possible, this review will focus on glycoengineering primarily for non-antibody motifs and briefly dis-cuss the glycoengineering approaches in antibodies

pat-2 Glycoengineering Strategies in CHO Cells

Among the numerous sugar moieties found in glycans, the terminal sialic acid (Neu5Ac) is considered particularly important for the lifespan of glycosylated proteins As an electro-negatively charged acidic 9-carbon moiety, sialic acid is α-glycosidically linked on the C3- or the C6-hydroxyl group of the terminal galactose in humans,

2.1 Sialylation

CHO Glycoengineering

through the action of α2,3-sialyltransferases (ST3) or the sialyltransferase-1 (ST6) [27–29] Terminal sialic acid residues can alter protein properties including biological activity and in vivo cir-culatory half-life Serving as a biological mask, the distal sialic acid can shield galactose residues that when exposed prompt a fast removal of the protein from blood circulation due to the endocyto-sis-mediated uptake by asialoglycoprotein receptors on hepatocytes [29–31] Therefore, in mammalian cells, it is generally desirable to maximize the distal sialic acid content of a glycoprotein to ensure its quality and consistency as an effective therapeutic [12]

α2,6-However, the sialic acid content of glycoproteins expressed in CHO cells can sometimes be incomplete, which is due to two oppos-ing cellular processes The first process consists of two steps—the biosynthesis of cytidine monophospho-sialic acid (CMP-SA) sub-strate and the transfer of sialic acid from this substrate onto a glycan catalyzed by a sialyltransferase The second process is the extracellu-lar removal of sialic acid by sialidase cleavage [32] Both these path-ways are targets for genetic engineering Hence, in the next sections,

we discuss genetic manipulation of the sialylation process, and divide

it into three parts: genetic engineering of sialylation pathways,

over-expression of N-acetylglucosaminyltransferase (GnT) genes, and

inhibition of sialidase activity and present a table to summarize the achievements of glycoengineering to improve protein sialylation (Table 1)

Genetic engineering of sialyltransferase enzymes is probably the most straightforward method to alter sialylation content in terms

of modifying the oligosaccharide biosynthesis reaction networks Sialyltransferases are ultimately responsible for introducing a Neu5Ac residue to the penultimate galactose residue

Six β-galactoside α2,3-sialyltransferases (ST3GAL1–6) and two β-galactoside α2,6-sialyltransferases (ST6GAL1–2) are responsible for forming these terminal sialic acids in mammalian cells Human glyco-proteins bear sialic acid residues in both α2,3- and α2,6-linkages, whereas only α2,3-terminal sialic acids are found in glycoproteins from CHO and BHK cells A report from our group revealed that three genes from the α2,3-sialyltransferase family (ST3GAL3, ST3GAL4, and ST3GAL6) are responsible for α2,3- sialylation in CHO cells using siRNA knockdown approaches, among which ST3GAL4 plays the critical role in dictating glycoprotein α2,3-sialylation [33] ST6GAL1 appears to prefer the Galβ1-4GlcNAc disaccharide sequence linked to

a protein, whereas ST6GAL2 shows a preference for free disaccharide Galβ1-4GlcNAc substrate in humans [34]

The overexpression of heterologous α2,6-sialyltransferase with or without recombinant α2,3-sialyltransferase serves to introduce linkages similar to those found in human cells and has been adapted to elevate the amounts of sialic acid on recombinant proteins [29] Since the first introduction of ST6GAL1 in CHO cells

Synthesize the CMP-sialic acid in the nucleus

UDP-GlcNAc 2-epimerase/ ManNAc kinase

Epimerization of GlcNAc to MAnNAc/ phosphor