Rapid high-throughput cloning and stable expression of antibodies in HEK293 cells

    Rapid high-throughput cloning and stable expression of antibodies in HEK293 cells Jared L. Spidel, Benjamin Vaessen, Yin Yin Chan, Luigi Grasso, J. Bradford Kline PII: DOI: Reference:

S0022-1759(16)30211-3 doi: 10.1016/j.jim.2016.09.007 JIM 12222

To appear in:

Journal of Immunological Methods

Received date: Revised date: Accepted date:

18 May 2016 21 September 2016 21 September 2016

Please cite this article as: Spidel, Jared L., Vaessen, Benjamin, Chan, Yin Yin, Grasso, Luigi, Kline, J. Bradford, Rapid high-throughput cloning and stable expression of antibodies in HEK293 cells, Journal of Immunological Methods (2016), doi: 10.1016/j.jim.2016.09.007

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Rapid High-Throughput Cloning and Stable Expression of Antibodies in HEK293 Cells Jared L. Spidela,*, Benjamin Vaessena, Yin Yin Chana, Luigi Grassoa, J. Bradford Klinea

PT

a

AC CE P

TE

D

MA

NU

SC

RI

Morphotek Inc., 210 Welsh Pool Road, Exton, PA, USA *Corresponding author. E-mail address: [email protected] (J. L. Spidel)

1

ACCEPTED MANUSCRIPT Abstract Single-cell based amplification of immunoglobulin variable regions is a rapid and

PT

powerful technique for cloning antigen-specific monoclonal antibodies (mAbs) for

RI

purposes ranging from general laboratory reagents to therapeutic drugs. From the initial screening process involving small quantities of hundreds or thousands of mAbs through

SC

in vitro characterization and subsequent in vivo experiments requiring large quantities of

NU

only a few, having a robust system for generating mAbs from cloning through stable cell line generation is essential. A protocol was developed to decrease the time, cost, and

MA

effort required by traditional cloning and expression methods by eliminating bottlenecks in these processes. Removing the clonal selection steps from the cloning process using

TE

D

a highly efficient ligation-independent protocol and from the stable cell line process by utilizing bicistronic plasmids to generate stable semi-clonal cell pools facilitated an

AC CE P

increased throughput of the entire process from plasmid assembly through transient transfections and selection of stable semi-clonal cell pools. Furthermore, the time required by a single individual to clone, express, and select stable cell pools in a highthroughput format was reduced from 4 to 6 months to only 4 to 6 weeks. Keywords: monoclonal antibody; high-throughput cloning; bicistronic; transfection; stable cell line 1

Introduction

The rapid production of monoclonal antibodies (mAbs) isolated from an immunized organism is essential for the development of therapeutic and diagnostic reagents and critical in the support of vaccine studies. Traditional methods for the production and

2

ACCEPTED MANUSCRIPT screening of mAbs involving hybridomas or EBV transduction (Aman et al. 1984; Beerli and Rader,2010; Kohler and Milstein,1975; Kozbor et al. 1982; Redmond et al. 1986;

PT

Stahli et al. 1980; Steinitz et al. 1977; Striebich et al. 1990; Strohl,2014) are inefficient

RI

due to genomic instability and low mAb expression of the resulting cell lines. To circumvent these limitations, technologies have been developed to clone variable

SC

domain genes into a phage, yeast, ribosomal, or mammalian display platform (Beerli

NU

and Rader,2010; Lim et al. 2014; Saggy et al. 2012; Strohl,2014). PCR amplified variable heavy (VH) and light (VL) regions from an immunized organism are randomly

MA

paired in a library that is screened for antigen binding, and, subsequently, the top binding VH-VL region pairs are subcloned into mammalian expression plasmids for

TE

D

large-scale production. While antibody display technologies are powerful techniques, generating a quality library is limited by the efficiencies of the random VH-VL pairing

AC CE P

(not all pairs result in optimal antigen binding) and cloning process and poor expression/enrichment of many relevant antigen specific V genes (Saggy et al. 2012). To overcome the limitations associated with hybridoma/immortal cell lines and combinatorial display platforms, new techniques have been developed to isolate and clone variable region cDNAs from antigen-specific single B cells into antibody expression vectors (Scheid et al. 2009; Smith et al. 2009; Tiller et al. 2009; Tiller et al. 2008; Tiller et al. 2007; Wardemann et al. 2003; Wrammert et al. 2008). Since the heavy (HC) and light chains (LC) are already paired, there would be no need to screen large libraries for correctly paired mAbs. Single B cell cloning simplifies the process to a single cloning and expression step, bypasses the multiple rounds of panning/screening and the subsequent subcloning into mammalian expression vectors required by display 3

ACCEPTED MANUSCRIPT techniques. Only a single round of screening is required to select the highest affinity mAbs whose recombinant production will be then accomplished by establishing high-

PT

producing stable cell lines. However, the number of antibodies capable of being cloned

RI

and expressed using this method is limited by traditional time- and effort-consuming

SC

methods.

Cloning PCR amplified variable domain genes into an expression vector typically

NU

requires growth on agar plates and screening of multiple clones that contain the desired

MA

insert. Plating hundreds of bacterial transformations on agar plates and screening multiple colonies from each plate is extremely time consuming and low throughput.

D

Further, generating stable cell lines or performing large-scale or multiple small-scale

TE

transient transfections to serve as a continuous source of recombinant mAb for

AC CE P

screening and preclinical development is impractical for dozens or hundreds of lead mAbs. An ideal mAb cloning and expression platform would contain a single high throughput cloning step followed by rapid generation of a continuous source of recombinant mAbs.

In this report, we describe a robust high-throughput mAb discovery and production system that integrates the advantages of high-efficiency cloning and stable pool selection methods, reducing the time and effort required to clone and express large numbers of amplified V genes. 2 2.1

Materials and methods Plasmid Construction

Monocistronic plasmids were generated from a customized pcDNA-3.1-based plasmid

4

ACCEPTED MANUSCRIPT where the HC or LC gene was cloned into a multiple cloning site between a CMV promoter and SV40 polyA signal. The HC plasmids contained the zeocin resistance

RI

blasticidin gene under the control of the SV40 promoter.

