Custom CRISPR-Cas9 products

CRISPR-Cas9 for gene editing

• of ADP Adiponectin Cardiovascular Series Human HEK293 Available Quality Control Materials, Calibration Materials, Immunization, etc.
• GDF15 Growth Differentiation Factor 15 Cardiovascular Series Human HEK293
• FLuA NP Influenza A-N protein Infectious Disease Series FLuA E.coli
• Available Quality control materials, calibrators, immunization

• Transfect cells with Karlan CRISPR plasmids with Cas9 and sgRNA for human, mouse, and rat.
• Karlan database of more than 65,000 human, mouse, and rat genes for genome editing using CRISPR. 

1.  FLuB NP B flow-N protein Infectious disease series FLuB E.coli Available quality control materials, calibrators, immunization, etc.
2.  TM Thrombomodulin Thrombus Human HEK293 Available Quality Control Materials, Calibration Materials, Immunization, etc.
3. proBNP B-type brain natriuretic peptide precursor cardiovascular series Human E.coli
4. Tissue-type plasminogen activator Tissue-type plasminogen activator Thrombus Human HEK293 Available Quality controls, calibrators, immunization, etc.
5. Plasminogen activator inhibitor 1 Plasminogen Activator Inhibitor-1 Thrombosis Human HEK293 Unavailable Quality Control Materials, Calibration Materials, Immunization, etc.
6. PCT Procalcitonin Inflammation Human E.coli Available Quality control materials, calibrators, immunization, etc.
7. Tau-441 Tau-441 protein Alzheimer’s Human E.coli
8.  cTnI/C complex Cardiac troponin I-C complex Cardiovascular series Human E.coli Available Quality control, calibrator, immune, etc.
9.  Plasminogen activator inhibitor 1 Plasminogen activator inhibitor-1 Thrombosis Human E.coli Available Quality control materials, calibrators, immunization, etc.
10. Plasminogen activator inhibitor 2 Plasminogen activator inhibitor-1 Thrombus Human E.coli Available Quality control materials, calibrators, immunization, etc.
11. HRSV Post-fusion glycoprotein F0 Respiratory syncytial virus post-fusion F0 protein Infectious disease series Human HEK293
12. HRSV prefusion glycoprotein F0 respiratory syncytial virus prefusion F0 protein
13. CRISPR Lentivirus

1. Genome integration of CRISPR elements using lentivirus. Cas9 and/or sgRNA packed in purified lentiviral particles at 10⁸TU/ml, ready to infect all cell types.

• CRISPR AAV

Episomal expression of CRISPR components with adeno-associated viral particles carrying Cas9 and/or sgRNA, excellent for tissue and animal transduction.

• Cas9 Stable Cell Lines

Premade Cas9-expressing stable cell lines are great for sgRNA library screening and other high-throughput CRISPR-Cas9 applications.

Coupling CRISPR interference with FACS enrichment: New
approach in glycoengineering of CHO cell lines for therapeutic
glycoprotein production
Katja Glinšek1 Lovro Kramer2 Aleksander Krajnc2 Eva Kranjc1 Nina Pirher2
Jaka Marušicˇ2 Leon Hellmann3 Barbara Podobnik2 Borut Štrukelj1
David Ausländer3 Rok Gaber2
1Faculty of Pharmacy, University of Ljubljana, Ljubljana, Slovenia
2Novartis Technical Research & Development, Biologics Technical Development, Lek Pharmaceuticals d.d., Mengeš, Slovenia
3Novartis Institutes for Biomedical Research, Basel, Switzerland
Correspondence
Borut Štrukelj, Faculty of Pharmacy, University
of Ljubljana, Aškerceva 7, SI-1000 Ljubljana, ˇ
Slovenia.
Email: borut.strukelj@ffa.uni-lj.si
David Ausländer, Novartis Institutes for
Biomedical Research, Klybeckstrasse 141,
CH-4057 Basel, Switzerland.
Email: david.auslaender@novartis.com
Rok Gaber, Novartis Technical Research &
Development, Biologics Technical
Development, Lek Pharmaceuticals d.d.,
Kolodvorska 27, SI-1234, Mengeš, Slovenia.
Email: rok.gaber@novartis.com
Funding information
Slovenian Ministry of Education Science and
Sport, Grant/Award Number:
OP20.04327(C3330-19-952017); EU -
European Regional Development Fund,
Grant/Award Number:
OP20.04327(C3330-19-952017)
Abstract
Difficulties in obtaining and maintaining the desired level of the critical quality
attributes (CQAs) of therapeutic proteins as well as the pace of the development are
major challenges of current biopharmaceutical development. Therapeutic proteins,
both innovative and biosimilars, are mostly glycosylated. Glycans directly influence
the stability, potency, plasma half-life, immunogenicity, and effector functions of
the therapeutic. Hence, glycosylation is widely recognized as a process-dependent
CQA of therapeutic glycoproteins. Due to the typically high heterogeneity of glycoforms attached to the proteins, control of glycosylation represents one of the most
challenging aspects of biopharmaceutical development. Here, we explored a new glycoengineering approach in therapeutic glycoproteins development, which enabled us
to achieve the targeted glycoprofile of the Fc-fusion protein in a fast manner. Coupling
CRISPRi technology with lectin-FACS sorting enabled downregulation of the endogenous gene involved in fucosylation and further enrichment of CHO cells producing
Fc-fusion proteins with reduced fucosylation levels. Enrichment of cells with targeted
glycoprofile can lead to time-optimized clone screening and speed up cell line development. Moreover, the presented approach allows isolation of clones with varying levels
of fucosylation, which makes it applicable to a broad range of glycoproteins differing in
target fucosylation level.
Abbreviations: BFP, blue fluorescent protein; BLI, bio-layer interferometry; CHO, Chinese hamster ovary cells; CQA, critical quality attribute; CRISPRi, CRISPR interference; dCas9, catalytically
inactive (“dead”) Cas9; FACS, fluorescence-activated cell sorting; FI, fluorescence intensity; Fut8, alpha-(1,6)-fucosyltransferase; KRAB, Krüppel associated box; LC-MS/MS, liquid
chromatography-tandem mass spectrometry; MFI, mean fluorescence intensity; PTM, post-translational modification; SadCas9, catalytically inactive (“dead”) Cas9 derived from Staphylococcus
aureus; SEM, standard error of mean; TSS, transcription start site.
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any
medium, provided the original work is properly cited and is not used for commercial purposes.
© 2022 The Authors. Biotechnology Journal published by Wiley-VCH GmbH.
Biotechnol. J. 2022;17:2100499. www.biotechnology-journal.com 1 of 13
https://doi.org/10.1002/biot.202100499
18607314, 2022, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/biot.202100499 by Health Research Board, Wiley Online Library on [24/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2 of 13 GLINŠEK ET AL.
KEYWORDS
cell line development, CHO cells, fucosylation, glycoengineering, therapeutic glycoproteins
1 INTRODUCTION
Therapeutic proteins represent one of the most promising types
of drugs in pharmaceutical development, displaying therapeutically
favorable properties, such as high target specificity, potency, and fewer
side effects.[1,2 ] The survey of the biopharmaceutical market approvals
between 2014 and 2018, showed the predominance of protein-based
therapeutics among biopharmaceutical approvals, which is likely to
remain the reality in the near future.[3] The same survey period
also witnessed the rise of biosimilar approvals. Biosimilars are highly
similar to the already marketed reference biologic approved by regulatory authorities. To gain approval, a proposed biosimilar needs
to demonstrate no clinically meaningful differences in terms of its
biological activity, safety, and efficacy when compared to the reference product.[4–6 ] As generic alternatives, biosimilars can be more
affordable compared to the original biologics. Biosimilar market entry
promotes competition, which lowers overall healthcare costs and
makes these otherwise costly types of drugs more accessible.[7] The
rising number of biosimilar approvals in the EU and US drives the competition between developers of innovative biologics and biosimilars.[3]
To stay competitive, both sites are constantly seeking approaches that
enable faster and more cost-effective drug development.
