Poly(beta-amino ester)s as high-yield transfection reagents for
recombinant protein production
Kathryn M. Luly1,2, Stephen J.
Lee2,3, Huilin Yang2,3, Wentao
Wang1,2, Seth D. Ludwig2,3, Haley E.
Tarbox4, David R. Wilson1,2,5,
Jordan J. Green1,2,3,5,6,7,8,9,10*, Jamie B.
Spangler1,2,3,6,7,8*
1Department of Biomedical Engineering, Johns Hopkins
University School of Medicine, 2Translational Tissue
Engineering Center, Johns Hopkins University School of Medicine,3Department of Chemical & Biomolecular Engineering,
Johns Hopkins University, 4Department of Chemistry,
Johns Hopkins University, 5Institute for
Nanobiotechnology, Johns Hopkins University,6Department of Oncology, Johns Hopkins University
School of Medicine, 7Bloomberg~Kimmel
Institute for Cancer Immunotherapy, Johns Hopkins University School of
Medicine, 8Sidney Kimmel Comprehensive Cancer Center,
Johns Hopkins University School of Medicine,9Departments of Neurosurgery and Ophthalmology, Johns
Hopkins University School of Medicine, 10Department of
Materials Science & Engineering, Johns Hopkins University
*Correspondence should be addressed to:
Jamie B. Spangler
400 N Broadway, Smith 5011, Baltimore, MD 21231
(443) 287-1708
jamie.spangler@jhu.edu
Jordan J. Green
400 N Broadway, Smith 5017, Baltimore, MD 21231
(410) 614-9113
green@jhu.edu
Grant numbers: R01EY031097, R01EB029455, R01CA228133, P41EB028239,
R01CA240339
Transient transfection is an essential tool for recombinant protein
production, as rapid screening for expression is conducted without
stable integration of genetic material into a target cell genome.
Poly(ethylenimine) (PEI) is the current gold standard for transient gene
transfer, but transfection efficiency and resulting protein yield are
limited by the polymer’s toxicity. This study investigated the use of an
alternative class of cationic polymers, poly(beta-amino ester)s (PBAEs),
for transient transfection of human embryonic kidney 293F (HEK) and
Chinese hamster ovary-S (CHO) cell suspensions. In both HEK and CHO
cells, several PBAEs demonstrated superior transfection efficiency and
production of a cytosolic reporter compared to PEI. This result extended
to secreted proteins, as a model PBAE increased the production of three
secreted antibodies compared to PEI at scales ranging from 20-2,000 mL.
In particular, non-viral gene transfer using the lead PBAE/plasmid DNA
nanoparticles led to robust transfection of mammalian cells across
different constructs, doses, volumes, and cell types. These results show
that PBAEs enhance transfection efficiency and increase protein yield
compared to a widespread commercially available reagent, making them
attractive candidates as reagents for use in recombinant protein
production.
Keywords: protein production, transient transfection, poly(beta-amino
ester)s, transfection reagents
Current research into chemical-based transfection methods focuses
largely on optimizing agents for use in the development and production
of recombinant proteins. Transient transfection, in which introduced
genetic material is not incorporated into the host genome, is especially
useful during the high-throughput design and screening of proteins
(e.g., candidate biologics) wherein stable expression is not needed.
While culture conditions and plasmid design have been popular targets
for optimization in transient transfection workflows (Backliwal et al.,
2008; Galbraith, Tait, Racher, Birch, & James, 2006), further research
into improved transfection reagents has even greater potential for
boosting protein yields. Chemical-based transient transfection relies on
condensation and encapsulation of plasmid DNA by a biocompatible
material into particles which are taken up by target cells; differences
in particle size can affect the method of cellular uptake, leading to
differences in transfection efficiency (Kim, Sunshine, & Green, 2014).
Particles must then escape the endosome and the encapsulating material
must degrade to allow for DNA release, nuclear translocation,
transcription and subsequent export, and finally translation and
processing into fully formed protein (Karlsson, Rhodes, Green, & Tzeng,
2020).