PT

gene under the control of the SV40 promoter, and the LC plasmids contained the

SC

Bicistronic plasmids were generated from a customized pcDNA-3.1-based plasmid. A modified encephalomyocarditis virus (EMCV) IRES with point mutations G375A and

NU

A484G to remove ApaI and HindIII restriction sites, respectively, was fused at the 5’ end

MA

of a cDNA conferring resistance to either blasticidin (B), hygromycin (H), neomycin (N), puromycin (P), or zeocin (Z), by using overlapping PCR, and cloned between the 3’ end

TE

pCIx, where x is B, H, N, P, or Z.

D

of a multiple cloning site and a SV40 polyA signal. The resulting plasmids were named

AC CE P

PCR was used to amplify a mouse immunoglobulin kappa leader with 3’ restriction sites that overlapped 5’ restriction sites added to human gamma and kappa constant (C) regions during PCR amplification. The leader and C regions were joined together by overlapping PCR and cloned into the multiple cloning site of pCIx vectors to make pCIxleader-C region vectors (Fig. 1). Leader

Leader

InFusion Cloning Site

CMV Promoter

InFusion Cloning Site

CMV Promoter

kappa Constant Region

gamma Constant Region

IRES

pCIx-ldr-kappa

IRES

pCIx-ldr-gamma

Selection Marker

5124 bp

polyA

5790 bp

Ampicillin Resistance

Ampicillin Resistance

Selection Marker

F1 ori

SV40 polyA

ORI pUC

F1 ori

ORI pUC

5

ACCEPTED MANUSCRIPT Figure 1. Plasmid schematics for pCIx-leader-C region vectors.

V regions cDNAs were amplified by RT-PCR from RNA isolated from B cells similar to

PT

Lightwood et al (2006) with primers containing 15 base pairs (bp) at the 5’ ends that

RI

were homologous to the signal sequence or C region. An In-Fusion HD cloning kit

SC

(Clontech) was used to clone the V regions between the leader and C regions in

NU

restriction enzyme-digested pCIx-leader-C region vectors (Fig. 2A). For high-throughput cloning of variable regions, following addition of SOC medium to

MA

the transformation and incubation at 37°C for 1 h, 50 μL of each transformation pool was added to 1 mL Terrific Broth in a 2-mL 96-well deep well plate, and shaken

D

overnight in a microtiter plate shaker at 37°C. Glycerol stocks were made from a portion

TE

of the cultures and stored at -80°C. The cultures were then processed for small-scale

AC CE P

plasmid DNA extraction (miniprep) with Qiagen’s QIAGEN Plasmid Plus 96 Miniprep Kit using an Eppendorf epMotion 5075 VAC. The semi-clonal DNA was sequenced to confirm the presence of the insert in the vector. To determine the efficiency of the In-Fusion reactions, several microliters of the glycerol stocks were added to LB agar plus ampicillin plates (Teknova) and streaked for isolation. Sixteen colonies from each plate were PCR-screened by resuspending the colony in 5 μL of water and adding 5 μL of Taq-Pro Complete (Denville), 0.1 μL of a forward primer homologous to the 5’ UTR, and 0.1 μL of a reverse primer homologous to the 5’ end of the IRES. Following PCR amplification, the products were analyzed on a 1% Tris-acetate-EDTA agarose gel. Point mutations were introduced in heavy chain plasmids using a QuikChange Lightning 6

ACCEPTED MANUSCRIPT kit (Agilent) according to the manufacturer’s protocol. 2.2

Cell Culture

PT

Human embryonic kidney (HEK) FreeStyle 293-F cultures were maintained in FreeStyle

RI

293 expression medium (ThermoFisher) according to the manufacturer’s instructions.

SC

Cells were transfected by seeding 1.8x107 cells in 30 mL medium in a 125-mL shake flask one day prior to transfection. On the day of transfection, 5 μg of each HC and LC

NU

plasmids were combined with 65 μg PEI (linear, MW 25000; Polysciences) in 1.5 mL

MA

OptiPro (Life Technologies). After incubation at room temperature for 15 minutes, the PEI-DNA mixture was added to 3x107 cells in 30 mL medium in a 125-mL shake flask.

D

Cultures were incubated at 37°C in 8% CO2 with shaking at 125 rpm for 8 to 10 days.

TE

Antibody-expressing stable pools were selected by adding 3 mL of transfectants to 12

AC CE P

mL cDMEM (complete DMEM supplemented with 10% FBS, ThermoFisher) in a T75 flask with the appropriate pair of selection drugs (5 μg/mL blasticidin, 100 μg/mL hygromycin, 3 μg/mL puromycin, 400 μg/mL zeocin [ThermoFisher], 400 μg/mL G418 [Calbiochem]) one to four days after transfection. After drug-resistant cells grew to confluency, the medium was replaced with FreeStyle 293 expression medium for 24 to 48 hours. Cells were physically dislodged by tapping the flask (trypsinization resulted in low viability, data not shown) and were then seeded at 6x105 cells/mL in 30 mL FreeStyle 293 expression medium in a 125-mL shake flask. Cultures were incubated at 37 °C in 8% CO2 with shaking at 125 rpm. After passaging the cells three times, terminal cultures were set up by seeding cells at 1x106 cells/mL in 30 mL FreeStyle 293 expression medium in a 125-mL shake flask. Cultures were incubated at 37 °C in 8% CO2 with shaking at 125 rpm for 10 days. It was found that when performing the drug 7

ACCEPTED MANUSCRIPT selection in a serum-containing static adherent culture, cells became drug resistant much faster than when selected for resistance in a serum-free suspension culture.

PT

There was no difference in mAb production between the two selection methods (data

RI

not shown).

SC

Clonal cell lines were obtained by seeding 96-well plates with an average of 0.5 cells per well in cDMEM. After 3 to 4 days, conditioned cDMEM from untransfected 293-F

NU

cells and drug selection to a final concentration of 200 μg/mL zeocin and 2.5 μg/mL

MA

blasticidin was added. After further incubation for several more days or weeks, wells containing a single colony were grown in 24-well or 6-well plates and eventually

D

expanded to T75 flasks. Once confluent, mAb expression of the clones was determined

TE

by seeding 1x106 cells in 3 mL of cDMEM supplemented with low-IgG FBS

AC CE P

(ThermoFisher) for 7 days. The amount of mAb in the medium was then quantitated. Clones were then adapted to FreeStlye293 medium and the amount of mAb expressed was quantitated as above. 2.3

Antibody Quantitation

Antibody titer yields were quantified using an Octet QK with Dip and Read Protein A Biosensors (ForteBio) according to the manufacture’s protocol. 3 3.1

Results High throughput, semi-clonal platform for constructing recombinant mAb expression plasmids

The biggest bottleneck in the cloning process of individual HCs and LCs is the selection of recombinant transformants on antibiotic agar plates and subsequent colony

8

ACCEPTED MANUSCRIPT screening. To streamline the cloning process, these steps were eliminated by utilizing the highly efficient ligation-independent In-Fusion HD cloning kit. In the In-Fusion HD

PT

reaction, reactions rely on homologous recombination between the vector and insert to

RI

circularize the vector, and, therefore, there is little to no background from religated vector. Furthermore, it was found that there was little to no background colonies

SC

resulting from undigested plasmid when the vector was digested with three restriction

NU

enzymes that cut between the leader and C region sequences (Fig. 2A; data not shown), thereby making it possible to grow and miniprep transformations as a semi-

MA

clonal pool.

