10974660, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jctb.7105 by Readcube (Labtiva Inc.), Wiley Online Library on [18/01/2023]. 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 Research Article Received: 1 December 2021 Revised: 25 April 2022 Accepted article published: 3 May 2022 Published online in Wiley Online Library: 30 May 2022 (wileyonlinelibrary.com) DOI 10.1002/jctb.7105 Bioprocess characterization of virus-like particle production with the insect cell baculovirus expression system at nanoparticle level Eduard Puente-Massaguer,a*† Irene González-Domínguez,a Florian Strobl,b,c Reingard Grabherr,c Gerald Striedner,c Martí Lecinad and Francesc Gòdiaa Abstract Background. Virus-like particles (VLPs) are a multivalent platform showing great promise for the development of vaccines, gene therapy, diagnostic, and drug delivery approaches. Particularly, HIV-1 Gag VLPs provide a robust and flexible scaffold for the presentation of a variety of antigens. The insect cell baculovirus expression vector system (BEVS) is nowadays one of the refer- ence systems to produce these complex nanoparticles, but information about VLP quality, quantity, stability, as well as cell per- formance is scarce, especially at bioreactor scale. Results. VLPs produced in the reference High Five and Sf9 insect cell lines share similar physicochemical properties, with VLPs produced in Sf9 cells showing lower levels of double stranded DNA and protein contaminants, and a higher degree of VLP assembly. Besides VLPs, other nanoparticle populations are divergently encountered in each cell line. Hi5 supernatants contain a higher load of extracellular vesicles, while Sf9 supernatants exhibit higher concentrations of baculovirus particles. Similar titers are achieved when comparing Gag to Gag-eGFP VLP production, with the size of most of the nanoparticles produced comprised at the 150–250 nm range. Eventually, Gag VLP production in a 2 L stirred tank bioreactor is successfully demonstrated, validating bioprocess transference to the final product candidate. Conclusions. This work provides two potentially valuable strategies for the production of HIV-1 Gag VLPs and a detailed anal- ysis of the different nanoparticle populations produced. © 2022 The Authors. Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). Supporting information may be found in the online version of this article. Keywords: insect cells; virus-like particle; HIV-1; baculovirus; bioreactor; super-resolution fluorescence microscopy INTRODUCTION human papillomavirus, and malaria.8 The versatility of VLPs is The baculovirus expression vector system (BEVS) has become one not restricted to vaccines, with successful results reported in gene of the gold standards for recombinant protein production. Since therapy,9,10 drug delivery,11,12 theragnostic,13 and in vitro its development in the early 1980s, the BEVS has experienced an unprecedented evolution1 and is nowadays considered a robust technology to produce high levels of recombinant proteins in * Correspondence to: E Puente-Massaguer, Departament d'Enginyeria Química, short times.2 A variety of products have been produced with this Biològica i Ambiental, Universitat Autònoma de Barcelona, Barcelona, Spain. system, ranging from very simple recombinant proteins3 to more E-mail: eduard.puente-massaguer@mssm.edu complex products including adeno-associated virus vectors4 and † Present address: Department of Microbiology, Icahn School of Medicine at virus-like particles (VLPs).5 Mount Sinai, 10029, New York, USA VLPs are self-assembling nanoparticles made of one or several structural proteins of the native virus. These nanoparticles closely a Departament d'Enginyeria Química, Biològica i Ambiental, Universitat resemble the original virus structure but, unlike wild-type viruses, Autònoma de Barcelona, Barcelona, Spain are devoid of viral genetic material.6 The main field of application b Austrian Centre of Industrial Biotechnology (acib GmbH), Vienna, Austria of VLPs is in vaccine development since they have shown to elicit robust and broad immune responses that are capable c Department of Biotechnology, University of Natural Resources and Life Sciences, Vienna, Austria of engaging B and T cells.7 Currently, several prophylactic VLP- 2456 based vaccines are commercially available for Hepatitis B, E, d IQS School of Engineering, Universitat Ramón Llull, Barcelona, Spain © 2022 The Authors. Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. 10974660, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jctb.7105 by Readcube (Labtiva Inc.), Wiley Online Library on [18/01/2023]. 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 VLP production with the BEVS in insect cells www.soci.org diagnostic approaches.14 Among the different VLP candidates developed with the Bac-to-Bac BV expression system (Thermo available, Gag-based VLPs have shown great promise in a variety Fisher Scientific).28 Both genes were under the control of the poly- of applications by providing a stable scaffold for membrane pro- hedrin (polh) promoter. tein presentation against infectious diseases,15 as an active cancer Sf9 cells were used for BV amplification by infection at 3 × 106 immunotherapy,16 and for protein delivery.17 cells mL−1 and a multiplicity of infection (MOI) of 0.1. BV contain- The production of enveloped VLPs is generally accomplished in ing supernatants were harvested at 96 h post infection (hpi) at mammalian or insect cell lines18 and through different expression 1000 × g for 5 min. The number of infectious BV particles/mL systems,19,20 with the insect cell/BEVS being generally unri- was measured in Sf9 cells by the plaque assay method in 6-well valled.21 Sf9, a clonal isolate derived from the parental Spodoptera plates (Nunc, Thermo Fisher Scientific).22 frugiperda cell line IPLB-Sf-21-AE, and High Five (Hi5, BTI-TN- 5B1-4) insect cells are generally the choice to produce recombi- Production of Gag-eGFP and Gag VLPs in shake flasks nant products with this system. Sf9 cells are known to achieve Hi5 and Sf9 cells were infected with BV-Gag or BV-Gag-eGFP in higher baculovirus (BV) titers, whereas Hi5 cells are typically used 125 mL Erlenmeyer shake flasks with 15 mL of Sf900III medium for recombinant protein production. Advances in the field of at conditions previously optimized. Briefly, exponentially growing nanoparticle characterization and quantification have opened Hi5 cells were infected with a MOI of 2.5 at 2 × 106 cells mL−1, and the opportunity to characterize the production of complex prod- Gag or Gag-eGFP VLPs produced harvested at 69 hpi by centrifu- ucts at nanoparticle level. Moreover, progress made in the field gation at 1000 × g for 5 min.22 For Sf9 cells, BV infection was per- of cell culture media development in recent years has put the formed with a MOI of 0.01 at 3.7 × 106 cells mL−1, with VLPs basis for the intensification and standardization of recombinant harvested at 80 hpi.23 protein production with the BEVS. In this sense, we recently assessed and