The majority of protein-based biopharmaceuticals, both innovative
and biosimilars, possess some form of post-translational modification
(PTM), among which glycosylation is the most common and complex
one.[8,9 ] Glycoforms directly influence the stability, potency, plasma
half-life, and immunogenicity of therapeutic proteins. Unlike protein
synthesis, protein glycosylation is not template driven and thus significant heterogeneity can arise from the choice of a host cell line and
the bioprocess set-up.[10 ] Hence, glycoproteins, including biopharmaceuticals, always carry a heterogeneous set of glycans.[11 ] Due to the
significant impact on the safety and efficacy of therapeutic proteins,
glycosylation is often considered a critical quality attribute (CQA) and
it needs to be systematically analyzed and controlled throughout the
entire drug development.[2,11 ] If maintaining the consistency of glycoforms from batch to batch is the major challenge for the developers of
innovative biologics, then demonstrating the similarity in glycosylation
is probably one of the most challenging aspects in the development of
biosimilars. Glycan modifications can cause conformational changes of
therapeutic proteins and through this modulate their effector function,
such as antibody-dependent cell-mediated cytotoxicity (ADCC) and
complement-dependent cytotoxicity (CDC).[12,13 ] Specifically, removal
of fucose from the glycosylated Fc domain of therapeutic monoclonal
antibodies (mAbs) results in enhanced ADCC activity.[14–16 ] Moreover,
it was shown that enhancement of Fc-dependent cellular cytotoxicity
by fucose removal is effective not only in the case of mAbs but also in
Fc-fusion proteins.[17 ] When a mode of action of therapeutic protein is
based on ADCC, fucosylation is chosen as CQA.[13 ]
Chinese hamster ovary (CHO) cells are the most commonly used
cell lines for manufacturing biopharmaceuticals.[3,10 ] In the past two
decades, a variety of different genetic modifications of CHO cell lines
were undertaken to achieve desired glycan profiles on protein therapeutics. Many studies have been focusing on designing cell lines
producing nonfucosylated antibodies. Different approaches were used
to achieve reduced levels of fucosylation or complete lack of fucosylation, from homologous recombination-mediated knock-out of α1,6-
fucosyltransferase (Fut8) gene,[18 ] small interfering RNAs (siRNAs)
for double knock-down of Fut8 and GDP-mannose 4,6-dehydratase
(Gmd) genes,[19 ] to zinc-finger nucleases (ZNFs) for deletion of the
Fut8 gene.[20 ] Secretion of almost completely nonfucosylated antibodies from CHO cells was also achieved by heterologous expression
of prokaryotic enzyme GDP-6-deoxy-D-lyxo-4-hexulose reductase
(RMD).[21 ] Besides, knockout of GDP-fucose transporter by using
different types of genome editing techniques was reported as an
alternative approach for fucose-free protein production.[22 ] The development of CRISPR/Cas technology has enabled a more efficient and
user-friendly genome editing tool that applies also to cell line development. Most of the studies used an active variation of Cas9 for Fut8 gene
knockout.[23–26 ]
Numerous studies of genetic manipulation of fucosylation offer
solutions for purposes, where fully nonfucosylated therapeutic proteins are desired. These only partially address challenges when the
target range of fucosylation level is predefined (e.g., in the development of biosimilars). Fucosylation levels can vary from 0% to more than
90%, in different therapeutic glycoproteins.[27–29 ] Currently, extensive
clone screening is necessary to find the final clone with fucosylation
levels in a predefined target range of the reference product. Thus,
there is an urgency to develop approaches for optimization of timeconsuming clone screening and speeding up cell line development of
therapeutic proteins with a broad range of fucosylation.
Here, we explored a new approach in CHO cell line glycoengineering
by combining CRISPR interference (CRISPRi)-mediated glycoenzyme
knockdown with subsequent FACS-mediated, lectin-based selection
(lectin-FACS) of desired cellular phenotypes. We evaluated CRISPRi
for modification of the expression of the endogenous gene involved
in the fucosylation pathway, which led to a significant decrease in
fucosylation on the therapeutic protein. In this exploratory study,
the transcription repression domain KRAB was fused to catalytically
inactive Staphylococcus aureus Cas9 (KRAB-SadCas9) for modifying
endogenous gene expression in CHO cells. To the best of our knowledge, this is the first application of KRAB-SadCas9 originating from
S. aureus in CHO cells. The model protein used in the presented
study is an Fc-fusion protein with multiple N-glycosylation sites on
the target-binding domain and an additional one on the Fc domain.
Repression of Fut8 gene expression in enriched clones resulted in significant downregulation of fucosylation on both target-binding and Fc
18607314, 2022, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/biot.202100499 by Health Research Board, Wiley Online Library on [24/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
GLINŠEK ET AL. 3 of 13
domain. Moreover, a distribution in fucosylation level obtained on isolated clones enabled »fine-tuning« of the desired glycoprofile. Taking
advantage of the specific fucose-binding properties of the Aleura aurantia lectin,[22 ] we could fluorescently stain glycosylated cell surface
proteins of engineered cells and use FACS to enrich a cell population exhibiting lower fucose levels corresponding to the target cell
phenotype. Finally, we believe that the presented glycoengineering
approach might reduce extensive clone screening and speed up cell line
development.
2 MATERIALS AND METHODS
2.1 Plasmid constructs
Transcriptional repressor KRAB-SadCas9-TagBFP was designed based
on previously published work[30 ] but some protein domains were
exchanged and their position swapped. In our design, KRAB repressor domain was placed to the N-terminus of the synthetic repressor which was then followed by GS linker-SV40NLS-Sa-dCas9-
nucleoplasminNLS fusion to which we added c-myc tag and another
SV40-NLS (GGSEQKLISEEDLPKKKRKVGSGSN) followed by FurinT2A self-cleaving peptide sequence[31 ] all flanked with GS linkers of
various lengths. TagBFP was used as an expression marker and was
placed immediately downstream of the splicing signal. The designed
amino acid sequence was reverse translated and codon-optimized for
expression in CHO cell lines. Total synthesis and cloning of synthetic
DNA into expression vector pcDNA3.1(+) Hygro was performed by
Geneart.
For sgRNA expression plasmid, a previously published optimized
S. aureus sgRNA scaffold expressed from human U6 promoter was
used.[32 ] Total synthesis and cloning of sgRNA expression cassette in
mammalian expression plasmid backbone with a puromycin resistance
was performed by Twist Bioscience. TSS of Fut8 was determined using
an internal RNAseq database. The standalone version of CCTop[33 ]
was used to search for patterns of (G)N (19) NNGRRT in the region
– 50 to + 300 bp relative to the TSS, described as the optimal region
for CRISPRi.[34 ] Several sgRNAs were evaluated and the guides targeting +60 bp and +250 bp relative to the identified TSS were used
in the present study (spacer sequences in Supplementary Table S1;
evaluation of individual sgRNAs in Figure S1). sgRNA expression plasmid bearing sequence that targets site absent from CHO genome
(nontargeting sgRNA) was used as a negative control in the present
study.
Biosimilar recombinant protein-expressing plasmid coding for the
Fc-fusion protein gene, neomycin resistance, and recombinant DHFR
gene constructed in Geneart (Life Technologies) was used in this study.
2.2 Generation of stable CRISPRi CHO cells
CHO-K1-derived proprietary parental cells were cultivated in proprietary, chemically defined serum-free medium in shake flasks and
incubated at 37◦C, 10% CO2 at 150 rpm (25 mm shaking amplitude). To
generate the stable dCas9 repressor expressing cell line, parental cells
were transfected with linearized plasmid DNA using Cell line Nucleofection kit V (Lonza) and Nucleofector 2b (Lonza) device, according to
manufacturer’s protocol. Cells were then split and seeded in 24-deep
well plate and incubated at 37◦C, 10% CO2 at 125 rpm (25 mm shaking
amplitude). On day 4, cells were treated with 4 μg mL-1 of hygromycin.
After 3 weeks of selection, single-cell cloning using Single-cell printer
(Cytena) was performed. Only BFP positive pools, measured by flow
cytometry (data not shown), were selected for cloning. Single cells were
seeded in 96-well plates and incubated at 37◦C, 10% CO2. When a
minimum of 60% confluence was achieved, cells were transferred and
expanded in 24-deep well plates. To measure the fluorescence intensity (FI) of BFP, expanded clones were centrifuged for 5 min at 180 × g,
washed twice with PBS, and transferred to a black 96-well plate. FI of
BFP was measured using microplate reader Infinite (Tecan), excited at
402 nm and emission measured at 464 nm. Clones with extremely high
or low FI were excluded from further evaluation due to potential cytotoxic effect[35,36 ] or too low expression, respectively. Of the remaining
clones, six clones with high FI and quantified dCas9 expression values
(Figure S2) were selected for further transfections.
2.3 Generation of stable CHO cells expressing
Fc-fusion protein and sgRNAs targeting Fut8
To establish CHO cells stably expressing Fc-fusion protein and CRISPRi
system, six selected clones stably expressing CRISPRi were pooled and
cotransfected with the plasmid expressing Fc-fusion protein and two
sgRNA plasmids targeting Fut8 at a ratio 2:1:1. For negative control,
nontargeting sgRNA plasmid was cotransfected with Fc-fusion protein
at a ratio of 1:1. In all cases, pooled clones were transfected with linearized plasmid DNA using SF Cell line 4D-Nucleofector X Kit S (Lonza)
and 4D-Nucleofector Core unit together with 4D-Nucleofector X unit
device (Lonza) according to manufacturer’s protocol. After nucleofection, each transfection mixture was split into three separate wells
in a 24-deep well plate, to generate biological triplicates and incubated at 37◦C, 10% CO2 at 125 rpm (25 mm shaking amplitude). For
the selection of pools expressing Fc-fusion protein and sgRNAs, cells
were treated with 400 μg mL-1 G418 and 3 to 5 μg mL-1 puromycin,
sequentially.