D

A.

TE

3’ Homologous Region

5’ Homologous Region

AC CE P

--AfeI--PmeI--EcoRV...GCCACCGGCGTGCACAGCgctgtttaaacgatATCCACCAAGGGCCC... nucleotide sequence ACCGGCGTGCACAGC...........gccATCCACCAAGGGCCC leader ... A

T

VH insert

G

V

H

S| x

x

PCR insert

human gamma

x |A

S

3’ Homologous Region

T

K

G

P ... amino acid sequence

5’ Homologous Region

--AfeI

EcoRV- PmeI----

restriction sites

GCCACCGGCGTGCACAGCgcttagatatctgtttAAACGAACTGTGGCT... nucleotide sequence ACCGGCGTGCACAGC................AAACGAACTGTGGCT leader ... A

T

VK insert G

V

H

S|

x

x

x

x

PCR insert

human kappa x

K |R

B.

9

T

V

A ... amino acid sequence

*

Figure 2. In-Fusion cloning of variable region inserts.

RI

*

SC

*

PT

ACCEPTED MANUSCRIPT

MA

NU

(A) HC and LC constant-region vectors were designed to contain a multiple cloning site between the leader sequence and the constant region. An in-frame stop codon was included in the multiple cloning site (underlined) to mitigate translation of proteins lacking a variable region insert. The HC constantregion vectors were cut with AfeI, PmeI, and EcoRV and LC vectors were cut with AfeI, EcoRV, and PmeI for linearization. Variable regions were PCR amplified with 15-bp 5’ and 3’ ends homologous to the vector. The variable region PCR fragments were fused to the vector by In-Fusion cloning.

TE

D

(B) Miniprepped pools of transformed E. coli were digested with restriction enzymes to determine the size of the antibody cDNA in the plasmid. Heavy chain plasmids with a variable region insert produced a ~1.4 kb fragment. Light chain plasmids that contain a variable region insert produced a ~700 bp fragment. Heavy and light chain plasmids without inserts produce 1 kb and 250 bp fragments, respectively. The wells alternate between HC and LC, except the last 8 samples on the bottom row are all LC plasmids. Arrows indicate fragments with no insert. Asterisks indicate a mixture of fragments with and without inserts.

AC CE P

To validate and quantify this method, 44 VH genes and 51 VL genes were RT-PCR amplified from clonal B cell populations and cloned into C region vectors. Minipreps of semi-clonal DNA pools were digested with restriction enzymes that cut 5’ and 3’ to the leader and C region, respectively, and analyzed on a 1% agarose gel. As predicted, HC plasmids with or without an insert yielded a ~1.4 kb or 1 kb fragment, respectively, and LC plasmids with or without an insert yielded a fragment at ~700 bp or 250 bp, respectively. The great majority of LC plasmids (98%) and HC plasmids (95.4%) contained the expected insert (Fig. 2B), while only few plasmids did not (arrows). Some of the HC minipreps (6.8%) contained a mixture of plasmids with and without inserts (asterisks). The percentage of plasmids with inserts in each transformation was further quantitated 10

ACCEPTED MANUSCRIPT by streaking out a small portion of the transformation on agar plates and screening individual colonies by PCR for the presence of an insert. The percentage of clones with

PT

inserts varied from 31% to 100% per transformation with an average of 71.6% for HC

RI

plasmids and 81.1% for LC plasmids (Table 1). The wide range of positive clones was likely due to incomplete digestion of the vector by the restriction enzymes as

SC

subsequent experiments utilizing a different preparation of vectors resulted in much

no. InFusion reactions

no. colonies screened/reaction

average % positive

Heavy Chain

42

D

MA

NU

higher percentage of positive clones (data not shown).

71.6%

± 3.7%

Light Chain

42

16

81.1%

± 3.2%

TE

16

SEM

AC CE P

Table 1. The average percentages of inserts present in 42 heavy chain or light chain plasmids.

This approach cannot only be applied to ligation-independent cloning, but also to wholeplasmid amplification-based mutagenesis of heavy and light chain plasmids using methods such as the QuikChange Lightning Site-Directed Mutagenesis kit. Several hundred site-directed HC plasmid mutants were generated and the transformants were miniprepped as semi-clonal pools as above. The semi-clonal DNA preps were each sequenced for the presence of the point mutation(s). All semi-clonal plasmid pools contained only the desired mutation(s) (data not shown). 3.2

Bicistronic plasmid design and expression

Even though a large number of mAbs can be rapidly produced via small-scale transient

11

ACCEPTED MANUSCRIPT transfections for screening purposes, scaling up selected mAbs post-screening to use for in vitro/in vivo studies can be more challenging. Although several techniques to

PT

enhance sufficient production for biological studies have been described over the last

RI

several years (Backliwal et al. 2008; Pham et al. 2006; Pham et al. 2003; Sun et al. 2006), these methods do not yield a process for stable cell line production to provide the

SC

continuous source of material required for support of longer term development of

NU

candidate therapeutic or diagnostic mAbs. Similar to high-throughput cloning, the major bottleneck for generating stable mammalian cell lines is screening for high-titer clones.

MA

This method typically selects for plasmid integration in transfected cells by performing limiting dilution of the cells in growth medium containing drug selection (Jostock and

AC CE P

(Fig. 1).

