Fluorescence-based method validation
HIV-1 VLP PRODUCTION PLATFORM BASED ON TRANSIENT TRANSFECTION OF HEK 293 SUSPENSION CULTURES IN BIOREACTORS Cervera, L., Gutiérrez-Granados, S., Gòdia, F., Segura, M.M. [email protected] [email protected]
Cell and Tissue Engineering Research Group Departament d’Enginyeria Química, Universitat Autònoma de Barcelona, Edifici Q, 08193 Bellaterra, Barcelona, Spain
OPTIMIZING HEK 293 CELL GROWTH USING DOE
ABSTRACT Robust and scalable Gag VLP production processes are required to generate VLPs for preclinical and eventually clinical testing. Transient gene expression offers a convenient route to accelerate the identification of candidate VLP vaccines as it allows for the production of a large number of product variants with relative ease. Considerable progress has been made in the past several years towards establishing large-scale transient transfection protocols. The human embryonic kidney 293 (HEK 293) is the preferred host system due to the many industrially relevant features this cell line offers including high transfectability, ability to grow in suspension, ability to grow to high cellular densities and adaptation to serum-free culture conditions. HEK 293 mammalian cell cultures were grown in suspension and serum-free conditions in disposable shake flasks and WAVE bioreactors. VLPs were generated by transient transfection using a Gag-GFP plasmid construct and polyethilenimine as transfection reagent. The GFP tagged VLPs were quantified using a spectrofluorometer. An optimized platform for the generation of Gag VLPs in mammalian suspension cultures is described in this work. To facilitate process optimization, a fast and reliable quantitation technique based on fluorescence was developed and validated. An in house culture medium supplemented with non-animal derived components was defined and the different transient transfection parameters were fine-tuned to allow maximum cell growth and VLP production. The great majority of Gag-GFP recovered from cell culture supernatants was shown to be correctly assembled into VLPs of the expected size and morphology. The scalability of this strategy was demonstrated using a 1L WAVE bioreactor The quantity and quality of Gag VLPs generated using the described platform is suitable for pre-clinical studies in mice. By using a GMP-compliant suspension cell line, disposable bioreactors and a culture medium devoid of animal-derived components, a fast translation into clinical trials is envisioned.
A
B
A SIMPLE AND RELIABLE QUANTITATION METHOD FOR FLUORESCENT GAG VLPS
C A
Fluorescence-based method
p24 ELISA
Gag-GFP fusion protein
Gag-GFP fusion protein
7 to 1000 RFU
10 to 300 pg/mL
Precision
~2% CV
~10% CV
Limit of detection
10 ng/mL
10 pg/mL
Time (96 samples)
1.5 h
4h
Price (96 samples)
7€
400 €
BB Specificity Linear range
Box-Behnken model accurately predicts HEK 293 cell densities HEK 293 cell growth as a function of the concentrations of Lipid mixture (X) vs. r-Transferrin (mg/L) (A), Lipid mixture (X) vs. r-Insulin (mg/L) (B) and r-Transferrin (mg/L) vs. r-Insulin (mg/L) (C) based on Box-Behnken experimental results. Optimal concentrations for each supplement were determined using this experimental design (D). D
E
Optimum levels of additives r-Insulin (mg/L)
19.8
r-Transferrin (mg/L)
1.6
Lipid mixture (X)
0.9
RFU: Relative fluorescence units; CV: coefficient of variation.
Growth kinetics of HEK 293 cells in optimized cell culture medium (E) The maximum cell concentration reached was 5.4x106 cells/mL, same value as predicted using the Box-Behnken model. Values presented are the mean ± SD (n=3).
1.5×108 VLPs/immunization
EXPERIMENTAL APPROACH OPTIMIZING GAG VLP PRODUCTION A
B FreeStyle medium
+ N H
n Transient transfection of HEK 293 suspension cultures
VLP harvest
Fluorescent Gag VLP characterization VLP size (TEM, NanoSight®)
VLP protein composition (SDS-PAGE, Western-blot)
VLP morphology (TEM) VLP characterization studies facilitated by fluorescence
B
5 µm
5 µm
Confocal fluorescence microscopy images of HEK 293 producer cells (A) Gag-GFP was accumulated in the vicinity of the plasma membrane where the budding process takes place. (B) Upon staining with Cell Mask™, co-localization of green (Gag-GFP molecules) and red (lipid membrane) fluorescence was observed in yellow. Cell nuclei were stained with Hoechst (blue).