optimized the production of Gag-eGFP VLPs with Production of Gag VLPs in bioreactor the BEVS in both insect cell lines cultured in shake flasks.22,23 Bio- A 2 L DASGIP® Bioblock glass bioreactor (Eppendorf, Hamburg, process validation using the final candidate product, its stability, Germany) equipped with three Rushton impellers was used for and the transference to larger volumes are key aspects for the suc- Hi5 and Sf9 cell cultivation in 0.6 L working volume. Aeration cess of any given bioprocess. It has been a long time since the last was performed through a sparger to maintain the dissolved attempts to transfer Gag VLP production with the BEVS to bioreac- oxygen (DO) at 30% oxygen of air saturation. The air flow rate tor scale,24 and advances made in the field provide an excellent was set at 1 L h−1 and temperature at 27 °C. Initial agitation opportunity to explore the production and quality of these conditions were set at 150 rpm for Hi5 cells and 100 rpm for nanoparticles. Sf9 cells. Agitation conditions were automatically adjusted in In this work, the physicochemical properties and stability of cascade control to aeration by the DASware control software HIV-1 VLPs produced in insect cells is initially addressed. Next, (Eppendorf) to maintain the DO setpoint at 30% oxygen of air the quantity and quality of the Gag-eGFP VLPs produced in Hi5 saturation. The pH was fixed at 6.4 for Hi5 cells and 6.2 for Sf9 and Sf9 cells, and its comparison to Gag VLP production is ana- cells and controlled with 20% w/w H3PO4 and 7.5% w/w lyzed at nanoparticle level. Finally, aiming to establish an insect NaHCO3. Antifoam C (Sigma Aldrich, Saint Louis, MO, USA) cell platform for the rapid production of Gag VLPs, a proof of con- was added to the cell culture by pulses to prevent foam cept in bioreactor is conducted. formation. Hi5 and Sf9 cells previously grown in 1 L Erlenmeyer shake flasks (Corning) were inoculated in the bioreactor at a final concentra- MATERIALS AND METHODS tion of 1 × 106 cells mL−1. Hi5 and Sf9 cells were infected with Insect cell lines and culture conditions BV-Gag at the same MOI used for shake flasks, when a viable cell Hi5 cells (cat. no. B85502, Thermo Fisher Scientific, Grand Island, concentration of 2 × 106 and 3.7 × 106 cell mL−1 was attained NY, USA) and Sf9 cells (cat. no. 71104, Merck, Darmstadt, respectively. Gag-VLPs were harvested by centrifugation at Germany) were grown in the low-hydrolysate animal 1000 × g for 5 min at the same conditions used in shake flasks. component-free Sf900III medium (Thermo Fisher Scientific). Hi5 and Sf9 cells were subcultured three times a week at a cell density Flow cytometry analysis of BV infection of 2–4 × 105 cells mL−125 and 4–6 × 105 cells mL−126 in 125 mL The percentage of Gag-eGFP expressing cells was assessed using disposable polycarbonate Erlenmeyer shake flasks (Corning, Steu- a BD FACS Canto II flow cytometer equipped with a 488 and ben, NY, USA), respectively. All cultures were grown in an orbital 635 nm laser configuration (BD Biosciences, San Jose, CA, USA). shaker at 130 rpm (Stuart, Stone, UK) and maintained at 27 °C. Cell For Gag expressing cells, a specific staining protocol was imple- count and viability were measured with the Nucleocounter NC- mented. Shortly, cells were harvested at different times, washed, 3000 (Chemometec, Allerød, Denmark) using acridine orange for and fixed in 4% p-formaldehyde (Sigma) for 10 min at 4 °C. After- cell detection and 40 ,6-diamidino-2-phenylindole (DAPI) wards, cells were washed and permeabilized using 0.1% (v/v) (Chemometec) to quantify non-viable cells. tween-20 (Sigma) for 15 min at room temperature (RT). Cells were washed again and blocked with 10% fetal bovine serum. Recombinant BVs and infection conditions After three washing cycles, cells were incubated with a rabbit The recombinant Autographa californica multicapsid nucleopoly- anti-HIV-1 p24 primary antibody (EnoGene, New York, NY, USA) hedrovirus (AcMNPV) encoding a Rev-independent full-length under mild rotation conditions for 1 h. Cells were then washed HIV-1 gag gene fused in frame to eGFP (Gag-eGFP)27 was con- and incubated with a chicken anti-rabbit IgG secondary conju- structed using the BaculoGold BV expression system gated to Alexa Fluor 467 (Thermo Fisher Scientific) under contin- (BD Biosciences, San Jose, CA, USA). The recombinant AcMNPV uous rotation for 2 h at RT in the dark. Cells were finally washed, encoding the full-length HIV-1 gag gene (GenBank accession resuspended in ice-cold phosphate buffered saline (PBS) solu- 2457 no. K03455.1) codon-optimized for insect cell expression was tion and analyzed in the APC PMT detector of the flow J Chem Technol Biotechnol 2022; 97: 2456–2465 © 2022 The Authors. wileyonlinelibrary.com/jctb Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). 10974660, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jctb.7105 by Readcube (Labtiva Inc.), Wiley Online Library on [18/01/2023]. 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 www.soci.org E Puente-Massaguer et al. cytometer. A total of 2 × 104 cells were analyzed per sample at a equipped with the Huygens deconvolution and LAS X software flow rate of 10 μL min−1. Single cells were gated according to and GPU array (Leica Microsystems, Wetzlar, Germany) at Servei side scatter (SSC-H) versus forward scatter (FSC-A) dot plots and d'Anatomia Patològica from Hospital Sant Joan de Déu.30 Briefly, Gag-eGFP (FITC-A) or Gag (APC-A) positive cells in comparison nanoparticles were stained with 0.1% v/v of CellMask™ and 0.1% to a non-transfected control depending on their mean FITC-A v/v of Hoechst 33342 or 0.05% v/v for Sf9-derived nanoparticles. or APC-A fluorescence intensity, respectively. Data acquisition Stained preparations were loaded onto a microscope slide and and analysis was performed with the BD FACSDIVA software adsorbed to the surface of the cover glass (Linea LAB, Barcelona, v.5.0 (BD Biosciences). Spain) for 30 min at RT. Analysis was conducted with 100× magni- fication (zoom 5) and a line average of 3 and 496 × 496 pixels with Gag-eGFP VLP characterization by ion exchange a HC PL APO CS2 100 X/1.40 OIL objective in the HyVolution2 chromatography mode (Leica). Five fields of 13 sections per each biological tripli- An ÄKTA pure 25 M2 with an S9 sample pump and fraction collec- cate were analyzed. Deconvolution was performed with the SVI tor F9-C (Cytiva, Marlborough, MA, USA) was used. Fifty milliliters Huygens Professional program and the best resolution strategy and 43 mL of clarified cell culture supernatants containing Gag- (Scientific Volume Imaging B.V., Hilversum, the Netherlands). Par- eGFP VLPs produced in Hi5 and Sf9 cells, respectively, were fil- ticle size distribution (PSD) analyses were performed with the 3D tered using a 0.8 μm Millex AA syringe filter (MilliporeSigma, Bed- module package in Imaris 8.2.1 (Bitplane, Oxford Instruments, ford, MA, USA) and loaded into 1 mL radial flow monolith column Zurich, Switzerland) at Servei de Microscòpia from Institut