TE

D

Knopf,2012), a process that is labor-intensive and requires between 4 and 6 months

An alternative method utilizes selection of stable pools of cells. Here, transfected cells are not subcloned and drug selection is applied to the pool of transfected cells (Ye,2012). Selection of a pool of mAb-expressing cells overcomes these hurdles since the expressing cell line is not subjected to multiple rounds of screening. To avoid generating stable cell pools with low antibody titers due to the abundance of non- or low-expressing antibiotic resistant cells, as seen in traditional approaches that utilize monocistronic expression plasmids (Ye et al. 2010), bicistronic plasmids were designed to transcriptionally link expression of a drug resistance protein for blasticidin (blasticidin S deaminase, BSD), G418 (neomycin phosphotransferase, NPT), hygromycin (hygromycin B phosphotransferase, HPH), puromycin (puromycin N-acetyltransferase, PAC), or zeocin (bleomycin resistance protein, BRP) to the expression of a HC or LC by 12

ACCEPTED MANUSCRIPT an EMCV IRES on a single mRNA transcript (Fig. 1). The HC and LC for human (mAb1) and chimeric rabbitV-humanC (mAb2) mAbs were cloned into both bicistronic pCIZ and

PT

pCIB vectors, respectively, as well as monocistronic zeocin-resistance and blasticidin-

RI

resistance vectors, respectively. Transient mAb expression was reduced only 20-30% when the HC and LC were cotranslated with antibiotic resistance genes from bicistronic

SC

mRNA when compared to monocistronic mRNA (Fig. 3A). However, mAb expression

NU

from stable pools was 2- to 3-fold higher for the bicistronic plasmids.

B2 0 0

MA

A 175 T ra n s ie n t S ta b le

150

****

*

D TE

75

****

400

****

m g /L m A b

300

200

100

2

2

b

b

A

A

m

m

ic

ic

n

n

o

o

tr

tr

ic b

o o

n

is

is c

is ic b

m

o

n

o

c

is

tr

tr

o

o

n

n

ic

ic

m

m

A

A

b

b

1

1

0

m

13

2

2 m

o

b A m ic n o tr is ic b

n

b

o

ic

c

is

is

tr

tr

o

o

n

n

ic

ic

m

m

A

A

b

b

1

1 m

o

n

b

n

o

C

b m ic n o tr is c

ic

c

o

is

is

tr

tr

o

o

n

n

ic

ic

m

m

A

A

b

b

2

2

0

m

m

o

n

b

o

ic

c

is

is

tr

tr

o

o

n

n

ic

ic

m

m

A

A

b

b

1

1

0

AC CE P

25

50

A

50

100

o

m g /L m A b

100

m g /L m A b

150

125

ACCEPTED MANUSCRIPT Figure 3. Comparing mAb expression from monocistronic and bicistronic mRNA.

NU

SC

RI

PT

The HC and LC of human (mAb1) and rabbit-human chimeric (mAb2) mAbs were cloned into monocistronic plasmids and expressed. The HC and LC genes were also cloned into bicistronic plasmids where the translation of the antibiotic resistance protein was from the same mRNA as the HC or LC via an IRES sequence separating the HC or LC and the antibiotic resistance gene. (A) The expression of mAb1 and mAb2 from monocistronic or bicistronic plasmids in transiently transfected 293-F cells using an Octet QK with protein A Biosensor tips 10-12 days post-transfection. Pools of 293-F cells stably expressing the mAbs were also analyzed for mAb expression using an Octet QK with protein A Biosensor tips 7-10 days after seeding the terminal cultures. (B) Individual clones were isolated from the stable pools, and the expression of each mAb in static culture was analyzed for expression as in (A). (C) Clonal cell lines were adapted to 293FreeStyle medium and culture in shake flasks, and the expression of each mAb was analyzed for expression as in (A). Average expression levels between the monocistronic and bicistronic cell lines in (B) and (C) was analyzed by Welch's unequal variances t-test. **** p <0.0001, * p = 0.0411

MA

Single cell clones were subsequently isolated from the stable pools via limiting dilution and tested for mAb titer yields. Up to forty clones were analyzed for mAb expression in

D

static culture (Fig. 3B). The average mAb titers for bicistronic clones were 1.5- to 2.5-

TE

fold higher than monocistronic clones. The highest titer clones producing mAb1 and

AC CE P

mAb2 were converted to serum-free medium, and analyzed for production yields in shaking cultures. Titers from clones using bicistronic expression plasmids were 2.2- to 2.4-fold higher than from monocistronic expression plasmids (Fig. 3C). 3.3

Pairing antibiotic resistance plasmids for maximal expression

To maximize mAb expression, the HC and LC genes are expressed by using separate plasmids. Therefore, obtaining stable cell lines that consistently expressed both chains requires different resistant genes on each plasmid. To determine the best combination of drug resistance genes that may result in the highest mAb titers from both transient transfections and stable pools, the HC and LC of mAb1 were cloned into vectors that contained each drug resistance gene. Each HC plasmid was cotransfected in a matrix format into FreeStyle 293-F cells with each LC plasmid. The media from transient

14

ACCEPTED MANUSCRIPT transfections were harvested after ten days, and the amount of secreted mAb was analyzed (Fig. 4A). Transient expression of the mAb was found to be influenced by the

PT

drug resistance gene in the LC plasmid as clones exhibiting the lowest titers (17-45 μg/mL) expressed NPT (pCIN) and HPH (pCIH) while those with the highest-titers (66-

RI

75 μg/mL) expressed PAC (pCIP). These data were observed in other studies as well

SC

(see below). There was no correlation between mAb expression and the drug resistance

NU

gene expressed by the HC plasmid.

B 300

MA

A T ra n s ie n t

300

S ta b le

T ra n s ie n t S ta b le

250

250

m g /L m A b

D TE

150

100

50

600

T ra n s ie n t

550

S ta b le

500

m g /L m A b

450 400

-P Z

-Z P

-N Z

-Z N

-H Z

-Z H

-B Z

-Z B

-P Z

-Z P

-Z

-N Z

N

-P

-N

H C - L C s e le c t io n g e n e

H C - L C s e le c t io n g e n e

C

150

0

N

-Z

-H

Z

-P

-H

-H

P

H

-N

N

H

-Z

-B

Z

B

-P

-B

P

-N

-B

-B

B

N

B

H

B

-H

0

AC CE P

50

P

100

H

m g /L m A b

200

200

350 300 250 200 150 100 50

-P Z

-Z P

-N Z

-Z N

-H Z

-Z H

-B Z

B

-Z

0

H C - L C s e le c t io n g e n e

Figure 4. Antibody expression using IRES-linked antibiotic selection plasmids. Bicistronic IRES plasmids encoding various antibiotic resistant genes (B – blasticidin, H – hygromycin, N – neomycin, P – puromycin, Z – zeocin) downstream of immunoglobulin HC or LC genes were co-

15

ACCEPTED MANUSCRIPT

PT

transfected into 293Freestyle cells, e.g. B-H is a blasticidin-HC plasmid cotransfected with a hygromycinLC plasmid. Various combinations of resistance genes and immunoglobulin HC and LCs were used including (A) human (mAb1), (B) rabbit-human chimeric (mAb2), or (C) rat-human chimeric (mAb3) mAbs. Antibody titers were analyzed using an Octet QK with protein A Biosensor tips 10-12 days posttransfection for transient transfections or 7-10 days after seeding the terminal cultures with stably selected cell lines.