2.4 Batch and fed-batch culture
For batch culturing, cells were seeded at 4 × 105 cells mL-1 in 3 mL
proprietary, chemically defined serum-free medium, supplemented
with 400 μg mL-1 hygromycin, 400 μg mL-1 G418, and 5 μg mL-1
puromycin and incubated at 37◦C, 10% CO2 at 125 rpm (25 mm shaking amplitude) for 6 days. On day 6, VCD and viability were determined,
samples were collected for qPCR analysis, cell harvesting was performed and supernatants were used for titer measurements and glycan
analysis.
18607314, 2022, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/biot.202100499 by Health Research Board, Wiley Online Library on [24/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4 of 13 GLINŠEK ET AL.
For fed-batch culturing, cells were seeded at 4 × 105 cells mL-1 in
3 mL proprietary, fed-batch medium, supplemented with 400 μg mL-1
hygromycin, 400 μg mL-1 G418, and 5 μg mL-1 puromycin and incubated at the 10% CO2 at 125 rpm (25 mm shaking amplitude) for 14
days with temperature shift from 37◦C to 33◦C on day 4. Proprietary
Glc and amino acid feed was added on days 4, 6, 7, 10, and 12. On
day 10, samples were taken for VCD, viability, titer measurements, and
qPCR analysis. On day 14, VCD and viability were measured before cell
harvesting. Supernatants were used for titer measurements and glycan
analysis.
2.5 Viable cell density and titer measurements
VCD and viability were measured on the ViCell XR Cell analyzer
(Beckman Coulter). Fc-fusion protein concentrations were determined with bio-layer interferometry using Octet QK (Fortebio). Before
measurements, Protein A biosensors (ForteBio) were rehydrated in
a cell culture medium for 10 min at room temperature. Samples
were diluted and measured for 120 s at 200 rpm. Absolute concentrations of biosimilar Fc-fusion protein were calculated using a
calibration curve prepared by serial dilution of reference Fc-fusion
protein.
2.6 Flow cytometry analysis
To label cells, 2 × 105 cells were centrifuged in a 96-well plate at 4◦C
for 5 min at 180 × g and washed twice with cold PBS. Next, a 1 h incubation in 200 uL PBS supplemented with 14 nM AAL-FITC (Vector Labs)
was performed in dark at 4◦C, followed directly by measuring of cells
without any additional washing steps. FI of FSC/SSC-gated live cells
was measured by Cytoflex (Beckman Coulter Life Sciences) using CytExpert software. A 488 nm laser to excite FITC and a 525/40 filter was
used.
2.7 FACS enrichment and cloning
Cell labeling was performed as described above, with minor modifications. One million cells were centrifuged in a FACS tube at 4◦C for 5 min
at 180 × g and washed twice with cold PBS. Next, a 1 h incubation in
1 mL PBS supplemented with 14 nM AAL-FITC (Vector Labs) was performed in dark at 4◦C, followed directly by measuring of cells without
any additional washing steps.
Cell sorting was performed with FACSAria (Becton Dickinson)
equipped with an automatic cell deposition unit (ACDU) using FACSDiva Software. A 488 nm laser was used to excite FITC and relative FI
of FITC was measured on a detector through a 530/30 filter. First, live
cells were FSC/SSC-gated and only these cells were used for sorting.
The 5% of cells with the lowest FIs were gated and sorted either in bulk
(2 × 104 cells) or as single cells in 96-well plates.
2.8 Total RNA isolation, reverse transcription,
and qPCR analysis
Before RNA isolation, cells were centrifuged at 4◦C for 10 min at 1300
× g and washed twice with cold PBS. RNA was isolated using E-Z 96
Total RNA kit (Omega Biotek) according to the manufacturer’s protocol
with additional steps, including homogenization and DNase I digestion.
Reverse transcription was done using High Capacity cDNA Reverse
Transcription kit (Applied Biosystems) according to the manufacturer’s
protocol for 1 μg of total RNA in 20 μL reaction.
qPCR was run on Lightcycler 480 (Roche) using predesigned
TaqMan Gene Expression Assays (Fut8: Cg04433064_m1, Vezt:
Cg04569010_g1 and Gapdh: Cg04424038_gH; Thermo Fischer
Scientific) together with SsoAdvanced universal probes supermix
(Bio-Rad) according to the manufacturer’s protocol. Three technical
replicates were run for each sample. Expression of Fut8 gene was
normalized to the expression of two housekeeping genes (Vezt and
Gapdh) and relative expression of Fut8 was calculated using the 2
−ΔΔCT method.[37 ]
2.9 Glycan analysis
2.9.1 Lectin based bio-layer interferometry (BLI)
Relative quantification of N-glycans present on the target-binding
domain was performed by lectin-based bio-layer interferometry using
OctetSystems (ForteBio). Before the lectin assay, Fc-fusion protein
concentrations were determined as described above, followed by sample dilutions to reach equal concentrations. First, Protein A biosensors
(ForteBio) were regenerated in 10 mM glycine. Next, Fc-fusion protein samples were loaded onto the Protein A biosensors at 200 rpm
for 120 s. Finally, Fc-fusion protein-loaded Protein A biosensors were
soaked into different lectin solutions, RPL-Fuc1, RPL-Sia1, RPL-Gal1,
RPL-Man2 (GlycoSelect), at 1000 rpm for 60 s, consecutively. Relative
amounts of glycans were calculated by normalization of a lectin binding response in a Fut8 knockdown sample with lectin binding response
detected in a control sample.
2.9.2 Liquid chromatography-tandem mass
spectrometry (LC-MS/MS) peptide mapping
All steps of sample preparation were performed in a fully automated
manner using Tecan Freedom EVO 200 liquid handling system (Tecan
Trading AG, Männedorf, Switzerland) equipped with Hettich Rotanta
460 centrifuge (Andreas Hettich GmbH & Co. KG, Tuttlingen, Germany). 28 μg of protein A-purified protein was denatured in buffer
exchange step using guanidine hydrochloride (Sigma-Aldrich, St. Louis,
MO) denaturing solution (pH 8) and Amicon Ultra-0.5 Centrifugal
Filters (molecular weight cut-off: 10000; Merck KGaA, Darmstadt,
Germany) placed onto a custom-made carrier plate in SBS (Society
18607314, 2022, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/biot.202100499 by Health Research Board, Wiley Online Library on [24/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
GLINŠEK ET AL. 5 of 13
for Biomolecular Screening) format. Reduction of disulfide bonds
was achieved by the addition of dithiothreitol (OneQuant DTT, GBiosciences, Page Avenue St. Louis, MO) and subsequent alkylation of
free thiol groups by iodoacetamide (OneQuant IAM, G-Biosciences).
Alkylation was quenched with DTT. Digestion of reduced and alkylated
protein was performed using MS grade Lysyl Endopeptidase (Lys-C;
FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) at 1:25
enzyme:substrate ratio in proteolysis buffer at pH 8. The digestion was
quenched with LC-MS grade trifluoroacetic acid (TFA; VWR Chemicals,
Radnor, PA).
LC-MS/MS analyses were performed using UHPLC Ultimate 3000
RS (Thermo Fischer Scientific, Waltham, MA) coupled to highresolution mass spectrometer Q Exactive Plus Orbitrap (Thermo
Fischer Scientific). A nominal load of 7.5 μg of the digested sample was
injected onto Ascentis Express Peptide ES-C18, 2.1 × 150 mm, 2.7 μm
column (Merck KGaA). Mobile phase A contained 0.1% TFA in water
and mobile phase B contained 0.1% TFA in acetonitrile. The peptides
were eluted from the column at 58◦C by applying a multistep gradient from 0% to 55% mobile phase B at a flow rate of 0.6 mL min-1.
A washing step was added in each analytical run. The autosampler
was kept at 5◦C. Heated electrospray ionization in positive ion mode
was used for the ionization. The MS capillary and heater temperatures
were maintained at 350◦C and 250◦C, respectively. The full MS spectra acquisition was performed at a resolution of 70, 000 with an AGC
target of 106 and a maximum injection time of 150 ms. For peptide
identification, data-dependent MS/MS was applied.
2.9.3 MS data processing and analysis
For identification and relative quantification of glycans per Nglycosylation site, the raw data files were processed with PepFinder
2.0 software (Thermo Fisher Scientific). For interpretation, obtained
relative abundances of glycans per site were initially normalized by
excluding the relative abundance of the respective nonglycosylated
site. Secondly, normalized relative abundances of glycans per site
were multiplied by the count (number) of monosaccharide units per
glycan and divided by the maximal theoretical count of monosaccharide units, which are achievable in a CHO cell line (5 mannoses,
4 N-acetylglucosamines, 4 galactoses, 4 N-acetylneuraminic acids,
4 N-glycolylneuraminic acids, 1 fucose). Core monosaccharide units
(2 N-acetylglucosamines and 3 mannoses) were not included in the
calculation as they were not targeted with the glycoengineering
experiments and remain constant in the glycan profile. Finally, the
relative abundances of monosaccharide units were summed to obtain
a relative abundance of monosaccharide per site. These are referred to
as glycan attributes: fucosylation, galactosylation, mannosylation and
sialylation.
