Methods
Dock and Boat Construct
Design
The Dock sites contain in order: a modified Long Terminal Repeat (mLTR)
which possesses very weak promoter activity, an enhanced packaging
region required for retroviral packaging, the Phi-C31 recombinaseatt achment sequence from the Phi-C31 P hage genome (attP),
and another mLTR (Figure 2).
The single gene Boat expression construct contains (in order) the
Phi-C31 recombinase att achment sequence from theStreptomyces b acterial genome (attB), the Glutamine
Synthetase (GS) gene, a strong mammalian promoter from the simian
cytomegalovirus immediate-early gene, the gene of interest (GOI), an
mRNA export element and a polyadenylation signal (Figure 2). The two
gene Boat expression construct adds a human cytomegalovirus
immediate-early gene promoter, intron, GOI, and polyadenylation sequence
at the end of the single gene Boat expression construct. The Boat
constructs are designed to utilize a promoter trap approach such that
random integration would not result in significant GS expression, since
the Boat lacks its own promoter to drive GS expression. However, upon
proper targeted integration into the Dock, the GS gene is situated
downstream of the mLTR which possesses weak promoter activity. Multiple
other methods are used to reduce GS transcription and translation,
allowing for survival of clones with high recombined Boat copy numbers
while clones containing few or no copies of integrated Boat are
outcompeted by the high producing cell lines.
Retrovector Production and Transduction to Create the Dock
Parental Cell
Line
GPEx technology leverages an HEK 293 cell line that constitutively
expresses MMLV gag, pro, and pol proteins (Bleck, 2005). The dock
construct described above was co-transfected with an envelope expression
plasmid. Retrovector was subsequently concentrated by
ultracentrifugation and used for transduction of the CHOZN®
(MilliporeSigma) Chinese Hamster Ovary parental cell line. Sequential
rounds of transduction were performed, and cells were routinely
maintained in media supplemented with 6 mM glutamine.
Production of GPEx® Lightning Technology
Pools
Coding sequences for commercially available biotherapeutics for
monoclonal IgG1-kappa antibodies and an Fc fusion protein were cloned
into either a single gene or two gene Boat expression construct. 3
million GPEx® Lightning Parental Dock cells were transfected with 2
micrograms (total) of Boat plasmid and recombinase plasmid DNA using
ExpiFectamineTM CHO (ThermoFisher Scientific). Cells
were allowed to recover in Ex-Cell® Advanced CHO Fed-Batch media
(MilliporeSigma) supplemented with 6 mM glutamine and 1% ClonaCell™-CHO
ACF Supplement (Stem Cell Technologies) until viability returned to
greater than 95% before glutamine was withdrawn. Viability was
monitored until the resulting selected cell pools returned to greater
than 95% viability and doubling times reached 18-24 hours, typically 16
to 22 days after transfection.
Single Cell Cloning
The Berkeley Lights Beacon® instrument was used for single cell cloning
(Kravitz, 2019). The individual clones from the GPEx® Lightning Dock
pool were ranked and selected based on desirable doubling time
(17.5-23.5 hours). All clones were exported onto 96 well plates,
expanded, frozen, and subjected to quantitative Polymerase Chain
Reaction (QPCR) to quantify the number of docks per cell. All Dock
clones were tested by PCR for retroviral components to ensure no viral
genes had been incorporated into the genome, and thus future cell lines
made from the Parental Dock line could not produce replication competent
retrovirus.
Clonal lines from pools expressing protein products were ranked and
exported based on the Beacon Spotlight assay which measures relative
titer through the Fc portion of the molecules. All clones were exported
onto 96 well plates, expanded, and frozen.
Quantification of
Recombination:
Genomic DNA from 3 million cells was isolated using a Qiagen DNEasy kit.