RI

The effect of antibiotic selection on the expression of mAb1 in stable pools was determined by quantitating the amount of secreted mAb ten days after seeding terminal

SC

cultures. In contrast to transient expression, mAb production did not correlate to the

NU

drug resistance gene in the LC plasmid. However, linking expression of BRP (pCIZ) to

MA

either the HC or the LC resulted in the highest mAb titers (Fig. 4A). To determine whether zeocin resistance results in the best universal expression , the

D

HC and LC plasmids of two additional mAbs (mAb2 and ratV-mouseC chimeric, mAb3)

TE

were co-transfected as pCIB and pCIZ, pCIH and pCIZ, pCIN and pCIZ, or pCIP and

AC CE P

pCIZ combinations. Transient expression of mAb2 and mAb3 was similar to mAb1 in that the lowest titers were observed with NPT- and HPH-expressing LC plasmids (Fig. 4B and 4C). There was little difference in transient expression between the other plasmid combinations (40-60 μg/mL and 36-51 μg/mL, respectively). Stable pools were selected, and the trends in expression of mAb2 and mAb3 were similar to mAb1 (Fig. 4B and 4C). Although the puromycin-resistant plasmids yielded the highest producing cell lines, a stable pool could not be obtained from the mAb3 HCpuromycin/LC-zeocin combination. Repeated attempts to generate a stable cell line from the mAb3 HC-puromycin/LC-zeocin combination were not successful even when the concentration of puromycin was reduced several fold. It was also observed that all mAb1, mAb2, and mAb3 stable cell lines made from any PAC-based plasmid took longer to grow to confluency following drug selection due to extreme sensitivity of 293-F 16

ACCEPTED MANUSCRIPT cells to puromycin (data not shown). Given the inconsistency of obtaining a stable cell line from puromycin-resistant plasmids, these plasmids were deemed least useful and

PT

not used for any further development.

RI

For all three mAbs listed above, the stable pool generated from the HC-blasticidin/LC-

SC

zeocin combination yielded the lowest-titers. Although stable mAb expression from HC plasmids containing hygromycin or neomycin was relatively high, the transient

NU

expression was rather low, limiting the usefulness of these plasmids. In addition, viable

MA

mAb3 stable pools were not obtained for the HC-hygromycin/LC-zeocin nor HCzeocin/LC-neomycin combinations (Fig. 4B). Based on these results we focused our

D

efforts on the HC-zeocin/LC-blasticidin plasmid combination platform since this format

Effects of increased or decreased selection pressure on mAb expression

AC CE P

3.4

TE

consistently yielded high mAb titers from both transient transfections and stable pools.

Since mAb expression was linked to drug resistance, it was possible that increasing the amount of antibiotics during selection could select for cells that express higher amounts of the antibiotic resistance proteins, and thereby higher levels of mAb. Varying concentrations of blasticidin and zeocin were added during the mAb1 selection process to evaluate this hypothesis. Unfortunately, increasing the concentration of zeocin up to 1200 μg/mL or blasticidin up to 15 μg/mL did not yield higher titer clones (Fig. 5A).

17

ACCEPTED MANUSCRIPT

A

B 300

300

250

250

200

200

5 /4 0 0 2 .5 /2 0 0 1 .2 5 /1 0 0

100 50

PT

150

150

100

50

0

0

5/800

5/1200

15/800

15/1200

28

LC/HC selection gene

56

SC

5/400

RI

m g /L m A b

mg/L mAb

0 /0

84

115

NU

d a y p o s t-s e le c tio n

Figure 5. Blasticidin and zeocin concentrations have little effect on mAb production.

D

MA

(A) HC/LC cotransfected 293FreeStyle cells were dually selected with varying concentrations of blasticidin (5 or 15 μg/ml) and zeocin (400, 800, or 1200 μg/ml). Terminal cultures from stably selected cell pools were analyzed after 7-10 days for antibody titers using an Octet QK with protein A Biosensor tips. (B) The concentration of antibiotics was reduced to one-half (2.5/200), one-quarter (1.25/100), or none (0/0) of the original concentration (5/400) and assayed every 4 weeks for antibody production as in (A).

TE

The stability of mAb expression by stable pools was analyzed by monitoring the

AC CE P

expression of mAb1 over time. After the eighth passage in 400 μg/mL zeocin and 5 μg/mL blasticidin, the pooled culture was split into three subcultures and cultured using decreasing amounts of drug selection. Decreasing the concentration of antibiotics by one-half or one-quarter did not affect mAb expression. Furthermore, removal of drug selection resulted in little decrease in mAb production for at least 56 days (15 passages, Fig. 5B). By day 84 (passage 29 for 5/400; passage 22 for the remainder) production began to decrease for all pools, though cells cultured without selection decreased more than the others. 3.5

Summary of Workflow

An exemplary workflow is detailed in Fig. 6. Herein the VH and VL regions are RT-PCR amplified from individual single B cell cultures and ligation-independently inserted into constant region vectors. The E. coli transformation for each reaction is grown in liquid 18

ACCEPTED MANUSCRIPT medium, rather than plated for isolation of individual colonies. This shortens the cloning process to 2 to 4 days from 1 to 2 weeks. The semi-clonal HC and LC DNAs cloned

PT

from each B cell culture are cotransfected and a portion of each transfection is selected

RI

as a pool in static serum-containing culture or serum-free shaking culture. Bypassing the need to select and screen individual colonies, the semi-clonal transfectants are

SC

selected within 7 to 11 days, as opposed to 4 to 5 weeks, and eliminates the expansion

NU

process, thereby save an additional 4 to 5 weeks. During the selection process the transient cultures can be screened for antigen binding, and the top mAbs picked for

MA

serum-free adaptation and expansion. The adaptation of the semi-clonal culture to serum-free medium requires no subsequent rescreening for mAb production, thereby

TE

D

saving more than 3 weeks. In total, our semi-clonal method cuts the time from cloning

AC CE P

through stable cell line development by 3 to 5 months.