Transmission electron microscopy (TEM) images of Gag-GFP VLPs (A) Negatively stained Gag-GFP VLPs and latex beads used for quantitation purposes (full and empty arrows, respectively). (B) High magnification image showing roughly spherical viral particles surrounded by a lipid envelope and containing electro-dense Gag core. The mean VLP diameter was determined to be 141 ± 22 nm (n=100) which is the expected size of Gag-GFP VLPs as they resemble HIV immature particles which are larger than wild-type HIV-1 virions (4). Gag-GFP VLP size distribution determined by Nanoparticle Tracking Analysis (NTA) VLPs were analyzed by NTA NanoSight® and the size distribution of 3 replicates is shown. The most frequent particle size value (statistical mode) was 149 ± 5 nm.
CONCLUSIONS
Scale up
Plasmid DNA and PEI polyplex formation
A
Fluorescence-based method validation Validation of the developed fluorescence-based quantitation assay was carried out according to ICH guidelines (3). Fluorescence measurements correlated well with p24 concentrations determined by ELISA, as shown by the correlation coefficient (A). All validation parameters met the criteria for analytical method validation (B).
GAG VLP PRODUCTION IN WAVE REACTOR
Generation of Gag-GFP VLPs
VLP CHARACTERIZATION
1L bioreactor C
D Supplemented FreeStyle medium + medium exchange
VLP quantitation (Fluorometry, p24 ELISA, NTA)
Transfection efficiency (Flow citometry)
VLP budding (Fluorescence microscopy, TEM)
750 immunizations
VLP concentration
Fluorescencebased method
p24 ELISA
NTA Nanosight®
VLP (particles/mL)
4.58×109
7.13×109
5.02×109
Mean VLP production was improved using optimal conditions of transient transfection Cell suspensions were transfected with 1 µg of DNA /mL of culture and a DNA to PEI mass ratio of 1:2. HEK 293 transfection efficiency (A, C) and VLP fluorescence intensity (B, D) in cell culture supernatants were measured . Mean values ± SD (n=3). The optimal conditions founded were attained when using supplemented FreeStyle medium with medium exchange performed just before transfection and a cell concentration of 2×106. With this optimization VLP production was improved 2.6-fold.
5.6×109 VLP/mL
VLP production in 1L Wave bioreactor Gag-GFP VLPs were produced in a 1L Wave bioreactor, using the optimal conditions found at small scale. Similar productivities have been obtained when compared to small scale production. The total cell concentration achieved was 1.5×106 cells/mL and final VLP concentration was 5.6×109 VLP/mL, after 72 hours post-transfection. Therefore it was proven than a single bioreactor run at 1L volume could provide enough VLPs to perform a preclinical experiment with ~180 mice (4 immunization/mouse) after purification steps.
1. A serum-free and animal-derived component free culture medium was optimized for HEK 293 cell growth. 2. Transient transfection protocol was optimized for Gag-GFP VLP production. 3. A quantitation method based on fluorescence for Gag-GFP VLPs was developed and validated. 4. Successful scale-up of the optimized process was attained using a Wave bioreactor, enabling preclinical studies to be performed in a short time. 5. Gag-GFP VLPs showed to be of an expected size and morphology consistent with immature HIV-1 particles.
REFERENCES 1. Cervera L, Gutiérrez-Granados S, Martínez M, Blanco J, Gòdia F, Segura MM. Generation of HIV-1 Gag VLPs by transient transfection of HEK 293 suspension cell cultures using an optimized animal-derived component free medium, J Biotechnol. 2013 Jul 20;166(4):152-65. 2. Gutiérrez-Granados S, Cervera L, Gòdia F, Carrillo J, Segura MM. Development and validation of a quantitation assay for GFP-tagged HIV-1 virus-like particles, J Virol Methods. 2013 Oct;193(1):85-95. 3. Validation of Analytical Procedures: Text and Methodology (2005). International Conference Harmonization (ICH). 4. Valley-Omar Z, Meyers AE, Shepard EG, Williamson AL, Rybicki EP. Abrogation of contaminating RNA activity in HIV-1 Gag VLPs. Virol J. 2011 Oct 6;8:462.
ACKNOWLEDGEMENTS We would like to thank Dr. Amine Kamen (NRC, Canada) for helpful discussions about this project and for kindly providing the cGMP HEK 293SF-3F6 cell line. The pGag-GFP plasmid was obtained through the NIH AIDS Reagent Program (Cat #11468). The recombinant insulin used to supplement Freestyle medium was generously provided by FeF Chemicals/Novo Nordisk, Køge, Denmark.