2.10 Statistical analysis
All statistical analyses were performed in GraphPad Prism 8 software
(GraphPad Prism Software Inc., La Jolla, CA). Statistical significance
was calculated with unpaired t-test, when two groups were compared,
and ordi nary one-way ANOVA with Dunnett’s test for multiple comparisons, where *p < 0.05, **p < 0.01, **p < 0.001, and ****p < 0.0001.
3 RESULTS
3.1 Generation of stable CHO cells expressing
CRISPRi system and Fc-fusion protein
We integrated the CRISPRi system into CHO cells for downregulation
of the endogenous gene involved in fucosylation. We first generated
stable clones expressing the repressor system, based on transcription repression domain KRAB fused to catalytically inactive Cas9
derived from Staphylococcus aureus (KRAB-SadCas9) (Figure 1A). To
avoid a potential undesirable phenotype of only one selected clone, we
pooled six stable CRISPRi clones and performed cotransfection with
sgRNA plasmids and Fc-fusion protein-expressing plasmid. To increase
chances for potent downregulation, we used two sgRNAs, both targeting region around TSS of Fut8 gene.[38,39 ] For negative control,
Fc-fusion protein was cotransfected with nontargeting control sgRNA
plasmid. After the selection process, stable pools were cultivated in
batch and fed-batch mode to explore whether the tested repressor system is functional in both processes. qPCR analysis confirmed
(Figure 1B and E) its functionality in both cases. The significant repression level was achieved in the pool with sgRNAs targeting Fut8 (Fut8
pool) relative to the control, with knockdown efficiency varying from
43% to 65% in batch (Figure 1B), and from 34% to 44% in fed-batch
(Figure 1E), respectively. Using two sgRNAs targeting region around
Fut8 TSS led to similar repression levels as observed in transfections
with individual gRNAs (Figure S1).
To check the effect of downregulation of Fut8 expression on the
fucosylation, we performed N-glycan analysis. Relative quantification of N-glycans attached to the target-binding domain of Fc-fusion
protein was done by lectin-based BLI, which provides an average glycosylation profile of the N-glycosylation sites on the target-binding
domain. BLI represents a high-throughput method for rapid quantification of glycans, and as such enables the analysis of a large number
of samples in a short time. Glycan analysis revealed only a marginal
difference in fucosylation level of the Fut8 pool compared to the control for batch (Figure 1C), and no difference for the fed-batch process
(Figure 1F).
Fc glycans are buried in the protein interior and therefore not
accessible to lectins when the Fc-fusion protein is in its native
state. Therefore, the described lectin-based BLI measurements could
be used to quantify glycans in a target-binding domain solely. To
check the glycosylation profile of the Fc domain, we performed LCMS/MS peptide mapping, which enables site-specific identification and
quantification of N-glycans. One biological replicate of each sample
was analyzed. In contrast to the target-binding domain, downregulation of Fut8 expression led to a pronounced decrease in fucosylation on the Fc domain. Fucose-containing glycans were reduced by
16.3% in the batch (Figure 1D) and 11.6% in the fed-batch process
(Figure 1G).
18607314, 2022, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/biot.202100499 by Health Research Board, Wiley Online Library on [24/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6 of 13 GLINŠEK ET AL.
FIGURE 1 Generation of stable CHO cells expressing CRISPRi system and Fc-fusion protein (A) Repressor system expressing plasmid.
Relative Fut8 gene expression in batch (B) and fed-batch process (E). Glycan analysis of target-binding domain performed by BLI in batch (C) and
fed-batch process (F). Glycan analysis of Fc-domain performed by LC-MS/MS peptide mapping in batch (D) and fed-batch (G). Dots represent
biological replicates. Error bars represent standard error of the mean (SEM) of three biological replicates. Statistical significance between control
pools and Fut8 pools was calculated using the unpaired t-test
3.2 Enrichment of low fucosylated cells
Analysis of Fut8 pools indicated that downregulation of Fut8 expression in pools leads to only a marginal decrease in fucosylation of the
target-binding domain and a more pronounced decrease in Fc fucosylation. We next examined if the pools contain populations of cells
with varying degrees of Fut8 downregulation and if the different levels of Fut8 expression levels lead to a wider range of fucosylation
on the model protein. We addressed this by enriching cells with low
fucosylation using FACS.
Before the enrichment, we stained Fut8 and control pools with
fluorescein-labeled Aleuria Aurantia lectins (AAL-FITC) with specificity
for α-1,6/ α-1,3 linked fucose. In the next step, the low 5% of FITC positive cells of Fut8 and control pools were sorted both as bulk and as
single cells. When the enriched pools and clones were expanded, we
again stained the cells with AAL-FITC and measured the FITC signal
(Figure 2A, Figure S3a).
Comparison of AAL-FITC FIs of enriched pools with the nonenriched pools (Figure 2B) showed a decrease of the mean fluorescence
intensity (MFI) for both the control and the Fut8 pool, demonstrating the lectin sorting itself allows isolation of populations of cells with
reduced fucosylation capacity. However, the relative decrease of cell
surface staining after enrichment was greater in the case of Fut8 pools
(41%) compared to control pools (29%). Furthermore, a clear tail of
cells with low fluorescence staining was observed in the Fut8 enriched
pool (Figure 2A, row 2, right panel), indicating the presence of a population of cells with very low surface fucosylation. Unexpectedly, the
enriched Fut8 clones with the lowest FI (see Figure 2A and suppl
Figure S3a) also exhibited wider histogram peaks compared to CTRL
clones, pointing to phenotype diversity at the time of measurement
in regard to the cell surface fucosylation in the clonally derived cell
population.
Comparison of FI of enriched clones showed that in the group of
clones derived from the control pool (Figure 2A), the MFI of clones
18607314, 2022, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/biot.202100499 by Health Research Board, Wiley Online Library on [24/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
GLINŠEK ET AL. 7 of 13
FIGURE 2 Enrichment of cells with low cell surface fucosylation. (A) AAL-FITC staining of Fut8 and control pools, and enriched pools and
clones. Only one out of three generated control pools and Fut8 pools were stained and used for further enrichment. Clones from each group that
were selected for further evaluation are labeled with an asterisk (see Figure S3 for only selected clones). (B) Quantification of MFI of FITC signal
from a). Dots represent measurements of each stained pool or clone. (C) qPCR quantification of Fut8 gene expression in control and Fut8 pools,
and selected enriched pools and clones cultivated in a 6-day batch. All three nonenriched control and Fut8 pools were cultivated. Dots represent
biological replicates of pools (bar 1,2,4,5) or individual clones (3 and 6). Error bars represent SEM of biological replicates of pools or individual
clones. Statistical significance between control and Fut8 enriched clones was calculated using the unpaired t-test
were in the range of 53% to 96% of the nonenriched control pool
(Figure 2B), whereas in the group of clones derived from the Fut8 pool
(Figure 2A), the MFI of clones were in the range of 15% to 78% of the
nonenriched control pool (Figure 2B), indicating the presence of clones
with substantially lower and more diverse fucosylation levels in the
latter group. Furthermore, on average the MFI of clones derived from
the Fut8 pool was significantly lower compared to the MFI of clones
derived from the control pool. More importantly, only in Fut8 enriched
clones could we find clones with ≥ 2.5-fold reduction of MFI (9 out of
24 number of clones, Figure 2), while in the control enriched clones
none of the clones had > 2.5-fold reduction of MFI (only one clone had
slightly reduced MFI, 1.9-fold reduction).
18607314, 2022, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/biot.202100499 by Health Research Board, Wiley Online Library on [24/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8 of 13 GLINŠEK ET AL.
For the comparison of repression levels that can be achieved by
CRISPRi and subsequent enrichment step and within our standard cell
line development (represented by nonenriched control pool), we further evaluated the Fut8 expression levels in the enriched pools and a
subset of enriched clones. For this purpose, random six clones were
selected from the control group, and six clones with the lowest AALFITC MFI were selected from the Fut8 group. Together with pools,
they were expanded in 6 day batch culture and quantification of Fut8
gene expression was performed by qPCR. As expected, based on surface fluorescence staining, Fut8 expression was downregulated after
the enrichment both in control pools and Fut8 pools as well as clones
(Figure 2C). In enriched control clones, Fut8 expression was in the
range between 70% and 82% compared to nonenriched control pool,
while in Fut8 enriched clones the expression was in the range between
14% and 45% compared to nonenriched control pool.