19

AC CE P

TE

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

Figure 6. Flowchart for high throughput cloning, transfection, and stable cell line selection.

4

Discussion

In this report, we have described a rapid, optimized high-throughput system for the cloning and expression of immunoglobulin VH and VL cDNAs by utilizing the same

20

ACCEPTED MANUSCRIPT plasmid for both transient mAb production as well as large-scale production by stably selected lines. Several groups have demonstrated either high-throughput methods for

PT

cloning mAbs (Dodev et al. 2014; Jones et al. 2010; Kurosawa et al. 2011; Liao et al.

RI

2009) or rapid generation of stable mAb-expressing cell lines, but to date there are no reports that couple these technologies into a single platform. Indeed, the methods and

SC

vectors used in high-throughput cloning methods are not amenable to rapid stable cell

NU

line production as they are based on transfection of PCR products (Liao et al. 2009), lack an antibiotic resistance gene (Dodev et al. 2014) or utilize monocistronic plasmids

MA

(Jones et al. 2010; Kurosawa et al. 2011). While, Ho et al. (Ho et al. 2012; Ho et al. 2013) demonstrated rapid generation of stable pools expressing monoclonal antibodies

TE

D

in CHO cells using a tricistronic plasmid where the HC, LC, and antibiotic resistance gene were translated from a single transcript, this method requires extra steps in the

AC CE P

cloning process. Two rounds of cloning are required to insert both the HC and LC into the plasmid. Alternatively, a single cloning step can be used to insert a single PCR fragment containing an IRES sequence flanked by the HC and LC. However, this method requires an extra round of PCR to create the aforementioned PCR product. Therefore, cloning the HC and LC into separate bicistronic expression plasmids provides the simplest and quickest means to generate expression plasmids and stable cell lines. One of the major advantages of this system is the ability to move quickly a lead mAb to large-scale production by creating stable cell pools using bicistronic expression plasmids that transcriptionally linking expression of the HC and LC with antibiotic resistance genes. In contrast to stable pools generated from monocistronic plasmids 21

ACCEPTED MANUSCRIPT (Ye et al. 2010), there is no need to sort cells for enrichment (Ye,2012; Ye et al. 2010) or subclone for high-producing cells (Jostock and Knopf,2012) since all antibiotic

PT

resistant cells theoretically express the mAb (Gurtu et al. 1996; Rees et al. 1996).

RI

Furthermore, this system is amenable to the cloning and expression of a large number of other types of recombinant proteins (e.g., surface antigens for high-throughput target

SC

cell line generation).

NU

Several steps have been eliminated to significantly enhance the cloning throughput

MA

while decreasing labor hours typically required for high-throughput mAb discovery. A few areas for improvement were experienced during this optimization process. Given

D

that 20-30% of plasmids lacking an insert in the miniprep pool, it was possible that mAb

TE

expression could be affected by either expression of aberrant leader-constant domain

AC CE P

protein or by diluting out the overall amount of insert-containing plasmid. To minimize these issues, a stop codon was engineered in the cloning site between the signal sequence and C region (Fig. 2A). A V region insert replaces this stop codon so that only productive mAbs are translated. In addition, it has been shown that non-coding DNA can supplement more than half of the coding DNA in a PEI-based transfection with little effect on recombinant protein expression (Kichler et al. 2005; Rajendra et al. 2012); data not shown). Even though the presence of insert-null vector likely has little effect on overall mAb expression within a pooled culture, one could employ a zero-background cloning strategy to ensure a cloning efficiency closer to 100% (Bernard,1995; Bernard et al. 1994; Lund et al. 2014; Miyazaki,2010; Parr and Ball,2003; Schefer et al. 2014; Wang et al. 2014). For recombinant mAb production, careful consideration was given to the choice of cells 22

ACCEPTED MANUSCRIPT used. While Chinese hamster ovary (CHO) cells are predominantly the cells of choice for the production of biopharmaceuticals, HEK cells are much more popular for transient

PT

transfection due to their inherent ease of transfection and high production from transient

RI

transfections (Butler and Meneses-Acosta,2012). Despite recent improvements of the protocols to increase the expression levels of mAbs in transient CHO transfections,

SC

these methods tend to involve additional steps, including diluting cells several hours

NU

after transfection, adding feed several hours or days post-transfection, decreasing the incubation temperature, and/or performing multiple rounds of transfections (Chusainow

MA

et al. 2009; Codamo et al. 2011; Pichler et al. 2011; Rajendra et al. 2015; Rajendra et al. 2011; Reisinger et al. 2009; Wulhfard et al. 2008). As these additional manipulations

TE

D

increase effort and cost, we decided that the use of HEKs would be most ideal for at least transient expression. It can be argued that the cell type used during mAb

AC CE P

characterization should be the same cell type used to manufacture the therapeutic mAb due to differences in post-translational modification differences (e.g. glycosylation). However, a recent report demonstrated that the mAbs produced by transient HEK transfections correlate with stable CHO cell lines in regards to mAb titers and the quality of the product (Diepenbruck et al. 2013). Given this observation and as the intended use of this system being the expression of mAbs for primary candidate selection, any differences between HEK-derived mAbs versus CHO-derived mAbs are negligible. Interestingly, our findings that the choice of antibiotic resistance gene pairings influences the mAb expression offers future opportunities to test other selection markers for even higher titers than is achieved in our current optimal system. Importantly, while our findings that resistance to zeocin was essential in obtaining a high expressing pool 23

ACCEPTED MANUSCRIPT were similar to Lanza et al (2013), we found, in contrast to Lanza et al, no advantage in maintaining drug selection after primary selection for up to eight weeks. Therefore, drug

PT

selection can be removed while scaling up the culture for mAb production, providing a

RI

distinct advantage over stable pools selected from monocistronic plasmids which

SC

require maintenance of drug selection during scale-up (Ye,2012). In conclusion, we outline an antibody cloning and screening system that can facilitate

NU

the cloning, expression, and screening of up to a thousand mAbs, which may be

MA

expandable using automation. By removing the bottlenecks described here, the time from isolating V-region cDNA to producing large-scale mAbs from stable pools has been

Acknowledgments

TE

5

D

reduced from 4 to 6 months to only 4 to 6 weeks.