de Neu- (CIMmultus™ QA, Sartorius, Göttingen, Germany). rociències (Universitat Autònoma de Barcelona).20 PSD histo- Mobile phase A and B consisted in 50 mM HEPES, pH 7.2, and grams were created with Microsoft Excel 2016 (Redmond, 50 mM HEPES, 2 M NaCl, pH 7.2, respectively. The column was WA, USA). equilibrated with 50 mM HEPES, 100 mM NaCl, pH 7.2 (5% v/v of buffer B) before loading. After column loading, a washing step Gag-eGFP quantification with 15 column volumes (CV) of equilibration buffer (5% v/v of The supernatants of insect cells infected with the BV-Gag-eGFP buffer B) was conducted. A 100–1000 mM NaCl (5–50% of buffer were recovered by centrifugation at 3000 × g for 5 min. Green B) linear gradient at a flow rate of 1 mL min−1 was used to assess fluorescence was measured with a Cary Eclipse fluorescence spec- the VLP elution profile (50 CV). Samples were loaded into the col- trophotometer (Agilent Technologies, Santa Clara, CA, USA) at RT umn using the sample pump. Conductivity and ultraviolet as follows: ⊗ex = 488 nm (5 nm slit), ⊗em = 500–530 nm (10 nm slit). (UV) absorbance at wavelengths of 280 and 260 nm were moni- Relative fluorescence units were calculated by subtracting fluores- tored using the software Unicorn (Cytiva).29 Total protein concen- cence unit values of fresh Sf900III medium. tration and double stranded DNA (dsDNA) quantification were conducted by the Bradford assay (BioRad Laboratories, Hercules, Gag quantification CA, USA) and QuantiTTM PicoGreen® dsDNA kit (Thermo Fisher An HIV-1 p24 enzyme-linked immunosorbent assay (ELISA) kit Scientific) according to manufacturer's instructions.29 (Sino Biological, Wayne, NJ, USA) was used to quantify the con- centration of Gag polyprotein. Supernatants were harvested by Nanoparticle quantification and characterization centrifugation at 3000 × g for 5 min. Samples were incubated in Nanoparticle tracking analysis SNCR buffer for 10 min at 70 °C, and in 1.5% Triton X-100 for VLP and total nanoparticle concentrations were measured by 10 min at 100 °C. The substrate solution was prepared by dissol- nanoparticle tracking analysis (NTA) using a NanoSight NS300 ving a SIGMAFAST OPD substrate tablet and one urea hydrogen (Malvern Panalytical, Malvern, UK). Fluorescent particles mea- peroxide tablet (MilliporeSigma) at a final concentration of sured were considered as VLPs, whereas all light-scattered parti- 0.4 mg mL−1 in deionized water. An HIV-1 p24 protein standard cles were considered as total nanoparticles. Shortly, samples of known concentration was also included for absolute Gag quan- from harvested supernatants at 3000 × g for 5 min were diluted tification. The reaction was stopped by adding a 625 mM H2SO4 in 0.22 μm-filtered DPBS and continuously injected into the solution after 10 min incubation. The absorbance was measured device chamber through a syringe pump at an average concen- at 492 nm with a reference wavelength at 630 nm in a Tecan Infi- tration of 108 particles mL−1 (20–60 particles frame−1). Videos of nite 200 Pro reader (Tecan, Männedorf, Switzerland). p24 concen- 60 s from independent triplicate measurements were analyzed tration values were corrected according to HIV-1 Gag molecular with the NanoSight NTA 3.2 software (Malvern Panalytical). weight (55 kDa). Flow virometry SDS-PAGE and western blot VLP and total nanoparticle concentrations were assessed by flow Gag and Gag-eGFP VLP containing supernatants were examined cytometry using a CytoFlex LX (Beckman Coulter, Brea, CA, USA). by SDS-PAGE and western blot for HIV-1 p24 (A2-851-100, Icosa- Gating of the different populations was made according to gen, Tartu, Estonia) and GP64 (A2980, Abcam, Cambridge, UK) SSC-A versus FITC-A dot plots and using fresh DPBS and Sf900III detection.31 medium samples as negative controls. Samples from superna- tants harvested at 3000 × g for 5 min were diluted in 0.22 μm- Cryo-transmission electron microscopy filtered DPBS and triplicate measurements from independent BV-Gag infected Sf9 and Hi5 cell supernatants at 80 and 69 hpi, samples were analyzed with the CytExpert 2.3 software respectively, were visualized with a cryo-transmission electron (Beckman Coulter). microscope. Two to three microliters of sample were blotted onto 400 mesh Holey carbon grids (Micro to Nano, Wateringweg, the Super-resolution fluorescence microscopy Netherlands) previously subjected to glow discharge in a PELCO The different nanoparticle populations in Gag and Gag-eGFP easiGlow discharge unit (Ted Pella Inc., Redding, CA, USA). Sam- 2458 supernatants were analyzed with a TCS SP8 confocal microscope ples were subsequently plunged into liquid ethane at −180 °C wileyonlinelibrary.com/jctb © 2022 The Authors. J Chem Technol Biotechnol 2022; 97: 2456–2465 Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). 10974660, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jctb.7105 by Readcube (Labtiva Inc.), Wiley Online Library on [18/01/2023]. 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 VLP production with the BEVS in insect cells www.soci.org Figure 1. Characterization of Gag-eGFP VLPs produced in shake flasks. Elution profiles of infected Hi5 (A) and Sf9 cell (B) supernatants using a monolithic ion exchange column. The black dotted line in the chromatogram indicates conductivity, whereas blue and red lines refer to ultraviolate (UV) absorbance at 280 and 260 nm, respectively. Fluorescence (C), protein (D), and dsDNA distribution (E) in P1, P2, and P3 elution fractions. SDS-PAGE (F) and HIV-1 p24 western blot (G) analysis of P1, P2, and P3 elution fractions. The P1 fraction was diluted 1:2 before loading the SDS-PAGE and western blot. The black arrow corresponds to the band of the full-length Gag-eGFP polyprotein. Conc, concentration; FU, fluorescence units; dsDNA: double-stranded DNA. Figure 2. Stability profiles of Gag-eGFP VLPs in harvested supernatants stored at 27, 4, −20 and −80 °C. (A) Analysis of Gag-eGFP fluorescence by spec- trofluorometry. (B) Analysis of VLP concentration by NTA. The first measurement in each condition was used as the reference to calculate relative percent- ages in subsequent readouts. (C) Size distribution of VLPs measured at day 56 post-freezing of the study [shown as a dashed square in Fig. 1(b)]. 2459 Fluorescent nanoparticles measured were considered as VLPs, whereas light-scattered particles were considered as total nanoparticles. NTA, nanoparticle tracking analysis; VLP, virus-like particle. J Chem Technol Biotechnol 2022; 97: 2456–2465 © 2022 The Authors. wileyonlinelibrary.com/jctb Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). 10974660, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jctb.7105 by Readcube (Labtiva Inc.), Wiley Online Library on [18/01/2023]. 