Collectively, these results demonstrated that the Fut8 pool contained clones with substantially downregulated Fut8 expression that
led to decreased cell surface fucosylation, while such clones were not
present in the control pool. These clones can be efficiently isolated
using AAL-FITC staining with subsequent enrichment of desired cell
populations using FACS. Furthermore, a two-round FACS yielded cells
with even more reduced surface fucosylation levels, demonstrating the
applicability of the method, although these pools were not examined
further in the evaluations (Figure S4). We additionally evaluated the
correlation of Fut8 expression and MFI (Figure S3b) and found a strong
correlation, indicating that repression observed on the mRNA level
results in similar repression on the protein level.Moreover, for the evaluation of the clonal distribution of Fut8 expression in a standard cloning
approach, additional sets of clones from the control pool and Fut8 pool
using standard cell cloning procedure without prior enrichment for low
fucosylation were generated. The average Fut8 expression obtained in
Fut8 clones is significantly lower compared to the expression in control clones (Figure S5), however, the relative change in Fut8 expression
between control and Fut8 clones is smaller in traditionally generated
clones, indicating there are limitations in achievable repression levels
by standard cloning approach.
3.3 Fed-batch of enriched pool and clones
Fluorescence measurements and Fut8 quantification in 6-day batch
showed that we successfully isolated cells with repressed Fut8 and lowered cell surface fucosylation using lectin-FACS enrichment. We next
evaluated enriched pools and six enriched clones originating from both
the Fut8 and control pool in the fed-batch process to determine the
impact of Fut8 gene regulation on the glycan profile of the Fc-fusion
therapeutic.
Quantification of Fut8 gene expression in fed-batch was performed
by qPCR on samples collected on day 10 of the fed-batch, 6 days
after the temperature shift to 33◦C and when the cells are in the
production phase of the bioprocess. The expression of the Fut8 gene
in the enriched Fut8 pool was significantly downregulated for 47%
to 55% relative to the nonenriched control (Figure 3A). No statistically significant change was observed in the enriched control pool.
However, the repression level in the enriched Fut8 pool was only
slightly improved compared to before the enrichment (Figure 1E).
These results are consistent with the fluorescence measurements and
Fut8 gene quantification in batches (Figure 2), where only a minor shift
of the fluorescence signal was observed in the case of the enriched Fut8
pool.
When we quantified the fed-batch expression of the Fut8 gene in
enriched clones originating from the Fut8 pool, significant downregulation of the gene expression was observed in all six enriched clones
(Figure 3A). The Fut8 expression level was downregulated for 41% to
75%, relative to the control. No statistically significant change in the
gene expression was observed in any of the enriched control clones.
This is in contrast with the batch results, where we observed lower Fut8
expression in the clones originating from the control pool (Figure 2C).
Next, we checked how the improved repression levels impacted
the fucosylation of the model protein. The fucosylation of the targetbinding domain was again evaluated by BLI. Average target-binding
fucosylation level of the protein produced by the enriched Fut8 pool
was comparable to the one from the nonenriched and enriched control pool (Figure 3B). However, analysis of the Fc domain fucosylation
level by LC-MS/MS peptide mapping revealed downregulation of fucosylation in the enriched Fut8 pool for 21% compared to the control
(Figure 3C).
In contrast to the enriched pool, the average fucosylation of the
target-binding domain was downregulated in most of the enriched
clones (Figure 3B). Fucosylation was reduced by 11% to 25% in the
clones with statistically significant downregulation (clones FA6, FB5,
FC6, and FD3 in Figure 2A, right panel), relative to the control pool.
In two out of six enriched clones (clones FA3 and FA5 in Figure 2A),
the target-binding fucosylation level was comparable to the level of the
control. Results of target-binding fucosylation obtained by BLI were
confirmed by LC-MS/MS peptide mapping (Figure S6).
The decrease in fucosylation was even more pronounced on the
Fc domain. Fucosylation was evaluated by LC-MS/MS peptide mapping on clones FB5 and FC6. Compared to the control pool it was
downregulated for 57 and 60%, respectively (Figure 3C).
Since we observed slightly weaker Fut8 repression levels in enriched
Fut8 pools and clones on day 10 of fed-batch compared to day 6
of batch, we also analyzed the fucosylation levels of the produced
Fc-fusion protein from batch samples.We indeed observed a more pronounced effect on lowering fucosylation both of target-binding (for up
to 52% in enriched clones) and Fc domain (for up to 89% in enriched
clones) (Figure S7) in batch samples compared to fed-batch, indicating that bioprocess parameters play a significant role in impacting the
Fut8-repressed glycoengineering outcome.
To check if manipulation of fucosylation affected other glycan
attributes, we also analyzed untargeted types of glycosylation (Figure
S8). We did not detect any significant changes on either of the domains
in any untargeted glycosylation type, neither before the enrichment
nor after it. However, higher distribution in galactosylation observed
on both domains is common in cell line development and it is in the
expected range of sample variance.[40 ]
18607314, 2022, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/biot.202100499 by Health Research Board, Wiley Online Library on [24/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
GLINŠEK ET AL. 9 of 13
FIGURE 3 qPCR and glycan analysis of fed batch-produced Fc-fusion therapeutic after the enrichment. 14-day fed-batch was performed with
enriched control and Fut8 pools, together with enriched control and Fut8 clones. Enriched control and Fut8 pool were split into two separate wells
before the fed-batch. (A) qPCR quantification of Fut8 gene expression, (B) glycan analysis of target-binding domain performed by BLI, and (C) Fc
domain performed by LC-MS/MS peptide mapping. Only two clones were selected for LC-MS/MS peptide mapping. Dots represent analyzed
biological replicates of pools or individual clones. Error bars represent SEM of biological replicates of pools or individual clones. Statistical
significance was calculated using ordinary one-way ANOVA with Dunnett’s test for multiple comparisons. For comparison, data from control pools
already shown in Figure 1 was used for calculation and presentation of relative Fut8 expression in enriched pools and clones
FIGURE 4 Cell growth performance. Cell growth performance in fed-batch evaluated by comparison of VCDs (A) and viability during the
process (B), and final titers (C). Error bars represent SEM of biological replicates of pools or individual clones. The line in Chart C represents titer
mean of individual clones. Statistical significance was calculated using ordinary one-way ANOVA with Dunnett’s test for multiple comparisons
We also evaluated how downregulation of Fut8 gene expression
and subsequent enrichment for low fucosylation influenced cell growth
performance. Importantly, repression of Fut8 gene expression and further enrichment did not impair cell growth performance (Figure 4).
However, we observed an unexpected wider distribution of titers in the
CTRL enriched clones that was not related to the decreased levels of
surface fucosylation (see Figure S9).
4 DISCUSSION
Achieving the consistency of the desired glycoprofile, while maintaining the speed of the development processes is a major challenge faced
by manufacturers of innovative therapeutic proteins as well as biosimilars. In the past 30 years, many different cell engineering approaches
have been developed to express therapeutic proteins with desired glycan profiles.[41 ] Regulation of fucosylation has drawn a lot of attention
among researchers since the discovery of its impact on the ADCC
activity.[19,22,42 ] While the majority of the glycoengineering studies
used different types of knockout techniques to remove fucose from glycans, less has been reported on method development for fine-tuning
fucosylation levels. The development of CRISPR/Cas9 enabled a fast
and more efficient genome-editing tool for gene knock-in or knockout.
When an inactivated version of Cas9 (dCas9) is fused to transcriptional effector domains, it serves as a tool for gene activation or gene
silencing, without disruption of the target gene.[43,44 ]
18607314, 2022, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/biot.202100499 by Health Research Board, Wiley Online Library on [24/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
10 of 13 GLINŠEK ET AL.
In this study we evaluated a catalytically inactive form of Cas9,
originating from Staphylococcus aureus fused to transcription repression domain KRAB. With stable integration of the CRISPRi system
targeting two sites upstream of TSS of the Fut8 gene, we were able
to downregulate its expression by an average of ∼ 50% in the pool
of CHO cells (Figure 1B and E). This is in contrast with the previous
report about S. pyogenes-derived dCas9-KRAB based CRISPRi system
not being effective in Fut8 gene repression,[45 ] pointing to a difference in the effectiveness of different CRISPRi systems as well as the
need for screening for the optimal sgRNA. The simultaneous targeting of multiple sgRNAs to the same gene (Figure 1) had a similar
effect as individual transfections (Figure S1) and did not lead to more
potent gene downregulation, as reported previously.[38,43 ] Knockdown
of gene expression by half had only a minor effect (∼12% to 16%
reduction) on Fc fucosylation (Figure 1D and G) and no effect on average target-binding fucosylation (Figure 1C and F). Nonetheless, slight
downregulation of fucosylation on the Fc domain indicated there might
be a population of cells producing low fucosylated Fc-fusion proteins in
the CRISPRi pool. Further AAL-lectin-based FACS enrichment of pools
with Fut8 gene repression indeed isolated cells with low surface fucosylation (Figure 2, Figure S3). Repression levels of enriched clones were
significantly improved, with the strongest downregulation for 75%
(Figure 3A). Effective gene repression led to downregulation of fucosylation on the target-binding domain (Figure 3B), ranging from ∼10%
to 25%. Even more pronounced effect was observed on the Fc domain,
where fucosylation was decreased by up to 60% (Figure 3C). Although
using IgG as a model protein, the authors of a recently published
study reported that knockdown of Fut8 gene expression for ∼ 80%
to 90% led to a decrease in fucosylation for only 30%,[45 ] reaffirming
that type of therapeutic protein as well as other cell line characteristic can significantly contribute to fucosylation variability.[40 ] The
model protein used in the study is Fc-fusion protein consisting of two
structurally and functionally distinct protein domains, with multiple
N-glycosylation sites on the target-binding domain and one on the Fcdomain. Each site is occupied by a different subset of glycans and this
heterogeneity between different N-glycan sites could be the result
of protein conformation influencing glycan accessibility for glycoenzymes. It was confirmed that site-to-site heterogeneity in N-glycan
maturation indeed largely arises from the sterical availability of glycan
residue.[46,47 ] This could also explain the difference in the fucosylation decrease between the Fc domain and the target-binding domain
observed in the case reported here. Thus, the repression of the Fut8
gene does not have the same effect on each protein domain, because
apart from gene expression also other factors, such as protein 3D
structure, could affect the glycoengineering outcome.
Short timelines of cell line development and appropriate product
quality are a top priority for biopharmaceutical developers. One of
the most time-consuming steps in cell line development represents a
selection of the production clone. Thousands of clones are screened to
find the optimal clone, producing sufficient amounts of recombinant
protein.[48 ] Even more clones are screened when particular product quality is desired. Coupling CRISPRi with lectin-FACS enrichment
enables isolation of clones with targeted glycosylation pattern (in our
case low fucosylation) (Figure 2) and with that increases the chance
to find a clone with a desired level of fucosylation. The importance of
enrichment steps for rapid isolation of cells with targeted glycoprofile was reported previously.[22,49,50 ] Unexpectedly, wider histogram
peaks were observed in Fut8 enriched clones with the lowest FI (see
Figure 2A and Figure S3a). While the reason for this is unclear, it suggests potential epigenetic variability around the Fut8 gene in individual
cells expressing KRAB-SadCas9. Although average Fut8 expression in
cell population remains stable through time (Figure S12) it seems that it
can vary in-between individual cells in the population due to unknown
effect on either the CRISPRi efficiency on the Fut8 repression or the
Fut8 repression efficiency on reducing surface fucosylation. Importantly, no untargeted glycan attributes were significantly affected by
Fut8 downregulation and subsequent enrichment (Figure S8). Interestingly, higher VCDs of the Fut8 pool and enriched samples were
observed compared to the control pool (Figure 4). However, higher
VCDs did not affect titers, except in the case of Fut8 enriched clones.
On the other hand, the tested CRISPRi system did not undermine cell
growth as it was reported for CRISPR-Cas13d.[45 ] High distribution of
titer and galactosylation observed in enriched clones can be attributed
to the clonal variation typically observed in CHO cell lines.[51 ]
Apart from achieving the desired quality of therapeutic protein,
stable production and consistent product quality need to be maintained through the whole manufacturing process to avoid potential
issues with the safety and efficacy of the therapeutic. The analysis of
long-term stability of Fut8 expression in enriched clones (Figure S12),
performed at various time points during the study showed that in the
majority (5 out of 6) of generated enriched clones, the repression levels
of the Fut8 gene remain stable for multiple doubling cycles, therefore
the proposed glycoengineering approach enables generation of clones
with long-term and stable Fut8 repression, although the stability study
(monitoring productivity and product quality) should be an integral
part of cell line development.
Advances in protein engineering are allowing drug developers to
tune desirable protein characteristics and optimize drug targeting as
well as the potency and functionality of therapeutic proteins.[52 ] Novel
tools enabling fine-tuning of glycosylation are especially valuable in
biosimilar development, as different biosimilar projects require adjustment of critical glycans to different target levels.[26,53 ] The approach
evaluated within this study enabled isolation of clones possessing a
wide range of fucosylation (Figure 3 and Figure S7), which makes it
applicable to different biosimilar projects that differ in a target range
of fucosylation.
However, using the KRAB-dCas9 system in CHO cells still exhibits
limitations when strong gene repression is needed (> 80%). A good correlation of Fut8 gene expression and fucosylation level, especially on
the Fc domain, (Figure S10) indicates that stronger repression would
result in a stronger decrease in fucosylation of the Fc-fusion protein.
In past years, new versions of Cas9 repressors were developed with
improved gene silencing efficiency in HEK293T.[54 ] The most potent
repressor from this study, dCas9-KRAB-MeCP2, was later tested
in CHO cells.[55 ] The initial repression levels of endogenous genes
involved in apoptosis (< 30%) using monopartite dCas9 repressors
18607314, 2022, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/biot.202100499 by Health Research Board, Wiley Online Library on [24/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
GLINŠEK ET AL. 11 of 13
were only minorly improved (< 50%) by the bipartite repressor. We
observed a similar repression efficiency of the Fut8 gene using another
type of bipartite repressor, KRAB-MBD2B[54 ] when tested in CHO
cells (Figure S11), and pools containing this repressor downregulated
cell surface fucosylation to a similar extent than KRAB-only repressor.
Results from studies using traditional technologies for gene repression
indicate that siRNA or shRNA technologies might be more applicable
when a stronger reduction of protein fucosylation is desired.[19,56 ] Furthermore, not only in the case of repression but also activation of gene
expression to a sufficient level for achieving pronounced effect on protein glycosylation using CRISPR technologies remains challenging in
CHO cells.[43,57 ] Within this exploratory study, we found some limitations of the CRISPRi technology for gene repression in CHO cells,
and in certain cases, alternative technologies for gene regulation, such
as shRNA technology, might be more applicable for use in cell line
development.
Finally, we observed a more pronounced effect on Fut8 gene repression and Fc-fusion protein fucosylation in the sample produced in batch
compared to fed-batch, with the most obvious difference observed
in clones (Figure 3 and Figure S7). Process parameters, such as temperature, pH, pO2, pCO2, osmolality, media, and feed composition can
impact the glycan structure.[58–60 ] Our data suggest that bioprocess
parameters might importantly contribute to the higher sensitivity of
fucosylation to Fut8 expression in batch mode than in fed-batch mode,
although this requires further study.
In conclusion, in this study we explored a new approach for CHO
cell line glycoengineering in the development of therapeutic glycoproteins. We showed that the combination of CRISPRi technology with
lectin-FACS enrichment leads to an accelerated selection of cells with
targeted fucosylation phenotype. Moreover, it enables tuning of the
desired level of fucosylation. Enrichment of difficult to obtain glycan
profile reduces the scale of clone screening and speeds up cell line
development. Altogether, this could lead to lower costs and reduced
time to the market, which results in expanded access to therapeutics for more patients. Lastly, we believe that the proposed approach
could be easily expanded and modified to manipulate other glycan
attributes.
ACKNOWLEDGMENTS
This work was supported by the Slovenian Ministry of Education
Science and Sport and EU – European Regional development Fund
(research project OP20.04327 (C3330-19-952017)). The authors
would like to thank Taja Zotler and Janja Skok for their technical support with cell culture work, Aljaž Bertalanic, Maja Semanjski and Tilen ˇ
Vidmar for their extensive support with glycan analysis and reviewing,
as well as Ana Golob and Tamara Cvijic for their technical support with ˇ
lectin preparation and BLI analysis. The authors are grateful to Jens
Pettelkau, Christina Ranninger, Silke Ruzek for their contribution to
LC-MS/MS peptide mapping method, as well as Samuel Äschlimann for
his support with sgRNA design and Sabine Lang for her support with
FACS experiments. The authors would also like to thank Zorica Dragic,
Thomas Jostock, Dominik Gaser, Ana Lenassi Zupan and Aleš Belic for ˇ
helpful discussions during the process of preparing this manuscript.
CONFLICT OF INTEREST
The authors declare no commercial or financial conflict of interest.
DATA AVA ILAB IL ITY STATEMENT
The data that support the findings of this study are available from the
corresponding author upon reasonable request.
REFERENCES
1. Solá, R. J., & Griebenow, K. (2010). Glycosylation of therapeutic
proteins. Biodrugs, 24, 9–21.
2. Delobel, A. (2021). Glycosylation of Therapeutic Proteins: A Critical
Quality Attribute, Humana Press Inc.
3. Walsh, G. (2018). Biopharmaceutical benchmarks 2018. Nature
Biotechnology, 36, 1136–1145.
4. European Medicines Agency. (2015). EMEA/CHMP/BMWP/
42832/2005,.
5. Welch, A. R. (2018). In Biosimilars, Vol. 34 (Eds: H. J. Gutka, H. Yang, S.
Kakar) Springer, Cham, Ch. 1.