AC CE P

We thank Marianne Henry for her work in RT-PCR amplifying VH and VL cDNA from single B cells cultures.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Reference List Aman P, Ehlin-Henriksson B, Klein G. 1984. Epstein-Barr virus susceptibility of normal human B lymphocyte populations. J Exp Med 159:208-220. Backliwal G, Hildinger M, Chenuet S, Wulhfard S, De JM, Wurm FM. 2008. Rational vector design and multi-pathway modulation of HEK 293E cells yield recombinant

24

ACCEPTED MANUSCRIPT antibody titers exceeding 1 g/l by transient transfection under serum-free conditions. Nucleic Acids Res 36:e96.

PT

Beerli RR, Rader C. 2010. Mining human antibody repertoires. MAbs 2:365-378.

RI

Bernard P. 1995. New ccdB positive-selection cloning vectors with kanamycin or

SC

chloramphenicol selectable markers. Gene 162:159-160.

NU

Bernard P, Gabant P, Bahassi EM, Couturier M. 1994. Positive-selection vectors using

MA

the F plasmid ccdB killer gene. Gene 148:71-74.

Butler M, Meneses-Acosta A. 2012. Recent advances in technology supporting

TE

894.

D

biopharmaceutical production from mammalian cells. Appl Microbiol Biotechnol 96:885-

AC CE P

Chusainow J, Yang YS, Yeo JH, Toh PC, Asvadi P, Wong NS, Yap MG. 2009. A study of monoclonal antibody-producing CHO cell lines: what makes a stable high producer? Biotechnol Bioeng 102:1182-1196. Codamo J, Munro TP, Hughes BS, Song M, Gray PP. 2011. Enhanced CHO cell-based transient gene expression with the epi-CHO expression system. Mol Biotechnol 48:109115. Diepenbruck C, Klinger M, Urbig T, Baeuerle P, Neef R. 2013. Productivity and quality of recombinant proteins produced by stable CHO cell clones can be predicted by transient expression in HEK cells. Mol Biotechnol 54:497-503.

25

ACCEPTED MANUSCRIPT Dodev TS, Karagiannis P, Gilbert AE, Josephs DH, Bowen H, James LK, Bax HJ, Beavil R, Pang MO, Gould HJ, Karagiannis SN, Beavil AJ. 2014. A tool kit for rapid

PT

cloning and expression of recombinant antibodies. Sci Rep 4:5885.

RI

Gurtu V, Yan G, Zhang G. 1996. IRES bicistronic expression vectors for efficient

SC

creation of stable mammalian cell lines. Biochem Biophys Res Commun 229:295-298.

NU

Ho SC, Bardor M, Feng H, Mariati, Tong YW, Song Z, Yap MG, Yang Y. 2012. IRESmediated Tricistronic vectors for enhancing generation of high monoclonal antibody

MA

expressing CHO cell lines. J Biotechnol 157:130-139.

D

Ho SC, Bardor M, Li B, Lee JJ, Song Z, Tong YW, Goh LT, Yang Y. 2013. Comparison

TE

of internal ribosome entry site (IRES) and Furin-2A (F2A) for monoclonal antibody

AC CE P

expression level and quality in CHO cells. PLoS One 8:e63247. Jones ML, Seldon T, Smede M, Linville A, Chin DY, Barnard R, Mahler SM, Munster D, Hart D, Gray PP, Munro TP. 2010. A method for rapid, ligation-independent reformatting of recombinant monoclonal antibodies. J Immunol Methods 354:85-90. Jostock T, Knopf HP. 2012. Mammalian stable expression of biotherapeutics. Methods Mol Biol 899:227-238. Kichler A, Leborgne C, Danos O. 2005. Dilution of reporter gene with stuffer DNA does not alter the transfection efficiency of polyethylenimines. J Gene Med 7:1459-1467. Kohler G, Milstein C. 1975. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256:495-497.

26

ACCEPTED MANUSCRIPT Kozbor D, Lagarde AE, Roder JC. 1982. Human hybridomas constructed with antigenspecific Epstein-Barr virus-transformed cell lines. Proc Natl Acad Sci U S A 79:6651-

PT

6655.

RI

Kurosawa N, Yoshioka M, Isobe M. 2011. Target-selective homologous recombination

SC

cloning for high-throughput generation of monoclonal antibodies from single plasma

NU

cells. BMC Biotechnol 11:39.

Liao HX, Levesque MC, Nagel A, Dixon A, Zhang R, Walter E, Parks R, Whitesides J,

MA

Marshall DJ, Hwang KK, Yang Y, Chen X, Gao F, Munshaw S, Kepler TB, Denny T, Moody MA, Haynes BF. 2009. High-throughput isolation of immunoglobulin genes from

D

single human B cells and expression as monoclonal antibodies. J Virol Methods

TE

158:171-179.

AC CE P

Lightwood DJ, Carrington B, Henry AJ, McKnight AJ, Crook K, Cromie K, Lawson AD. 2006. Antibody generation through B cell panning on antigen followed by in situ culture and direct RT-PCR on cells harvested en masse from antigen-positive wells. J Immunol Methods 316:133-143.

Lim BN, Tye GJ, Choong YS, Ong EB, Ismail A, Lim TS. 2014. Principles and application of antibody libraries for infectious diseases. Biotechnol Lett 36:2381-2392. Lund BA, Leiros HK, Bjerga GE. 2014. A high-throughput, restriction-free cloning and screening strategy based on ccdB-gene replacement. Microb Cell Fact 13:38. Miyazaki K. 2010. Lethal ccdB gene-based zero-background vector for construction of shotgun libraries. J Biosci Bioeng 110:372-373. 27

ACCEPTED MANUSCRIPT Parr RD, Ball JM. 2003. New donor vector for generation of histidine-tagged fusion proteins using the Gateway Cloning System. Plasmid 49:179-183.

PT

Pham PL, Kamen A, Durocher Y. 2006. Large-scale transfection of mammalian cells for

RI

the fast production of recombinant protein. Mol Biotechnol 34:225-237.

SC

Pham PL, Perret S, Doan HC, Cass B, St-Laurent G, Kamen A, Durocher Y. 2003.

NU

Large-scale transient transfection of serum-free suspension-growing HEK293 EBNA1 cells: peptone additives improve cell growth and transfection efficiency. Biotechnol

MA

Bioeng 84:332-342.