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 www.soci.org E Puente-Massaguer et al. using a Leica EM GP cryo workstation (Leica) and observed in a Amino acid concentrations were determined by HPLC after Jeol JEM-2011 TEM electron microscope operating at 200 kV derivatization in a reversed-phase Eclipse Plus C18 column (Jeol Ltd., Akishima, Tokyo, Japan). During imaging, samples were (Agilent) at 40 °C according to manufacturer's instructions maintained at −173 °C and pictures were taken using a CCD mul- (Agilent). The flow rate was adjusted to 0.64 mL min−1 and two tiscan camera model no. 895 (Gatan Inc., Pleasanton, CA, USA). solvents (solution A and B) were used in the mobile phase. Solu- tion A consisted of 10 mM K2HPO4 and 10 mM K2B4O7, and solu- Analysis of insect cell metabolism tion B of a 45/45/10% v/v/v mix of acetonitrile, methanol and Glucose, lactate, and phosphate concentrations were measured water, respectively. Amino acids were detected at 266/305 nm by ion-exclusion chromatography using a sulfonated polystyrene for fluorenylmethoxycarbonyl derivates and at 450 nm for o- divinyl benzene column (Aminex HPX-87H, Bio-Rad, Hercules, CA, phthalaldehyde derivates. The final amino acid concentration USA) in an Agilent 1200 series HPLC system (Agilent). A 0.01 N was quantified using an internal standard calibration. The stan- H2SO4 solution was used as the mobile phase with a flow rate of dard deviation associated with the measurement of amino acid 0.45 mL min−1. All measurements were performed in an AZURA concentrations was 4 ± 1%. UV/VIS detector (Knauer, Berlin, Germany) with a refractive index detector temperature of 35 °C. The standard deviation of the technique was determined as 0.31% for glucose, 0.26% for lactate, RESULTS AND DISCUSSION and 1.01% for phosphate measurement. The phosphate uptake Physicochemical properties and quality assessment rate was calculated taking into consideration the amount of phos- of HIV-1 VLPs phate present in the medium and also the volume of H3PO4 We have recently proven that the BV infection conditions of Hi522 added for pH control. and Sf9 cells23 can be successfully optimized to produce high Figure 3. Analysis of Gag and Gag-eGFP VLPs. Imaging and assessment of size distribution by SRFM of the different nanoparticle populations in Hi5 (A) and Sf9 (B) supernatants infected with BV-Gag-eGFP. Nanoparticle lipid membranes were stained with CellMask™ (red) and nucleic acids with Hoechst (blue). Imaging and evaluation of size distribution by SRFM of the different nanoparticle populations in Hi5 (C) and Sf9 (D) supernatants infected with BV- Gag. (E) HIV-1 p24 (upper) and GP64 (lower) western blot analysis of Gag-eGFP and Gag VLP productions. The green mark indicates the Gag-eGFP band, 2460 the black mark shows the Gag band, and the white mark illustrates the GP64 band. Cryo-TEM micrographs of Gag VLPs produced in Hi5 (F) and Sf9 cells (G). NA, nucleic acid; NP, nanoparticle; SRFM, super-resolution fluorescence microscopy; VLP, virus-like particle. wileyonlinelibrary.com/jctb © 2022 The Authors. J Chem Technol Biotechnol 2022; 97: 2456–2465 Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). 10974660, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jctb.7105 by Readcube (Labtiva Inc.), Wiley Online Library on [18/01/2023]. 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 VLP production with the BEVS in insect cells www.soci.org HIV-1 VLP titers. However, the physicochemical properties and fluorescence elution pattern between cell lines could be attrib- quality of these nanoparticles produced in both insect cell lines uted to the high ratio of unassembled Gag-eGFP monomer versus is still to be characterized. To this purpose, fluorescently tagged Gag-eGFP VLPs in the Hi5 supernatant,22 since free monomer has VLPs (Gag-eGFP VLPs) were used for an easier sample monitoring. been reported to elute earlier than VLPs.29 Apart from the differ- Different parameters were assessed by ion exchange chromatog- ent content of unassembled Gag-eGFP monomer between both raphy (IEX) in the unprocessed supernatant samples, including cells, similar HIV-1 VLP elution profiles in monolithic columns were Gag-eGFP polyprotein (fluorescence units, FU), total protein con- observed in this work compared to other production platforms, centration (μg mL−1), and dsDNA concentration (ng mL−1). A with HIV-1 VLPs produced in HEK 293, Tnms42 and CHO cells elut- five-fold increase in FU was measured in Hi5 compared to Sf9 ing at conductivities of 27–49, ∼20–45, and 45–90 mS cm−1, supernatants, with Sf9 supernatant samples exhibiting lower respectively.29,32,33 levels of contaminating protein and dsDNA (Fig. 1). The low Hi5 cell viability at harvest might explain the high amounts of dsDNA and proteins encountered in the supernatant. After IEX, there was Stability of HIV-1 VLPs no presence of VLPs in the flowthrough, supporting the complete Product stability is an important parameter to consider towards capture of these nanoparticle by the monolith column (data not extending the shelf-life of the product of interest. This is especially shown). Three different fractions, P1–P3, were pooled in each relevant when it comes to enveloped particulate products such as run according to UV 260 and 280 nm chromatograms. The three VLPs. To this purpose, the supernatant of VLPs produced with the elution peaks were obtained at conductivities of 22, 38, and BEVS was split, aliquoted, and maintained at four different condi- 51 mS cm−1 in the Hi5 supernatant [Fig. 1(A)], with similar conduc- tions (27, 4, −20 and −80 °C) for 2 months. Gag-eGFP VLPs were tivity values of 25, 31 and 51 mS cm−1 for P1–P3 in the Sf9 super- used with the aim to discriminate VLPs from other nanoparticle natant [Fig. 1(B)]. Most of the protein present in both Hi5 (80%) populations co-expressed with the BEVS.27 Different analytical and Sf9 supernatants (50%) eluted at lower salt concentrations techniques, including spectrofluorometry and NTA were [Fig. 1(C)], whereas dsDNA eluted in P1 in the Hi5 and in P3 in employed (Fig. 2). VLP fluorescence was generally maintained the Sf9 supernatant, respectively [Fig. 1(D)]. The dsDNA elution over the 2-month period independently of the storage conditions profile of the Hi5 supernatant sample was comparable to that used [Fig. 2(A)]. However, VLP characterization in native condi- observed in Tnms42 cells,32 while it was similar to the elution pro- tions by NTA revealed that the structural integrity of VLPs was file measured in CHO cells for the Sf9 supernatant.33 Considering only preserved at 4 and −80 °C [Fig. 2(B)], suggesting that VLPs the different level of contaminants in each unpurified superna- stored at 27 and −20 °C were probably disassembling. Enhanced tant, as well as the strong positive charge of the quaternary amine protease activity could explain the instability of VLPs stored at monolithic column, a competition for binding between negative 27 °C, while particle disruption due to external and internal ice charged dsDNA and VLPs could be the reason behind the differ- formation around the VLP envelope might explain the low stabil- ences in dsDNA elution profiles. ity at −20 °C.34,35 Additionally, changes in the morphology were The presence of Gag-eGFP VLPs was analyzed by spectrofluo- detected in