Transfection reagent structure and buffering capacity have been
demonstrated to influence DNA uptake and escape, making these properties
particularly consequential in reagents for transient transfection
workflows (Sunshine, Peng, & Green, 2012). Maximizing protonability,
for example, facilitates endosomal swelling and consequent rupture via
the “proton sponge” effect (Boussif et al., 1995; Bus, Traeger, &
Schubert, 2018). Cationic polymers have typically been among the most
promising transfection reagents; their charge-based association with DNA
into particles offers protection from degradation and offers sufficient
buffering capacity to facilitate endosomal escape following cellular
uptake (Sunshine et al., 2012).
Poly(ethyleneimine) (PEI) is a commercially available cationic polymer
used extensively as a transfection reagent that has a high density of
protonatable amines, giving rise to high buffering capacity and
efficient endosomal escape (Boussif et al., 1995). PEI of average
molecular weight 25 kDa is most frequently used in transfection
workflows, but its toxicity limits transfection efficiency and,
consequently, protein yield (Breunig, Lungwitz, Liebl, & Goepferich,
2007; Yang, Li, Goh, & Li, 2007). Previously, PEI has been conjugated
to polyethylene glycol (Petersen et al., 2002) and arginine modified
oligo(-alkylaminosiloxane) [P(SiDAAr)n] (Morris & Sharma, 2010) to
mitigate cytotoxicity.
A promising alternative to PEI, poly(beta-amino ester)s (PBAEs) are a
class of cationic polymers used to facilitate efficient gene transferin vitro (Bishop, Kozielski, & Green, 2015). PBAEs are composed
of an acrylate base monomer, an amine sidechain, and a terminal
end-capping group, each of which can be varied to create a vast library
of materials (Akinc, Lynn, Anderson, & Langer, 2003). Hydrolyzable
ester linkages allow for degradation of the PBAEs in transfection
conditions which allows for use of the polymers at high weight ratios
relative to other non-biodegradable materials, maximizing density of
buffering amines to facilitate endosomal escape (Sunshine et al., 2012).
Their biodegradability also obviates the need for medium replacements or
additions, themselves contributors to cell death, which are common where
PEI is utilized (Galbraith et al., 2006). These linear polymers are
synthesized from inexpensive, commercially available reagents using a
two-step polymerization method (Fig. S1A) and are stable long term when
stored dry at -20°C (Wilson et al., 2019).
Given the high transfection efficacy observed with PBAEs in variousin vitro contexts, we sought to investigate the use of PBAE
nanoparticles for transient transfection of suspension cultures in
intracellular and secreted protein production workflows (Fig. 1A). We
selected four PBAEs with varying base (B), sidechain (S), and end-cap
(E) structures to evaluate in comparison with linear 25 kDa PEI:
B4-S4-E6 (4-4-6); B4-S5-E7 (4-5-7); B4-S5-E39 (4-5-39); and B5-S3-E6
(5-3-6) (Fig. 1B-C, S1B). Physiochemical characterization of PBAE and
PEI nanoparticles in serum-free transfection media indicated that PBAE
nanoparticles maintained a smaller size in transfection conditions
(approximately 200-350 nm) whereas PEI nanoparticles were prone to
aggregation, resulting in sizes over 1 µm (Fig. 1D, S1C). Previous
studies indicated that PEI nanoparticles were prone to aggregation in
serum-free media due to interactions with salts and a lack of adsorbed
proteins that can help stabilize discrete particles and prevent
clustering (Ogris et al., 1998; Pezzoli, Giupponi, Mantovani, &
Candiani, 2017). Analysis of surface charge revealed that PBAE
nanoparticles maintained a positive zeta potential in transfection
conditions, whereas PEI nanoparticles exhibited a near neutral surface
charge (Fig. 1E, S1C). Shielded surface charge of PEI particles may
limit interactions with a charged cell membrane, thus hindering cellular
uptake.
To determine the optimal DNA dose for production of cytosolic mCherry
using various polymer-based transfection agents, we selected a
representative PBAE, 2-(3-aminopropylamino)ethanol end-capped
poly(1,4-butanediol diacrylate-co-4-amino-1-butanol) (referred to here
as 4-4-6), and compared this PBAE to 25 kDa PEI at doses ranging from
0.5 to 4 µg/mL DNA. The polymers were compared in two mammalian cell
lines frequently employed for protein expression: human embryonic kidney
293F (HEK) cells and Chinese hamster ovary-S (CHO) cells. Evaluation of
mCherry fluorescence over a span of 5 days indicated that peak mCherry
expression occurred using the 4-4-6 polymer at 2 µg/mL and 4 µg/mL DNA
doses in HEK and CHO cells, respectively (Fig. S1A). Notably, peak
mCherry expression in PEI-based transfections was not comparable to that
attained by 4-4-6 at any dose. Subsequent time course studies using the
optimized DNA dose that compared additional PBAE structures demonstrated
significantly increased mCherry expression using 4-4-6, 4-5-7, and
4-5-39 in HEK cells, and using 4-4-6 and 4-5-7 in CHO cells, compared to
PEI-mediated transfection (Fig. 2A, 2C). Transfection efficiency, as
measured by mCherry-positive cells on day 5, was significantly increased
with all PBAEs tested in HEK cells and with 4-4-6, 4-5-7, and 4-5-39 in
CHO cells (Fig. 2B, 2D, S2C). Fluorescence microscopy confirmed the
increase in mCherry expression (Fig. 2E, S2D). Cell viability, assessed
via MTS assay 24 h following transfection, indicated that PBAEs
(especially 4-5-39 and 5-3-6) showed greater toxicity than PEI, though
notably at a 20- to 30-fold higher weight ratio (Fig. S2B). Importantly,
this reduced viability did not result in inferior mCherry expression
relative to PEI, with 4-4-6 and 4-5-7 demonstrating superior expression
in both HEK and CHO cells (Fig. 2A, C).
To demonstrate that the results of these fluorescent protein expression
experiments were replicable at scales relevant to the development and
production of secreted proteins, we transfected HEK cell cultures of
varying volumes with DNA encoding the recombinant antibody 10H2
(Chuntharapai, Lee, Hébert, & Kim, 1994; Patent No. WO/2020/243489,
2020) using either 4-4-6- or PEI-based particles. Based on SDS-PAGE
image analysis of small-scale dose titrations (Fig. S3A-D), DNA was
dosed at 1 µg/mL for secreted proteins in HEK cells, with polymer weight
adjusted accordingly. At volumes ranging from 20-200 mL, transfection
with 4-4-6 yielded between 4.5-fold and 8.2-fold more protein than did
transfection with PEI (Fig. 3A). To further demonstrate the scalability
of enhanced protein expression using PBAEs, 2 L cultures of HEK cells
were transfected with DNA encoding either 10H2 or the bispecific
antibody BS2 (Patent No. WO/2020/243489, 2020) using either 4-4-6 or
PEI. Both the 10H2 and BS2 antibodies were recovered in significantly
higher quantities (4.9-fold and 5.6-fold higher, respectively) when
4-4-6 was utilized compared to PEI (Fig. 3B-C). Superiority of PBAE
nanoparticles was reproducible across cell lines; transfection of CHO
cells with DNA encoding the recombinant antibody 602 (Krieg, Letourneau,
Pantaleo, & Boyman, 2010; Létourneau et al., 2010; Patent No.
WO/2020/264321, 2020) at an optimized dose of 4 µg/mL resulted in
3.4-fold more protein recovered when 4-4-6 was used compared to PEI
(Fig. S3E-G).
Taken together, experiments with both cytosolic and secreted proteins
demonstrated that PBAEs lead to significantly enhanced protein yields
compared to leading commercial reagent PEI in two cell lines that are
widely used for protein production. Storage stability and
straightforward synthesis from inexpensive chemical monomers further
strengthen their attractiveness for use in recombinant protein
production across batch scales. Overall, the favorable properties of
PBAEs combined with the results herein suggest that these polymers hold
promise as superior reagents for transient transfection that can
significantly improve protein production workflows.