D

Pichler J, Galosy S, Mott J, Borth N. 2011. Selection of CHO host cell subclones with

TE

increased specific antibody production rates by repeated cycles of transient transfection

AC CE P

and cell sorting. Biotechnol Bioeng 108:386-394. Rajendra Y, Hougland MD, Alam R, Morehead TA, Barnard GC. 2015. A high cell density transient transfection system for therapeutic protein expression based on a CHO GS-knockout cell line: Process development and product quality assessment. Biotechnol Bioeng 112:977-986. Rajendra Y, Kiseljak D, Baldi L, Hacker DL, Wurm FM. 2011. A simple high-yielding process for transient gene expression in CHO cells. J Biotechnol 153:22-26. Rajendra Y, Kiseljak D, Manoli S, Baldi L, Hacker DL, Wurm FM. 2012. Role of nonspecific DNA in reducing coding DNA requirement for transient gene expression with CHO and HEK-293E cells. Biotechnol Bioeng 109:2271-2278.

28

ACCEPTED MANUSCRIPT Redmond MJ, Leyritz-Wills M, Winger L, Scraba DG. 1986. The selection and characterization of human monoclonal antibodies to human cytomegalovirus. J Virol

PT

Methods 14:9-24.

RI

Rees S, Coote J, Stables J, Goodson S, Harris S, Lee MG. 1996. Bicistronic vector for

SC

the creation of stable mammalian cell lines that predisposes all antibiotic-resistant cells

NU

to express recombinant protein. Biotechniques 20:102-110. Reisinger H, Steinfellner W, Katinger H, Kunert R. 2009. Serum-free transfection of

MA

CHO cells with chemically defined transfection systems and investigation of their

D

potential for transient and stable transfection. Cytotechnology 60:115-123.

TE

Saggy I, Wine Y, Shefet-Carasso L, Nahary L, Georgiou G, Benhar I. 2012. Antibody isolation from immunized animals: comparison of phage display and antibody discovery

AC CE P

via V gene repertoire mining. Protein Eng Des Sel 25:539-549. Schefer Q, Hallmann S, Grotzinger C. 2014. Knockin' on pHeaven's door: a fast and reliable high-throughput compatible zero-background cloning procedure. Mol Biotechnol 56:449-458.

Scheid JF, Mouquet H, Feldhahn N, Seaman MS, Velinzon K, Pietzsch J, Ott RG, Anthony RM, Zebroski H, Hurley A, Phogat A, Chakrabarti B, Li Y, Connors M, Pereyra F, Walker BD, Wardemann H, Ho D, Wyatt RT, Mascola JR, Ravetch JV, Nussenzweig MC. 2009. Broad diversity of neutralizing antibodies isolated from memory B cells in HIV-infected individuals. Nature 458:636-640.

29

ACCEPTED MANUSCRIPT Smith K, Garman L, Wrammert J, Zheng NY, Capra JD, Ahmed R, Wilson PC. 2009. Rapid generation of fully human monoclonal antibodies specific to a vaccinating

PT

antigen. Nat Protoc 4:372-384.

RI

Stahli C, Staehelin T, Miggiano V, Schmidt J, Haring P. 1980. High frequencies of

SC

antigen-specific hybridomas: dependence on immunization parameters and prediction

NU

by spleen cell analysis. J Immunol Methods 32:297-304.

Steinitz M, Klein G, Koskimies S, Makel O. 1977. EB virus-induced B lymphocyte cell

MA

lines producing specific antibody. Nature 269:420-422.

D

Striebich CC, Miceli RM, Schulze DH, Kelsoe G, Cerny J. 1990. Antigen-binding

TE

repertoire and Ig H chain gene usage among B cell hybridomas from normal and

AC CE P

autoimmune mice. J Immunol 144:1857-1865. Strohl WR. 2014. Antibody discovery: sourcing of monoclonal antibody variable domains. Curr Drug Discov Technol 11:3-19. Sun X, Goh PE, Wong KT, Mori T, Yap MG. 2006. Enhancement of transient gene expression by fed-batch culture of HEK 293 EBNA1 cells in suspension. Biotechnol Lett 28:843-848. Tiller T, Busse CE, Wardemann H. 2009. Cloning and expression of murine Ig genes from single B cells. J Immunol Methods 350:183-193.

30

ACCEPTED MANUSCRIPT Tiller T, Meffre E, Yurasov S, Tsuiji M, Nussenzweig MC, Wardemann H. 2008. Efficient generation of monoclonal antibodies from single human B cells by single cell RT-PCR

PT

and expression vector cloning. J Immunol Methods 329:112-124.

RI

Tiller T, Tsuiji M, Yurasov S, Velinzon K, Nussenzweig MC, Wardemann H. 2007.

SC

Autoreactivity in human IgG+ memory B cells. Immunity 26:205-213.

NU

Wang J, Xu R, Liu A. 2014. IRDL cloning: a one-tube, zero-background, easy-to-use, directional cloning method improves throughput in recombinant DNA preparation. PLoS

MA

One 9:e107907.

D

Wardemann H, Yurasov S, Schaefer A, Young JW, Meffre E, Nussenzweig MC. 2003.

AC CE P

301:1374-1377.

TE

Predominant autoantibody production by early human B cell precursors. Science

Wrammert J, Smith K, Miller J, Langley WA, Kokko K, Larsen C, Zheng NY, Mays I, Garman L, Helms C, James J, Air GM, Capra JD, Ahmed R, Wilson PC. 2008. Rapid cloning of high-affinity human monoclonal antibodies against influenza virus. Nature 453:667-671.

Wulhfard S, Tissot S, Bouchet S, Cevey J, De JM, Hacker DL, Wurm FM. 2008. Mild hypothermia improves transient gene expression yields several fold in Chinese hamster ovary cells. Biotechnol Prog 24:458-465. Ye J. 2012. Stable transfection pools for large quantity of protein production. Methods Mol Biol 899:221-225.

31

ACCEPTED MANUSCRIPT Ye J, Alvin K, Latif H, Hsu A, Parikh V, Whitmer T, Tellers M, de la Cruz Edmonds MC, Ly J, Salmon P, Markusen JF. 2010. Rapid protein production using CHO stable

AC CE P

TE

D

MA

NU

SC

RI

PT

transfection pools. Biotechnol Prog 26:1431-1437.

32