all conditions except −80 °C [Fig. 2(C)]. So far, it is still rometry (FU), SDS-PAGE, and HIV-1 p24 western blot in each not well known whether variations in VLP morphology can alter pooled fraction [Fig. 1(E)–(G)]. The Gag-eGFP polyprotein was their immunogenicity profile. In any case, this data indicates that detected in all fractions analyzed, most of it eluting at a low con- Gag VLPs produced with the BEVS can be safely stored at 4 and ductivity in the Hi5 supernatant (93%), and at a low-intermediate −80 °C for mid to long term purposes without the need of further salt concentration in the Sf9 supernatant (60%). Differences in the purification. Table 1. Quantification of Gag and Gag-eGFP VLP production titers and baculovirus particles in Hi5 and Sf9 cells infected with the BEVS in shake flasks. Nanoparticle quantification was performed by NTA and flow virometry, and i baculovirus particles were measured by the plaque assay method Condition Quantification method Product Fluorescent particles (109 particles mL−1) Total particles (109 particles mL−1) Hi5 NTA Gag n.a. 495 ± 60 Gag-eGFP 34 ± 2 426 ± 61 Flow virometry Gag n.a. 5.9 ± 0.5 Gag-eGFP 2.2 ± 0.1 9.5 ± 0.3 Sf9 NTA Gag n.a. 128 ± 24 Gag-eGFP 37 ± 9 181 ± 15 Flow virometry Gag n.a. 5.9 ± 0.3 Gag-eGFP 3.3 ± 0.3 8.2 ± 0.9 Sf900III medium NTA n.a. n.a. 44.8 ± 1.9 Flow virometry 0.1 ± 0.0 Condition Product IBV particles (107 pfu mL−1) Hi5 Gag 10.0 ± 2.7 Gag-eGFP 6.0 ± 1.4 Sf9 Gag 31.0 ± 13.5 Gag-eGFP 76.7 ± 51.3 2461 IBV, infectious baculovirus particle; n.a., not available. J Chem Technol Biotechnol 2022; 97: 2456–2465 © 2022 The Authors. wileyonlinelibrary.com/jctb Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). 10974660, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jctb.7105 by Readcube (Labtiva Inc.), Wiley Online Library on [18/01/2023]. 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 www.soci.org E Puente-Massaguer et al. Comparison of Gag-eGFP and Gag VLPs concentration of BV particles in Sf9 over Hi5 supernatants Gag tagging with the fluorescent reporter protein eGFP is a useful (Table 1) given that these cells generally produce higher amounts strategy in bioprocess development since discrimination of BVs. This is in agreement with the higher frequency of nucleic between other nanoparticle populations and VLPs is possible. This acid detection in the 300–500 nm range since BVs contain large allows a better control and understanding of nanoparticle-based genomes up to 180 kb.36 production platforms such as VLP expression with the BEVS. Nev- A similar size distribution was observed in Hi5 Gag VLP superna- ertheless, once the bioprocess has been defined, the production tants, with most of the nanoparticles located in the 150–250 nm of the final candidate without the fluorescent tag has to be size range [Fig. 3(C) and (D)]. Specific detection of Gag-eGFP and assessed. Super-resolution fluorescence microscopy, a novel tech- Gag in Hi5 and Sf9 supernatants was also confirmed by western nique to characterize different nanoparticle populations blot, while the presence of the GP64 BV-associated protein also simultaneously,20 was applied to characterize Gag-eGFP and indicated the co-expression of BVs [Fig. 3(E)]. As Gag VLPs were Gag VLP productions. Most of the Gag-eGFP VLPs produced in not tagged, VLPs could not be differentiated from other nanopar- both insect cell lines were detected at the 150–250 nm range, as ticles, specifically extracellular vesicles (EVs), which fall in the same expected [Fig. 3(A) and (B)]. However, Sf9-derived Gag-eGFP VLPs size range and have been described to be co-expressed with VLPs proved to be more heterogeneous with a fraction of these VLPs in animal cell lines.30 These nanoparticles are generally produced with sizes in the 300–400 nm range. A higher number of nanopar- at higher levels than VLPs, representing from 4 to 13-fold in Hi5 ticles was also measured at the 300–500 nm range in Sf9 superna- supernatants and from 2 to 5-fold in Sf9 supernatants,26 with var- tants. This could be related to the three to ten-fold higher iability attributed to the quantification methodology employed Figure 4. Transference of HIV-1 VLP production from shake flasks to bioreactor. Cell growth and viability profiles of Hi5 (A) and Sf9 cells (B). The black arrow indicates the time of recombinant baculovirus infection. Evolution of Gag and Gag-eGFP expression in Hi5 (C) and Sf9 cells (D) by flow cytometry after infection with recombinant baculoviruses. (E) Quantification of Gag production by enzyme-linked immunosorbent assay in the supernatant of infected insect cells cultured in bioreactor at the time of harvest. The HIV-1 p24 protein was included as standard of known concentration. (F) Measurement of total nanoparticles by NTA in the supernatant of infected insect cells cultured in bioreactor at the time of harvest. The concentration 2462 of infectious baculovirus particles in these conditions was also determined by the plaque assay method. IBV, infectious baculovirus; NTA, nanoparticle tracking analysis; R, bioreactor; SF, shake flask. wileyonlinelibrary.com/jctb © 2022 The Authors. J Chem Technol Biotechnol 2022; 97: 2456–2465 Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). 10974660, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jctb.7105 by Readcube (Labtiva Inc.), Wiley Online Library on [18/01/2023]. 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 VLP production with the BEVS in insect cells www.soci.org (Table 1). The higher load of BV particles in the Sf9 supernatant observed in Hi5 cells. After infection, an increase in stirring speed was also appreciated in Gag VLP preparations [Fig. 3(D)], with a was observed in Hi5 cells, peaking at ∼24 hpi, which could be larger nanoparticle population frequency in the 300–500 nm related to a higher oxygen demand owing to the BV infection pro- range compared to the Hi5 supernatant. Of note, it is possible that cess since cell growth was completely arrested.40 As for Sf9 cells, the concentration of BV particles was higher than the titers the maximum oxygen demand was recorded at ∼40 hpi [Fig. 5 obtained by the plaque assay method due to the existence of (A) and (B)], coinciding with the peak in viable cell concentration defective BV particles.37 The presence of nucleic acids in particles and complete infection. In these conditions, Hi5 and Sf9 cells of similar sizes to VLPs was detected in Gag and Gag-eGFP prepa- yielded 1.86 and 1.28 mg L−1 of Gag polyprotein, with a higher rations, possibly indicating that nanoparticles within this size amount of contaminant BV particles in the latter [Fig. 4(E) and range, VLPs, EVs or both, might contain or be associated with (F)]. VLP titers in both platforms were higher than those achieved nucleotide molecules.38,39 Further studies are needed to discrim- in insect cells by stable gene expression,41-43 and transient transfec- inate which nanoparticle populations are subjected to incorpo- tion of plasmid DNA.44,45 On the other hand, a larger load of con- rate or