6. FDA. (2015) FDA-2011-D-0605.
7. Gherghescu, I., & Delgado-Charro, M. B. (2020). The biosimilar landscape: An overview of regulatory approvals by the EMA and FDA.
Pharmaceutics, 13, 48.
8. Walsh, G. (2010). Post-translational modifications of protein biopharmaceuticals. Drug Discovery Today, 15, 773–780.
9. Walsh, G., & Jefferis, R. (2006). Post-translational modifications in the
context of therapeutic proteins. Nature Biotechnology, 24, 1241–1252.
10. Tejwani, V., Andersen, M. R., Nam, J. H., & Sharfstein, S. T. (2018). Glycoengineering in CHO cells: Advances in systems biology. Biotechnology
Journal, 13, 1700234.
11. Zhang, P., Woen, S., Wang, T., Liau, B., Zhao, S., Chen, C., Yang, Y., Song,
Z.,Wormald,M. R., Yu, C., & Rudd, P.M. (2016). Challenges of glycosylation analysis and control: An integrated approach to producing optimal
and consistent therapeutic drugs. Drug Discover Today, 21, 740–765.
12. Liu, L. (2015). Antibody glycosylation and its impact on the pharmacokinetics and pharmacodynamics of monoclonal antibodies and
Fc-fusion proteins. Journal of Pharmaceutical Sciences, 104, 1866–1884.
13. Hajba, L., Szekrényes, Á., Borza, B., & Guttman, A. (2018). On the glycosylation aspects of biosimilarity. Drug Discovery Today, 23, 616–625.
14. Iida, S., Kuni-Kamochi, R., Mori, K., Misaka, H., Inoue, M., Okazaki,
A., Shitara, K., & Satoh, M. (2009). Two mechanisms of the enhanced
antibody-dependent cellular cytotoxicity (ADCC) efficacy of nonfucosylated therapeutic antibodies in human blood. BMC cancer, 9,
58.
15. Pereira, N. A., Chan, K. F., Lin, P. C., & Song, Z. (2018). The less-ismore in therapeutic antibodies: Afucosylated anti-cancer antibodies
with enhanced antibody-dependent cellular cytotoxicity. MAbs, 10,
693–711.
16. Shields, R. L., Lai, J., Keck, R., O’connell, L. Y., Hong, K., Meng, Y. G.,
Weikert, S. H. A., & Presta, L. G. (2002). Lack of fucose on human
IgG1 N-linked oligosaccharide improves binding to human Fc-RIII and
antibody-dependent cellular toxicity. Journal of Biological Chemistry,
277, 26733–26740.
17. Shoji-Hosaka, E., Kobayashi, Y., Wakitani, M., Uchida, K., Niwa, R.,
Nakamura, K., & Shitara, K. (2006). Enhanced Fc-dependent cellular
cytotoxicity of Fc fusion proteins derived from TNF receptor II and
LFA-3 by fucose removal from Asn-linked oligosaccharides. Journal of
Biochemistry, 140, 777–783.
18. Yamane-Ohnuki, N., Kinoshita, S., Inoue-Urakubo, M., Kusunoki, M.,
Iida, S., Nakano, R., Wakitani, M., Niwa, R., Sakurada, M., Uchida,
K., Shitara, K., & Satoh, M. (2004). Establishment ofFUT8 knockout
Chinese hamster ovary cells: An ideal host cell line for producing completely defucosylated antibodies with enhanced antibody-dependent
cellular cytotoxicity. Biotechnology and Bioengineering, 87, 614–622.
18607314, 2022, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/biot.202100499 by Health Research Board, Wiley Online Library on [24/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
12 of 13 GLINŠEK ET AL.
19. Imai-Nishiya, H., Mori, K., Inoue, M., Wakitani, M., Iida, S., Shitara,
K., & Satoh, M. (2007). Double knockdown of α1,6-fucosyltransferase
(FUT8) and GDP-mannose 4,6-dehydratase (GMD) in antibodyproducing cells: A new strategy for generating fully non-fucosylated
therapeutic antibodies with enhanced ADCC. Bmc Biotechnology [Electronic Resource], 7, 84.
20. Malphettes, L., Freyvert, Y., Chang, J., Liu, P. - Q., Chan, E., Miller,
J. C., Zhou, Z., Nguyen, T., Tsai, C., Snowden, A. W., Collingwood,
T. N., Gregory, P. D., & Cost, G. J. (2010). Highly efficient deletion
of FUT8 in CHO cell lines using zinc-finger nucleases yields cells
that produce completely nonfucosylated antibodies. Biotechnology and
Bioengineering, 106, 774–783.
21. Von Horsten, H. H., Ogorek, C., Blanchard, V., Demmler, C., Giese,
C., Winkler, K., Kaup, M., Berger, M., Jordan, I., & Sandig, V. (2010).
Production of non-fucosylated antibodies by co-expression of heterologous GDP-6-deoxy-D-lyxo-4-hexulose reductase. Glycobiology, 20,
1607–1618.
22. Chan, K. F., Shahreel, W., Wan, C., Teo, G., Hayati, N., Tay, S. J., Tong,
W. H., Yang, Y., Rudd, P. M., Zhang, P., & Song, Z. (2016). Inactivation
of GDP-fucose transporter gene ( Slc35c1 ) in CHO cells by ZFNs,
TALENs and CRISPR-Cas9 for production of fucose-free antibodies.
Biotechnology Journal, 11, 399–414.
23. Schmieder, V., Bydlinski, N., Strasser, R., Baumann, M., Kildegaard,
H. F., Jadhav, V., & Borth, N. (2018). Enhanced genome editing tools
for multi-gene deletion knock-out approaches using paired CRISPR
sgRNAs in CHO cells. Biotechnology Journal, 13, 1700211.
24. Amann, T., Hansen, A. H., Kol, S., Hansen, H. G., Arnsdorf, J.,
Nallapareddy, S., Voldborg, B., Lee, G. M., Andersen, M. R., &
Kildegaard, H. F. (2019). Glyco-engineered CHO cell lines producing
alpha-1-antitrypsin and C1 esterase inhibitor with fully humanized
N-glycosylation profiles. Metabolic Engineering, 52, 143–152.
25. Ronda, C., Pedersen, L. E., Hansen, H. G., Kallehauge, T. B., Betenbaugh,
M. J., Nielsen, A. T., & Kildegaard, H. F. (2014). Accelerating genome
editing in CHO cells using CRISPR Cas9 and CRISPy, a web-based
target finding tool. Biotechnology and Bioengineering, 111, 1604–1616.
26. Chang, M. M., Gaidukov, L., Jung, G., Tseng, W. A., Scarcelli, J. J.,
Cornell, R., Marshall, J. K., Lyles, J. L., Sakorafas, P., Chu, A. - H.,
Cote, K., Tzvetkova, B., Dolatshahi, S., Sumit, M., Mulukutla, B. C.,
Lauffenburger, D. A., Figueroa, B., Summers, N. M., Lu, T. K., &
Weiss, R. (2019). Small-molecule control of antibody N-glycosylation
in engineered mammalian cells. Nature Chemical Biology, 15, 730–736.
27. Tobinai, K., Klein, C., Oya, N., & Fingerle-Rowson, G. (2017). A review of
Obinutuzumab (GA101), a novel Type II Anti-CD20 monoclonal antibody, for the treatment of patients with B-Cell malignancies. Advances
in Therapy, 34, 324–356.
28. Li, C., Rossomando, A., Wu, S. -. L., & Karger, B. L. (2013). Comparability analysis of anti-CD20 commercial (rituximab) and RNAi-mediated
fucosylated antibodies by two LC-MS approaches. MAbs, 5, 565–
575.
29. Goletz, C., Lischke, T., Harnack, U., Schiele, P., Danielczyk, A., Rühmann,
J., & Goletz, S. (2018). Glyco-engineered anti-human programmed
death-ligand 1 antibody mediates stronger CD8 T Cell activation than
its normal glycosylated and non-glycosylated counterparts. Frontiers in
Immunology, 9, 16.
30. Gao, Y., Xiong, X., Wong, S., Charles, E. J., Lim, W. A., & Qi, L. S. (2016).
Complex transcriptional modulation with orthogonal and inducible
dCas9 regulators. Nature Methods, 13, 1043–1049.
31. Lin, J. ’., Neo, S. H., Ho, S. C. L., Yeo, J. H. M., Wang, T., Zhang, W., Bi, X.,
Chao, S. -. H., & Yang, Y. (2017). Impact of signal peptides on Furin-2A
mediated monoclonal antibody secretion in CHO cells. Biotechnology
Journal, 12, 1700268.
32. Chen, B., Hu, J., Almeida, R., Liu, H., Balakrishnan, S., Covill-Cooke, C.,
Lim, W. A., & Huang, B. (2016). Expanding the CRISPR imaging toolset
with Staphylococcus aureus Cas9 for simultaneous imaging of multiple
genomic loci. Nucleic Acids Research, 44, e75.