Cell and Tissue Engineering Research Group Departament d’Enginyeria Química, Universitat Autònoma de Barcelona, Edifici Q, 08193 Bellaterra, Barcelona, Spain
OPTIMIZING HEK 293 CELL GROWTH USING DOE
ABSTRACT Robust and scalable Gag VLP production processes are required to generate VLPs for preclinical and eventually clinical testing. Transient gene expression offers a convenient route to accelerate the identification of candidate VLP vaccines as it allows for the production of a large number of product variants with relative ease. Considerable progress has been made in the past several years towards establishing large-scale transient transfection protocols. The human embryonic kidney 293 (HEK 293) is the preferred host system due to the many industrially relevant features this cell line offers including high transfectability, ability to grow in suspension, ability to grow to high cellular densities and adaptation to serum-free culture conditions. HEK 293 mammalian cell cultures were grown in suspension and serum-free conditions in disposable shake flasks and WAVE bioreactors. VLPs were generated by transient transfection using a Gag-GFP plasmid construct and polyethilenimine as transfection reagent. The GFP tagged VLPs were quantified using a spectrofluorometer. An optimized platform for the generation of Gag VLPs in mammalian suspension cultures is described in this work. To facilitate process optimization, a fast and reliable quantitation technique based on fluorescence was developed and validated. An in house culture medium supplemented with non-animal derived components was defined and the different transient transfection parameters were fine-tuned to allow maximum cell growth and VLP production. The great majority of Gag-GFP recovered from cell culture supernatants was shown to be correctly assembled into VLPs of the expected size and morphology. The scalability of this strategy was demonstrated using a 1L WAVE bioreactor The quantity and quality of Gag VLPs generated using the described platform is suitable for pre-clinical studies in mice. By using a GMP-compliant suspension cell line, disposable bioreactors and a culture medium devoid of animal-derived components, a fast translation into clinical trials is envisioned.
A
B
A SIMPLE AND RELIABLE QUANTITATION METHOD FOR FLUORESCENT GAG VLPS
C A
Fluorescence-based method
p24 ELISA
Gag-GFP fusion protein
Gag-GFP fusion protein
7 to 1000 RFU
10 to 300 pg/mL
Precision
~2% CV
~10% CV
Limit of detection
10 ng/mL
10 pg/mL
Time (96 samples)
1.5 h
4h
Price (96 samples)
7€
400 €
BB Specificity Linear range
Box-Behnken model accurately predicts HEK 293 cell densities HEK 293 cell growth as a function of the concentrations of Lipid mixture (X) vs. r-Transferrin (mg/L) (A), Lipid mixture (X) vs. r-Insulin (mg/L) (B) and r-Transferrin (mg/L) vs. r-Insulin (mg/L) (C) based on Box-Behnken experimental results. Optimal concentrations for each supplement were determined using this experimental design (D). D
E
Optimum levels of additives r-Insulin (mg/L)
19.8
r-Transferrin (mg/L)
1.6
Lipid mixture (X)
0.9
RFU: Relative fluorescence units; CV: coefficient of variation.
Growth kinetics of HEK 293 cells in optimized cell culture medium (E) The maximum cell concentration reached was 5.4x106 cells/mL, same value as predicted using the Box-Behnken model. Values presented are the mean ± SD (n=3).
1.5×108 VLPs/immunization
EXPERIMENTAL APPROACH OPTIMIZING GAG VLP PRODUCTION A
B FreeStyle medium
+ N H
n Transient transfection of HEK 293 suspension cultures
VLP harvest
Fluorescent Gag VLP characterization VLP size (TEM, NanoSight®)
VLP protein composition (SDS-PAGE, Western-blot)
VLP morphology (TEM) VLP characterization studies facilitated by fluorescence
B
5 µm
5 µm
Confocal fluorescence microscopy images of HEK 293 producer cells (A) Gag-GFP was accumulated in the vicinity of the plasma membrane where the budding process takes place. (B) Upon staining with Cell Mask™, co-localization of green (Gag-GFP molecules) and red (lipid membrane) fluorescence was observed in yellow. Cell nuclei were stained with Hoechst (blue).