associate to nucleic acids. taminant EVs in the same size range as VLPs was observed in Hi5 Gag VLP supernatants from both insect cell lines were analyzed cell cultures. Despite several advances to separate these contami- by cryo-transmission electron microscopy (cryo-TEM). The native nant species from Gag VLPs have been accomplished in the last structure of Gag VLPs could be observed but differentiation years,32,46 further research is still required. Moreover, it is not well of VLPs from EVs was more difficult in the Hi5 supernatant known how these specimens might impact the final application [Fig. 3(F) and (G)]. This is possibly due to the higher load of EVs of Gag VLPs, either by boosting their immunogenicity47 or by co-expressed in Hi5 cells and also to the larger amounts of reducing it.48 If BV particles negatively impact the final application proteins (Fig. 2), which increase the background signal and reduce of Gag VLPs as a pharmaceutical product, Hi5 cells might be a bet- the quality of the imaging. Compared to previous studies with Gag- ter option for VLP production. If this is not the case, Sf9 cells might eGFP VLPs, Gag VLPs proved to be less heterogenous and with a have an advantage since the presence of contaminating nucleic higher degree of internal arrangement.30 The absence of the eGFP acids, host cell proteins and unassembled Gag, and EVs is lower, fusion protein might result in the formation of more structured simplifying their purification. In any case, both platforms achieve nanoparticles closely resembling immature HIV-1 virions. superior Gag VLP yields in comparison to other systems employed to produce these nanoparticles at bioreactor scale,49-51 and a three- Gag VLP production in bioreactor to four-fold increase with respect to the most recent optimization The production of Gag VLPs with the insect cell/BEVS was assessed study for Gag VLP production with the insect cell/BEVS.52 Assuming in a stirred tank bioreactor to test the feasibility of this platform for a mouse is immunized with two doses of 5 μg of Gag VLPs in a large scale VLP manufacture. Similar cell growth and infection prime-boost regimen16 and not considering losses in process puri- kinetics were measured between shake flasks and bioreactors for fication, a 1 L bioreactor culture of Sf9 or Hi5 cells would enable to Gag as well as Gag-eGFP production, indicating the successful immunize around 100–200 mice, respectively. transference of the conditions optimized in shake flasks to bioreac- The analysis of cell metabolism is an important aspect to consider tor scale [Fig. 4(A)–(D)]. The higher MOI used in Hi5 cells completely in bioprocess development since strategies for process intensifica- arrested cell growth after infection, whereas Sf9 cells continued to tion can be implemented. To this end, the metabolism of both grow for the next 48 hpi. The same reason applies to the earlier insect cell lines cultured at bioreactor scale after BV infection was decrease in cell viability and also the faster infection kinetics analyzed. Hi5 cells consumed larger levels of all the different Figure 5. Analysis of bioreactor parameters with the insect cell/BEVS. Stirring speed, dissolved oxygen concentration, pH, and volume of acid/base added in Hi5 (A,C) and Sf9 cell (B,D) cultures. Black arrows indicate the time point of recombinant baculovirus infection. Profiles of glucose, lactate, phos- 2463 phate (E,F) and main amino acids produced/consumed (G,H) in Hi5 and Sf9 cultures, respectively. Ala, alanine; Asn, asparagine; Asp, aspartic acid; DO, dis- solved oxygen; Gln, glutamine; Glu, glutamic acid; RPM, revolutions per minute; Ser, serine. J Chem Technol Biotechnol 2022; 97: 2456–2465 © 2022 The Authors. wileyonlinelibrary.com/jctb Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). 10974660, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jctb.7105 by Readcube (Labtiva Inc.), Wiley Online Library on [18/01/2023]. 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 www.soci.org E Puente-Massaguer et al. metabolites assessed in comparison to Sf9 cells [Fig. 5(E)–(H)], with ACKNOWLEDGEMENTS glucose, asparagine, and glutamine as the most highly consumed The authors thank Alois Jungbauer (University of Natural (Supporting Information, Fig. S1). Asparagine is reported as the Resources and Life Sciences, Vienna, Austria) for facilitating the principal nitrogen source for Hi5 cells and its uptake rate can be access to ÄKTA pure system, and Paula Alves (Instituto de Biologia even higher than glucose.53 After BV infection, asparagine was Experimental e Tecnológica, Oeiras, Portugal) for providing the not consumed as rapid as previously observed,41 possibly due to Hi5 cell line. We also appreciate the support of Miriam Klausberger the low concentration in the medium at the time of infection, on the protocol for Gag staining, Krisztina Koczka for providing resulting in complete asparagine exhaustion between 24–48 hpi the BV-Gag stock, Sahar Masoumeh Ghorbanpour with ELISA and coinciding with the decline in cell viability. Sf9 cells exhibited quantification, Martí de Cabo (Servei de Microscòpia, Universitat a more balanced metabolism with reduced uptake fluxes of the dif- Autònoma de Barcelona) with cryo-TEM, Mónica Roldán (Unitat ferent metabolites analyzed (Supporting Information, Fig. S1). Glu- de Microscòpia Confocal, Servei d'anatomia Patològica, Institut cose, the principal carbon source, and glutamine, an important Pediàtric de Malalties Rares, Hospital Sant Joan de Déu) with nitrogen source in insect cell cultures, were the main compounds SRFM, and Jorge Fomaro and Ángel Calvache (Beckman Coulter) consumed by Sf9 cells. None of them was limiting by the end of for facilitating the access to the CytoFlex LX. Eduard Puente- the production phase, though the low concentration of glutamine Massaguer (FPU15/03577) and Irene González-Domínguez in the medium (∼1 mM) could have possibly influenced its uptake (FPU16/02555) are recipients of FPU grants from Ministerio de by Sf9 cells. Interestingly, serine was depleted at 48 hpi, coinciding Educación, Cultura y Deporte of Spain. The research group is rec- with the peak of maximum viable cell concentration and complete ognized as 2017 SGR 898 by Generalitat de Catalunya. BV infection. Thus, asparagine and serine supplementation at the time of infection could be an option to extend the productive time of Hi5 and Sf9 cells, respectively, and further increase VLP yields. SUPPORTING INFORMATION The control of bioprocess parameters in the bioreactor was suc- Supporting information may be found in the online version of this cessfully achieved in both insect cell lines. However, differences in article. the behavior of Hi5 and Sf9 cells were observed [Fig. 5(C) and (D)]. The addition of sodium bicarbonate to the Sf9 cell culture by the software control loop at the beginning of the bioreactor opera- tion indicated the acidification of the culture medium. The accu- REFERENCES mulation of organic acids derived from cell metabolism such as 1 Stolt-Bergner P, Benda C, Bergbrede T, Besir H, Celie PHN, Chang C et al., Baculovirus-driven protein expression in insect cells: a bench- glutamic and aspartic acid, both with pKa values below the cell marking study. J Struct Biol 203:71–80 (2018). culture pH, and CO2 production from cell respiration could be 2 Contreras-Gómez A, Sánchez-Mirón A, García-Camacho F, Molina- the reasons behind medium acidification. On the contrary, the Grima E and Chisti Y, Protein production using the baculovirus- addition of phosphoric acid to the Hi5 cell culture was detected. insect cell expression system. Biotechnol Prog 30:1–18 (2014). 