33. Stemmer, M., Thumberger, T., Del Sol Keyer, M.,Wittbrodt, J., & Mateo,
J. L. (2015). CCTop: An intuitive, flexible and reliable CRISPR/Cas9
target prediction tool. Plos One, 4, e0124633.
34. Gilbert, L. A., Horlbeck, M. A., Adamson, B., Villalta, J. E., Chen, Y.,
Whitehead, E. H., Guimaraes, C., Panning, B., Ploegh, H. L., Bassik, M.
C., Qi, L. S., Kampmann, M., & Weissman, J. S. (2014). Genome-scale
CRISPR-mediated control of gene repression and activation. Cell, 159,
647–661.
35. Li, R., Xia, X., Wang, X., Sun, X., Dai, Z., Huo, D., Zheng, H., Xiong, H., He,
A., & Wu, X. (2020). PLoS Biology, 18, 11.
36. Yu, K.-R., Corat, M. A. F., Metais, J.-Y., & Dunbar, C. E. (2016). 564.
The cytotoxic effect of RNA-guided endonuclease Cas9 on human
hematopoietic stem and progenitor cells (HSPCs). Molecular Therapy,
24, S225–S226.
37. Livak, K. J., & Schmittgen, T. D. (2001). Analysis of relative gene expression data using real-time quantitative PCR and the 2−ΔΔCT method.
Methods (San Diego, Calif.), 25, 402–408.
38. Chavez, A., Tuttle, M., Pruitt, B. W., Ewen-Campen, B., Chari, R., TerOvanesyan, D., Haque, S. J., Cecchi, R. J., Kowal, E. J. K., Buchthal, J.,
Housden, B. E., Perrimon, N., Collins, J. J., & Church, G. (2016). Comparison of Cas9 activators in multiple species. Nature Methods, 13,
563–567.
39. Karottki, K. J. L. C., Hefzi, H., Xiong, K., Shamie, I., Hansen, A. H., Li, S.,
Pedersen, L. E., Li, S., Lee, J. S., Lee, G. M., Kildegaard, H. F., & Lewis, N.
E. (2020). Awakening dormant glycosyltransferases in CHO cells with
CRISPRa. Biotechnology and Bioengineering, 117, 593–598.
40. Van Berkel, P. H. C., Gerritsen, J., Perdok, G., Valbjørn, J., Vink, T.,
Van De Winkel, J. G. J., & Parren, P. (2009). N-linked glycosylation is
an important parameter for optimal selection of cell lines producing
biopharmaceutical human IgG. Biotechnology Progress, 25, 244–251.
41. Ma, B., Guan, X., Li, Y., Shang, S., Li, J., & Tan, Z. (2020). Protein glycoengineering: An approach for improving protein properties. Frontiers
Chemistry, 8, 622.
42. Kanda, Y., Imai-Nishiya, H., Kuni-Kamochi, R., Mori, K., Inoue, M.,
Kitajima-Miyama, K., Okazaki, A., Iida, S., Shitara, K., & Satoh, M.
(2007). Establishment of a GDP-mannose 4,6-dehydratase (GMD)
knockout host cell line: A new strategy for generating completely nonfucosylated recombinant therapeutics. Journal of Biotechnology, 130,
300–310.
43. Karottki, K. J. L. C., Hefzi, H., Xiong, K., Shamie, I., Hansen, A. H., Li, S.,
Pedersen, L. E., Li, S., Lee, J. S., Lee, G. M., Kildegaard, H. F., & Lewis, N.
E. (2020). Awakening dormant glycosyltransferases in CHO cells with
CRISPRa. Biotechnology and Bioengineering, 117, 593–598.
44. Lee, J. S., Grav, L. M., Lewis, N. E., & Faustrup Kildegaard, H. (2015).
CRISPR/Cas9-mediated genome engineering of CHO cell factories:
Application and perspectives. Biotechnology Journal, 10, 979–994.
45. Shen, C.-C., Lin, M.-W., Nguyen, B. K. T., Chang, C.- W., Shih, J.-R.,
Nguyen, M. T. T., Chang, Y.-H., & Hu, Y.-C. (2020). CRISPR-Cas13d for
gene knockdown and engineering of CHO cells. ACS Synth Biology, 9,
2808–2818.
46. Suga, A., Nagae, M., & Yamaguchi, Y. (2018). Analysis of protein landscapes around N-glycosylation sites from the PDB repository for
understanding the structural basis of N-glycoprotein processing and
maturation. Glycobiology, 28, 774–785.
47. Thaysen-Andersen, M., & Packer, N. H. (2012). Site-specific glycoproteomics confirms that protein structure dictates formation of N-glycan
type, core fucosylation and branching. Glycobiology, 22, 1440–1452.
48. Lai, T., Yang, Y., & Ng, S. (2013). Advances in mammalian cell line
development technologies for recombinant protein production. Pharmaceuticals, 6, 579–603.
49. Lin, N., Mascarenhas, J., Sealover, N. R., George, H. J., Brooks, J., Kayser,
K. J., Gau, B., Yasa, I., Azadi, P., & Archer-Hartmann, S. (2015). Chinese hamster ovary (CHO) host cell engineering to increase sialylation
of recombinant therapeutic proteins by modulating sialyltransferase
expression. Biotechnology Progress, 31, 334–346.
18607314, 2022, 7, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/biot.202100499 by Health Research Board, Wiley Online Library on [24/03/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
GLINŠEK ET AL. 13 of 13
50. Lalonde, M.-E., Koyuturk, I., Brochu, D., Jabbour, J., Gilbert, M.,
& Durocher, Y. (2020). Production of α2,6-sialylated and nonfucosylated recombinant alpha-1-antitrypsin in CHO cells. Journal of
Biotechnology, 307, 87–97.
51. Hansen, H. G., Pristovšek, N., Kildegaard, H. F., & Lee, G. M. (2017).
Improving the secretory capacity of Chinese hamster ovary cells by
ectopic expression of effector genes: Lessons learned and future
directions. Biotechnology Advances, 35, 64–76.
52. Lagassé, H. A. D., Alexaki, A., Simhadri, V. L., Katagiri, N. H., Jankowski,
W., Sauna, Z. E., & Kimchi-Sarfaty, C. (2017). Recent advances in
(therapeutic protein) drug development. F1000Research, 6, 113.
53. Mishra, N., Spearman, M., Donald, L., Perreault, H., & Butler, M.
(2020). Comparison of two glycoengineering strategies to control the
fucosylation of a monoclonal antibody. Journal of Biotechnology X, 5,
100015.
54. Yeo, N. C., Chavez, A., Lance-Byrne, A., Chan, Y., Menn, D., Milanova, D.,
Kuo, C.-C., Guo, X., Sharma, S., Tung, A., Cecchi, R. J., Tuttle,M., Pradhan,
S., Lim, E. T., Davidsohn, N., Ebrahimkhani, M. R., Collins, J. J., Lewis, N.
E., Kiani, S., & Church, G. M. (2018). An enhanced CRISPR repressor for
targeted mammalian gene regulation. Nature Methods, 15, 611.
55. Xiong, K., Marquart, K. F., Cour Karottki, K. J., Li, S., Shamie, I., Lee, J. S.,
Gerling, S., Yeo, N. C., Chavez, A., Lee, G. M., Lewis, N. E., & Kildegaard,
H. F. (2019). Reduced apoptosis in Chinese hamster ovary cells via
optimized CRISPR interference. Biotechnology and Bioengineering, 116,
1813.
56. Beuger, V., Künkele, K.-P., Koll, H., Gärtner, A., Bähner, M., Burtscher,
H., & Klein, C. (2009). Short-hairpin-RNA-mediated silencing of fucosyltransferase 8 in Chinese-hamster ovary cells for the production
of antibodies with enhanced antibody immune effector function.
Biotechnology and Applied Biochemistry, 53, 31.
57. Marx, N., Grünwald-Gruber, C., Bydlinski, N., Dhiman, H., Ngoc
Nguyen, L., Klanert, G., & Borth, N. (2018). CRISPR-based targeted epigenetic editing enables gene expression modulation of the silenced
beta-galactoside Alpha-2,6-Sialyltransferase 1 in CHO cells. Biotechnology Journal, 13, 1700217.
58. Brunner, M., Fricke, J., Kroll, P., & Herwig, C. (2017). Investigation of
the interactions of critical scale-up parameters (pH, pO2 and pCO2)
on CHO batch performance and critical quality attributes. Bioprocess
Biosystems Engineering, 40, 251.
59. Wallner, J., Sissolak, B., Sommeregger, W., Lingg, N., Striedner, G., &
Vorauer-Uhl, K. (2019). Lectin bio-layer interferometry for assessing
product quality of Fc-glycosylated immunoglobulin G. Biotechnology
Progress, 35, e2864.
60. Popp, O., Moser, S., Zielonka, J., Rüger, P., Hansen, S., & Plöttner, O.
(2018). Development of a pre-glycoengineered CHO-K1 host cell line
for the expression of an