QPCR using SYBR® green dye (Thermo Fisher Scientific) was employed to
estimate the number of specific recombination events into the Dock
sites. AttR formation is the result of site-specific recombination
between attP in the genome (Dock) and attB on the transgene construct
(Boat). QPCR was performed to quantify attR using a forward primer in
the attP sequence in the Dock and a reverse primer in the attB sequence
in the Boat. Amplification using this primer pair only detects the
transgene construct when it is recombined into a Dock but does not
detect empty Dock, random integration, integration into a pseudo-attP
site in the genome (Thyagarajan, 2005), or free plasmid. Cycle threshold
(Ct) values from the attR primer set were subtracted from the Ct value
from an internal reference gene primer set to determine the delta-Ct
values. Those values were used to estimate copy number using the
equation: copy number =2^(delta-Ct). The amplification efficiency of
both primer sets was determined to be 90-100% via a template DNA
sequential dilution assay.
Fed-Batch Production
Generic Fed-Batch Strategy 1
(GFB1)
Cells were inoculated at 600,000 cells per mL in Ex-Cell® Advanced CHO
Fed-Batch media (MilliporeSigma) and incubated in a humidified (70-80%)
shaking incubator with 5% CO2 and temperature of 37°C
(shift to 34°C on day 4). Cultures were fed every other day starting on
day 2 with 6.25% (V:V) of a feed blend containing 66% Ex-Cell®
Advanced CHO Feed 1 and 33% Cellvento® 4Feed (MilliporeSigma). Glucose
was monitored daily and supplemented if below 5 g/L. Cultures were
terminated when viabilities were ≤ 70% or at the end of day 20. Cell
concentration and viability were measured by a Roche Cedex HiRes®
instrument and cell metabolites, glucose, and IgG titer were measured by
a Roche Cedex Bio HT® instrument.
Generic Fed-Batch Strategy 2
(GFB2)
Methodology was as described in GFB1 except media and feeds were
replaced with ActiProTM media and 5%
CellBoostTM 7a and 0.5% CellBoostTM7b feeds (Cytiva).
Production in Ambr® 250
Bioreactors
All methodologies were the same as GFB2 except Ambr® 250 (Sartorius)
vessels were inoculated at 6.5 x 105 cells/mL, pH set
point was 7.0 with deadband of -0.2 and +0.5. Baseline agitation was set
to 20 watts/m3 and increased as needed to reach the
dissolve oxygen set point of 30%. Actual rates ranged from 340-500 RPM.
Stability Studies
Pools or clones were continually passaged in Ex-Cell® Advanced CHO
Fed-Batch media (MilliporeSigma) or ActiProTM (Cytiva)
media and frozen weekly. All generations were simultaneously thawed and
evaluated for production.
Glycosylation Analysis
Sample Preparation
Protein-A purified protein was denatured in 1%
RapiGestTM (Waters) solution, followed by the
simultaneous release of N-glycans using RapidTMPNGaseF (New England BioLabs) and derivatization with RapiFluor-MS
(Waters). The GlycoWorks™ SPE Reagents-Automation kit (Waters) and
GlycoWorks™ HILIC µElution plate (Waters) were used to purify and
concentrate the released and fluorescently labeled N-glycans.
Chromatography
Conditions
These N-glycans were separated and analyzed using a Waters Acquity UPLC
(Ultra Performance Liquid Chromatography) H-Class equipped with
fluorescence detector, an Acquity® Glycan BEH130 Amide, 2.1x150mm
(Waters) column, and acetonitrile (mobile phase A) and 50mM ammonium
formate (mobile phase B). A gradient of 25% to 46% B 0.4 mL/min in
35.0 min, then a gradient shift to 100% B 0.2 mL/min to 36.5 min,
followed by an isocratic step at 100% B 0.2 mL/min from 36.5 to 39.5
min, then a gradient shift to 25% B 0.2 mL/min to 43.1min, and then
equilibration at 25% B 0.4 mL/min from 47.6 min to 55.0 min. Column
temperature was 60°C and a 10 µL injection volume was used.
Glycan Detection and Identification
N-glycans were fluorescently detected (λex = 265nm,
λem = 425nm) at a sample rate of 10 Hz and mass
confirmed using a Waters Xevo® G2-XS Q-ToF mass spectrometer.
RapiFluor-MSTM Calibration Ladder was injected (1 µL)
at the beginning and the end of the sequence to determine the glycan
units (GU) for the samples tested and assist in assignment of identity.
The calculated glycan units for each observed species in conjunction
with mass confirmation were compared to the Waters RFMS Glycan Library
for released N-glycan identification.