3 Weber W and Fussenegger M, Insect Cell-Based Recombinant Protein Ammonia formation, a by-product of insect cell metabolism, Production. Springer Heidelberg, Berlin, pp. 263–277 (2009). could explain medium basification since significantly higher 4 Merten OW and Gaillet B, Viral vectors for gene therapy and gene mod- asparagine and glutamine uptake rates were measured in this cell ification approaches. Biochem Eng J 108:98–115 (2016). line.54 Alanine was also produced in both bioreactor cultures 5 Dai X, Jian C, Na L, Wang X, Dai Y and Li D, Production and characteri- peaking in the 15–20 mM range. The formation of this amino acid zation of Hantaan virus-like particles from baculovirus expression system. Biochem Eng J 152:107373 (2019). is a feature observed in animal cell metabolism under glucose 6 Ding X, Liu D, Booth G, Gao W and Lu Y, Virus-like particle engineering: from excess conditions.55 Lactate, another by-product of insect cell rational design to versatile applications. Biotechnol J 13:1–7 (2018). metabolism, was not produced after BV infection, which is proba- 7 Zepeda-Cervantes J, Ramírez-Jarquín JO and Vaca L, Interaction bly associated to pH and dissolved oxygen concentration mainte- between virus-like particles (VLPs) and pattern recognition recep- tors (PRRs) from dendritic cells (DCs): toward better engineering of nance along the culture.56 VLPs. Front Immunol 11:1100 (2020). 8 Nooraei S, Bahrulolum H, Hoseini ZS, Katalani C, Hajizade A, Easton AJ et al., Virus-like particles: preparation, immunogenicity and their roles as nanovaccines and drug nanocarriers. J Nanobiotechnol 19:59 (2021). CONCLUSIONS 9 Biddlecome A, Habte HH, McGrath KM, Sambanthamoorthy S, Wurm M, Sykora MM et al., Delivery of self-amplifying RNA vaccines in in vitro This study provides a detailed analysis of the insect cell/BEVS reconstituted virus-like particles. PLoS One 14:e0215031 (2019). and shows the versatility of this system to produce HIV-1 VLPs. 10 Prel A, Caval V, Gayon R, Ravassard P, Duthoit C, Payen E et al., Highly VLPs produced in the reference Hi5 and Sf9 insect cell lines efficient in vitro and in vivo delivery of functional RNAs using new exhibit similar physicochemical properties and are stable when versatile MS2-chimeric retrovirus-like particles. Mol Ther Methods stored at −80 °C for 2 months with little impact on their mor- Clin Dev 2:15039 (2015). 11 Finbloom J, Aanei I, Bernard J, Klass S, Elledge S, Han K et al., Evaluation phology. Gag VLP titers are similar to those obtained for Gag- of three morphologically distinct virus-like particles as nanocarriers eGFP VLPs, with nanoparticle sizes in the 150–250 nm range for convection-enhanced drug delivery to glioblastoma. Nanomater- closely resembling immature HIV-1 virions. A high degree of ials. 8:1007 (2018). heterogeneity and complexity is detected in VLP productions, 12 Peyret H, Gehin A, Thuenemann EC, Blond D, El Turabi A, Beales L et al., Tandem fusion of hepatitis B core antigen allows assembly of virus- with large concentrations of EVs, dsDNA and total protein con- like particles in bacteria and plants with enhanced capacity to tent in Hi5 supernatants, whereas high loads of BV particles are accommodate foreign proteins. PLoS One. 10:e0120751 (2015). encountered in Sf9 supernatants. The successful transference 13 Enomoto T, Kawano M, Fukuda H, Sawada W, Inoue T, Haw KC et al., of Gag VLP production from shake flasks to stirred tank bioreac- Viral protein-coating of magnetic nanoparticles using simian virus tor is demonstrated, providing two different strategies to pro- 40 VP1. J Biotechnol 167:8–15 (2013). 14 Theillet G, Martinez J, Steinbrugger C, Lavillette D, Coutard B, duce these complex nanoparticles in large volumes with 2464 Papageorgiou N et al., Comparative study of chikungunya virus-like controlled operational conditions. particles and pseudotyped-particles used for serological detection wileyonlinelibrary.com/jctb © 2022 The Authors. J Chem Technol Biotechnol 2022; 97: 2456–2465 Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI). 10974660, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/jctb.7105 by Readcube (Labtiva Inc.), Wiley Online Library on [18/01/2023]. 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 VLP production with the BEVS in insect cells www.soci.org of specific immunoglobulin M. Virology 529:195–204 (2019). https:// 35 González-Domínguez I, Lorenzo E, Bernier A, Cervera L, Gòdia F and doi.org/10.1016/j.virol.2019.01.027. Kamen A, A four-step purification process for gag vlps: from culture 15 Fontana D, Garay E, Cevera L, Kratje R, Prieto C and Gòdia F, Chimeric supernatant to high-purity lyophilized particles. Vaccine 9:1154 (2021). vlps based on hiv-1 gag and a fusion rabies glycoprotein induce spe- 36 van Oers M and Vlak J, Baculovirus genomics. Curr Drug Targets 8: cific antibodies against rabies and foot-and-mouth disease virus. 1051–1068 (2007). Vaccine 9:251 (2021). 37 Wickham TJ, Davis T, Granados RR, Hammer DA, Shuler ML and 16 Nika L, Cuadrado-Castano S, Arunkumar GA, Grünwald-Gruber C, Wood HA, Baculovirus defective interfering particles are responsible McMahon M, Koczka K et al., An HER2-displaying virus-like particle for variations in recombinant protein production as a function of vaccine protects from challenge with mammary carcinoma cells in multiplicity of infection. Biotechnol Lett 13:483–488 (1991). a mouse model. Vaccine 7:1–19 (2019). 38 Valley-Omar Z, Meyers AE, Shephard EG, Williamson A-L and 17 Kaczmarczyk SJ, Sitaraman K, Young HA, Hughes SH and Chatterjee DK, Rybicki EP, Abrogation of contaminating RNA activity in HIV-1 gag Protein delivery using engineered virus-like particles. Proc Natl Acad VLPs. Virol J 8:462 (2011). Sci U S A 108:16998–17003 (2011). 39 Fatima F and Nawaz M, Vesiculated long non-coding RNAs: offshore 18 Lavado-García J, Jorge I, Boix-Besora A, Vázquez J, Gòdia F and packages deciphering trans-regulation between cells, cancer pro- Cervera L, Characterization of HIV-1 virus-like particles and determi- gression and resistance to therapies. Noncoding RNA 3:10 (2017). nation of gag stoichiometry for different production platforms. Bio- 40 Lecina M, Soley A, Gràcia J, Espunya E, Lázaro B, Cairó JJ et al., Applica- technol Bioeng 118:2660–2675 (2021). tion of on-line OUR measurements to detect actions points to 19 Alvim RGF, Itabaiana I and Castilho LR, Zika virus-like particles (VLPs): improve baculovirus-insect cell cultures in bioreactors. J Biotechnol stable cell lines and continuous perfusion processes as a new poten- 125:385–394 (2006). tial vaccine manufacturing platform. Vaccine 37:6970–6977 (2019). 