Transmission electron microscopy (TEM) images of Gag-GFP VLPs (A) Negatively stained Gag-GFP VLPs and latex beads used for quantitation purposes (full and empty arrows, respectively). (B) High magnification image showing roughly spherical viral particles surrounded by a lipid envelope and containing electro-dense Gag core. The mean VLP diameter was determined to be 141 ± 22 nm (n=100) which is the expected size of Gag-GFP VLPs as they resemble HIV immature particles which are larger than wild-type HIV-1 virions (4). Gag-GFP VLP size distribution determined by Nanoparticle Tracking Analysis (NTA) VLPs were analyzed by NTA NanoSight® and the size distribution of 3 replicates is shown. The most frequent particle size value (statistical mode) was 149 ± 5 nm.
CONCLUSIONS
Scale up
Plasmid DNA and PEI polyplex formation
A
Fluorescence-based method validation Validation of the developed fluorescence-based quantitation assay was carried out according to ICH guidelines (3). Fluorescence measurements correlated well with p24 concentrations determined by ELISA, as shown by the correlation coefficient (A). All validation parameters met the criteria for analytical method validation (B).
GAG VLP PRODUCTION IN WAVE REACTOR
Generation of Gag-GFP VLPs
VLP CHARACTERIZATION
1L bioreactor C
D Supplemented FreeStyle medium + medium exchange
VLP quantitation (Fluorometry, p24 ELISA, NTA)
Transfection efficiency (Flow citometry)
VLP budding (Fluorescence microscopy, TEM)
750 immunizations
VLP concentration
Fluorescencebased method
p24 ELISA
NTA Nanosight®
VLP (particles/mL)
4.58×109
7.13×109
5.02×109
Mean VLP production was improved using optimal conditions of transient transfection Cell suspensions were transfected with 1 µg of DNA /mL of culture and a DNA to PEI mass ratio of 1:2. HEK 293 transfection efficiency (A, C) and VLP fluorescence intensity (B, D) in cell culture supernatants were measured . Mean values ± SD (n=3). The optimal conditions founded were attained when using supplemented FreeStyle medium with medium exchange performed just before transfection and a cell concentration of 2×106. With this optimization VLP production was improved 2.6-fold.
5.6×109 VLP/mL
VLP production in 1L Wave bioreactor Gag-GFP VLPs were produced in a 1L Wave bioreactor, using the optimal conditions found at small scale. Similar productivities have been obtained when compared to small scale production. The total cell concentration achieved was 1.5×106 cells/mL and final VLP concentration was 5.6×109 VLP/mL, after 72 hours post-transfection. Therefore it was proven than a single bioreactor run at 1L volume could provide enough VLPs to perform a preclinical experiment with ~180 mice (4 immunization/mouse) after purification steps.
1. A serum-free and animal-derived component free culture medium was optimized for HEK 293 cell growth. 2. Transient transfection protocol was optimized for Gag-GFP VLP production. 3. A quantitation method based on fluorescence for Gag-GFP VLPs was developed and validated. 4. Successful scale-up of the optimized process was attained using a Wave bioreactor, enabling preclinical studies to be performed in a short time. 5. Gag-GFP VLPs showed to be of an expected size and morphology consistent with immature HIV-1 particles.
REFERENCES 1. Cervera L, Gutiérrez-Granados S, Martínez M, Blanco J, Gòdia F, Segura MM. Generation of HIV-1 Gag VLPs by transient transfection of HEK 293 suspension cell cultures using an optimized animal-derived component free medium, J Biotechnol. 2013 Jul 20;166(4):152-65. 2. Gutiérrez-Granados S, Cervera L, Gòdia F, Carrillo J, Segura MM. Development and validation of a quantitation assay for GFP-tagged HIV-1 virus-like particles, J Virol Methods. 2013 Oct;193(1):85-95. 3. Validation of Analytical Procedures: Text and Methodology (2005). International Conference Harmonization (ICH). 4. Valley-Omar Z, Meyers AE, Shepard EG, Williamson AL, Rybicki EP. Abrogation of contaminating RNA activity in HIV-1 Gag VLPs. Virol J. 2011 Oct 6;8:462.
ACKNOWLEDGEMENTS We would like to thank Dr. Amine Kamen (NRC, Canada) for helpful discussions about this project and for kindly providing the cGMP HEK 293SF-3F6 cell line. The pGag-GFP plasmid was obtained through the NIH AIDS Reagent Program (Cat #11468). The recombinant insulin used to supplement Freestyle medium was generously provided by FeF Chemicals/Novo Nordisk, Køge, Denmark.