41 Puente-Massaguer E, Grau-Garcia P, Strobl F, Grabherr R, Striedner G, 20 González-Domínguez I, Puente-Massaguer E, Cervera L and Gòdia F, Lecina M et al., Accelerating HIV-1 VLP production using stable high Quantification of the HIV-1 virus-like particle production process five insect cell pools. Biotechnol J 16:1–12 (2021). by super-resolution imaging: from VLP budding to nanoparticle 42 Puente-Massaguer E, Grau-Garcia P, Strobl F, Grabherr R, Striedner G, analysis. Biotechnol Bioeng 117:1929–1945 (2020). Lecina M et al., Stable Sf9 cell pools as a system for rapid HIV-1 21 Puente-Massaguer E, Cajamarca-Berrezueta B, Volart A, González- virus-like particle production. J Chem Technol Biotechnol 96:3388– Domínguez I and Gòdia F, Transduction of HEK293 cells with Bac- 3397 (2021). Mam baculovirus is an efficient system for the production of HIV-1 43 Vidigal J, Fernandes B, Dias MM, Patrone M, Roldão A, Carrondo MJT virus-like particles. Viruses 14:636 (2022). et al., RMCE-based insect cell platform to produce membrane pro- 22 Puente-Massaguer E, Lecina M and Gòdia F, Integrating nanoparticle teins captured on HIV-1 gag virus-like particles. Appl Microbiol Bio- quantification and statistical design of experiments for efficient technol 102:655–666 (2018). HIV-1 virus-like particle production in high five cells. Appl Microbiol 44 Puente-Massaguer E, Gòdia F and Lecina M, Development of a non- Biotechnol 104:1569–1582 (2020). viral platform for rapid virus-like particle production in Sf9 cells. 23 Puente-Massaguer E, Lecina M and Gòdia F, Application of advanced J Biotechnol 322:43–53 (2020). quantification techniques in nanoparticle-based vaccine develop- 45 Puente-Massaguer E, Strobl F, Grabherr R, Striedner G, Lecina M and ment with the Sf9 cell baculovirus expression system. Vaccine 38: Gòdia F, PEI-mediated transient transfection of high five cells at biore- 1849–1859 (2020). actor scale for HIV-1 VLP production. Nanomaterials 10:1–16 (2020). 24 Cruz PE, Cunha A, Peixoto CC, Clemente J, Moreira JL and 46 Reiter K, Aguilar PP, Wetter V, Steppert P, Tover A and Jungbauer A, Separa- Carrondo MJT, Optimization of the production of virus-like particles tion of virus-like particles and extracellular vesicles by flow-through and in insect cells. Biotechnol Bioeng 60:408–418 (1998). heparin affinity chromatography. J Chromatogr A 1588:77–84 (2019). 25 Puente-Massaguer E, Lecina M and Gòdia F, Nanoscale characteriza- 47 Heinimäki S, Tamminen K, Malm M, Vesikari T and Blazevic V, Live bacu- tion coupled to multi-parametric optimization of Hi5 cell tran- lovirus acts as a strong B and T cell adjuvant for monomeric and olig- sient gene expression. Appl Microbiol Biotechnol 102:10495– omeric protein antigens. Virology 511:114–122 (2017). 10510 (2018). 48 Thompson CM, Petiot E, Mullick A, Aucoin MG, Henry O and Kamen A, 26 Puente-Massaguer E, Saccardo P, Ferrer-Miralles N, Lecina M and Critical assessment of influenza VLP production in Sf9 and HEK293 Gòdia F, Coupling microscopy and flow cytometry for a comprehen- expression systems. BMC Biotechnol 15:1–12 (2015). sive characterization of nanoparticle production in insect cells. Cyto- 49 Sakuragi S, Goto T, Sano K and Morikawa Y, HIV type 1 Gag virus-like metry A 97:921–932 (2020). particle budding from spheroplasts of saccharomyces cerevisiae. 27 Hermida-Matsumoto L and Resh MD, Localization of human immuno- Proc Natl Acad Sci U S A 99:7956–7961 (2002). deficiency virus type 1 Gag and Env at the plasma membrane by 50 Fernandes B, Vidigal J, Correia R, Carrondo MJT, Alves PM, Teixeira AP confocal imaging. J Virol 74:8670–8679 (2000). et al., Adaptive laboratory evolution of stable insect cell lines for 28 Nika L, Wallner J, Palmberger D, Koczka K, Vorauer-Uhl K and improved HIV-Gag VLPs production. J Biotechnol 307:139–147 (2020). Grabherr R, Expression of full-length HER2 protein in Sf9 insect cells 51 Fernandes B, Correia R, Sousa M, Carrondo MJT, Alves PM and and its presentation on the surface of budded virus-like particles. Roldão A, Integrating high cell density cultures with adapted labora- Protein Expr Purif 136:27–38 (2017). tory evolution for improved Gag-HA virus-like particles production 29 Pereira Aguilar P, González-Domínguez I, Schneider TA, Gòdia F, in stable insect cell lines. Biotechnol Bioeng 118:2536–2547 (2021). Cervera L and Jungbauer A, At-line multi-angle light scattering 52 Pillay S, Meyers A, Williamson AL and Rybicki EP, Optimization of chi- detector for faster process development in enveloped virus-like par- meric HIV-1 virus-like particle production in a baculovirus-insect cell ticle purification. J Sep Sci 42:2640–2649 (2019). expression system. Biotechnol Prog 25:1153–1160 (2009). 30 González-Domínguez I, Puente-Massaguer E, Cervera L and Gòdia F, 53 Monteiro F, Bernal V and Alves PM, The role of host cell physiology in Quality assessment of virus-like particles at single particle level: a the productivity of the baculovirus-insect cell system: fluxome anal- comparative study. Viruses 12:223–235 (2020). ysis of Trichoplusia ni and Spodoptera frugiperda cell lines. Biotech- 31 Steppert P, Burgstaller D, Klausberger M, Kramberger P, Tover A, nol Bioeng 114:674–684 (2017). Berger E et al., Separation of HIV-1 gag virus-like particles from vesic- 54 Rhiel M, Mitchell-Logean CM and Murhammer DW, Comparison of Tri- ular particles impurities by hydroxyl-functionalized monoliths. J Sep choplusia ni BTI-Tn-5b1-4 (high five) and Spodoptera frugiperda Sf-9 Sci 40:979–990 (2017). insect cell line metabolism in suspension cultures. Biotechnol Bioeng 32 Reiter K, Aguilar PP, Grammelhofer D, Joseph J, Steppert P and 55:909–920 (1997). Jungbauer A, Separation of influenza virus-like particles from baculovirus 55 Benslimane C, Elias CB, Hawari J and Kamen A, Insights into the central by polymer grafted anion-exchanger. J. Sep. Sci 43:2270-2278 (2020). metabolism of Spodoptera frugiperda (Sf-9) and Trichoplusia ni BTI- 33 Steppert P, Burgstaller D, Klausberger M, Berger E, Aguilar PP, Schneider TA Tn-5B1-4 (Tn-5) insect cells by radiolabeling studies. Biotechnol Prog et al., Purification of HIV-1 gag virus-like particles and separation of other 21:78–86 (2005). extracellular particles. J Chromatogr A 1455:93–101 (2016). 56 Strobl F, Ghorbanpour SM, Palmberger D and Striedner G, Evaluation 34 Lynch A, Meyers AE, Williamson A-LL and Rybicki EP, Stability studies of of screening platforms for virus-like particle production with the HIV-1 Pr55gag virus-like particles made in insect cells after storage in baculovirus expression vector system in insect cells. Sci Rep 10: various formulation media. Virol J 9:210 (2012). 1065 (2020). 2465 J Chem Technol Biotechnol 2022; 97: 2456–2465 © 2022 The Authors. wileyonlinelibrary.com/jctb Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI).