Applied Microbiology and Biotechnology (2020) 104:1569–1582 https://doi.org/10.1007/s00253-019-10319-x BIOTECHNOLOGICAL PRODUCTS AND PROCESS ENGINEERING Integrating nanoparticle quantification and statistical design of experiments for efficient HIV-1 virus-like particle production in High Five cells Eduard Puente-Massaguer 1 & Martí Lecina 2 & Francesc Gòdia 1 Received: 17 October 2019 / Revised: 4 December 2019 / Accepted: 13 December 2019 / Published online: 6 January 2020 # Springer-Verlag GmbH Germany, part of Springer Nature 2020 Abstract The nature of enveloped virus-like particles (VLPs) has triggered high interest in their application to different research fields, including vaccine development. The baculovirus expression vector system (BEVS) has been used as an efficient platform for obtaining large amounts of these complex nanoparticles. To date, most of the studies dealing with VLP production by recom- binant baculovirus infection utilize indirect detection or quantification techniques that hinder the appropriate characterization of the process and product. Here, we propose the application of cutting-edge quantification methodologies in combination with advanced statistical designs to exploit the full potential of the High Five/BEVS as a platform to produce HIV-1 Gag VLPs. The synergies between CCI, MOI, and TOH were studied using a response surface methodology approach on four different response functions: baculovirus infection, VLP production, VLP assembly, and VLP productivity. TOH and MOI proved to be the major influencing factors in contrast with previous reported data. Interestingly, a remarkable competition between Gag VLP production and non-assembled Gag was detected. Also, the use of nanoparticle tracking analysis and flow virometry revealed the existence of remarkable quantities of extracellular vesicles. The different responses of the study were combined to determine two global optimum conditions, one aiming to maximize the VLP titer (quantity) and the second aiming to find a compromise between VLP yield and the ratio of assembled VLPs (quality). This study provides a valuable approach to optimize VLP production and demonstrates that the High Five/BEVS can support mass production of Gag VLPs and potentially other complex nanoparticles. Keywords High Five cells . Virus-like particle . Response surface methodology . Multiple-criteria decision analysis . Nanoparticle Introduction since they do not contain genetic material from the virus itself (Andersson et al. 2018). Among the different proteins able to Virus-like particles (VLPs) have emerged as promising nano- self-assemble as VLPs, the group-specific antigen (Gag) particles, particularly in the field of new recombinant vaccines polyprotein has demonstrated a wide applicability in different (Mohsen et al. 2017). Compared with other conventional vac- research areas. These include their use as immunogens cines such as live-attenuated or inactivated viruses, VLPs can (Young and Ross 2006), drug delivery vehicles (Deo et al. elicit potent antibody and T cell responses in a safe manner 2015), scaffolds for surface antigen presentation (Vidigal et al. 2018), and research surrogates (Buonaguro et al. Electronic supplementary material The online version of this article 2013). A common need in all these applications is the under- (https://doi.org/10.1007/s00253-019-10319-x) contains supplementary standing of the production system itself, which is paramount material, which is available to authorized users. to define the most adequate conditions for VLP production. The baculovirus species Autographa californica multiple * Eduard Puente-Massaguer eduard.puente@uab.cat nucleopolyhedrovirus (AcMNPV) has been successfully used for the production of a variety of VLP types. The ease of 1 handling, the high productivities associated and the capacity Departament d’Enginyeria Química, Biològica i Ambiental, Universitat Autònoma de Barcelona, Cerdanyola del Vallès, to accommodate large quantities of DNA have made the Barcelona, Spain baculovirus expression vector system (BEVS) a very attrac- 2 IQS School of Engineering, Universitat Ramón Llull, tive platform (Fernandes et al. 2013; Gómez-Sebastián et al. Barcelona, Spain 2014). From the different available insect cell lines, High Five 1570 Appl Microbiol Biotechnol (2020) 104:1569–1582 cells have been reported to produce high VLP titers with a conditions encompassing high production yields with an ac- reduced level of contaminant baculovirus (Senger et al. ceptable ratio of correctly assembled Gag in the form of VLPs, 2009). In fact, Cervarix®, the first VLP-based human papil- but at the same time maintaining high productivities. A lomavirus vaccine approved, is produced using the High multiple-criteria decision analysis (MCDA) was implemented Five/BEVS. to combine the RSM-optimized responses into a global opti- Recent advances in the field of nanoparticle quantification mal set of conditions based on two criteria: quantity and qual- and characterization have brought the opportunity to deepen ity. The two optimal production conditions were successfully into the understanding of the BEVS. In the frame of initiatives validated, and the quantity optimum was compared with other such as Process Analytical Technologies (PAT), the applica- Gag VLP production strategies using the High Five/BEVS. tion of new and more sophisticated characterization technolo- Finally, the correct formation and morphology of the VLPs gies will be a powerful tool to increase the understanding of produced was characterized using cryogenic transmission how the process conditions influence the quality of the final electron microscopy (cryo-TEM). The results presented in this product. Nanoparticle tracking analysis (NTA) and flow work represent an advance with respect to current literature virometry are gaining interest as tools to monitor the produc- related to Gag VLP production in the High Five/BEVS and tion of different particles, including VLPs (Nika et al. 2019; provide useful insights into nanoparticle-based process Pereira Aguilar et al. 2019). To date, most of the studies deal- characterization. ing with VLP production are performed with indirect quanti- fication methods based on monomer detection, which hinder the appropriate assessment of the final product and the yields Materials and methods obtained (Haynes et al. 2009; Pillay et al. 2009; Visciano et al. 2011). In addition, the simultaneous production of extracellu- Cell line and culture maintenance lar vesicles (EV), which frequently fall in the size range of VLPs, is not often considered and may lead to an erroneous High Five (cat. num. B85502, Thermo Fisher Scientific, Grand estimation of VLP titers. With the aim to deepen in the char- Island, NY, USA) and Sf9 cells (cat. num. 71104, Merck, acterization of the different existing nanoparticle populations, Darmstadt, Germany) were cultured in suspension in the low- the eGFP was fused in frame to the gag gene from the human hydrolysate content and animal origin-free Sf900III medium immunodeficiency virus serotype 1 (HIV-1). (Thermo Fisher Scientific). Both cell lines were routinely main- Considering the information from previous reports, we de- tained at exponential growth phase in 125-mL disposable poly- cided to apply these novel techniques for optimizing the Gag carbonate Erlenmeyer flasks (Corning, Steuben, NY, USA) in VLP production process using the High Five/BEVS platform. 15 mL of medium and subcultured three times a week at a Different methods are available for process improvement, density of 2–4 × 105 cells/mL for High Five cells (Puente- from the classical one-variable-at-a-time approach to more Massaguer et al. 2018) and 4–6 × 105 cells/mL for Sf9 cells sophisticated statistical design of experiments (DoE) (Puente-Massaguer et al. 2020). All cultures were shaken at (Puente-Massaguer et al. 2019). So far, VLP production stud- 130 rpm using an orbital shaker (Stuart, Stone, UK) and main- ies have focused either on the former approach (Krammer tained at 27 °C in an incubator. Cell count and viability were et al. 2010; Pushko et al. 2017) or on simple factorial designs performed using the Nucleocounter NC-3000 (Chemometec, (Pillay et al. 2009; Pastor et al. 2019), limiting the analysis of Allerod, Denmark) according to manufacturer’s instructions. higher order effects between the variables influencing VLP production. This is of special relevance since the use of lower Baculovirus stock preparation and titration or higher MOI than strictly required might reduce the maxi- mum VLP yield obtained or be detrimental to the system, The recombinant Autographa californica multicapsid respectively. Here, we applied a response surface methodolo- nucleopolyhedrovirus (AcMNPV), containing the HIV-1 gy (RSM) to evaluate the impact of cell concentration at in- Gag matrix protein fused in frame to eGFP (BV-GageGFP) fection (CCI), multiplicity of infection (MOI), and time of (Hermida-Matsumoto and Resh 2000) and under the control harvest (TOH) on VLP production. To go a step further in of the polyhedrin promoter, was constructed using the process understanding, the inclusion of several responses dur- BaculoGold system (BD Biosciences, San Jose, CA, USA). ing process optimization has shown successful results and Baculovirus amplification was performed in the Sf9 cell line applicability in different research areas (Pinzi et al. 2011; (cat. num. 71104, Merck, Darmstadt, Germany) cultured in Honary et al. 2014; Bukzem et al. 2016). Particularly, the three Sf900III medium by infection at 3 × 106 cells/mL, a multiplic- responses considered in addition to VLP production were ity of infection (MOI) of 0.1, and harvest at 96 hours post VLP assembly, baculovirus infection, and VLP productivity. infection (hpi). These are of special relevance in processes with recombinant Titration of baculovirus infectious particles was conducted baculovirus since it is desirable to define production using the plaque assay method. Briefly, 2 mL of Sf9 cells at a Appl Microbiol Biotechnol (2020) 104:1569–1582 1571 concentration of 0.5 × 106 cell/mL was plated per well in 6- v/v of Hoechst (Thermo Fisher Scientific) to visualize the lipid well plates (Nunc, Roskilde, Denmark) and left to attach for membrane and cell nucleus, respectively. A washing step prior 30 min at RT. Then, the Sf900III medium was aspired and Sf9 to observation was performed by centrifugation at 300×g dur- cells were infected with the addition of 1 mL per well of serial ing 5 min, and then cells were resuspended in fresh Sf900III dilutions (10−4 to 10−9) of the BV-GageGFP samples. Six well medium. Samples were placed in 35-mm glass bottom Petri plates were incubated at very mild mixing conditions during dishes with a 14-mm microwell (MatTek Corporation, 1 h to homogeneously distribute the baculoviruses. Ashland, MA, USA) for visualization. The VLP production Afterwards, baculovirus-containing supernatants were aspired process was monitored during 6.5 min at 48 hpi using the and 2 mL of a type VII-A agarose (Merck) solution warmed at Resonant scanner mode of the confocal microscope (Leica). 42 °C was added per well to cover the infected cells. The agarose solution was prepared by mixing an ultrapure water solution containing 3% w/v agarose with Sf900III medium GageGFP quantification supplemented with 10% fetal bovine serum (FBS) at a 1:1 ratio. After agarose cooling, 2 mL of Sf900III medium sup- The amount of GageGFP produced by High Five infected plemented with 10% FBS was added per well and the plates cells was measured intra- and extracellularly by spectrofluo- were incubated for 5 days at 27 °C without agitation. The rometry. Supernatants of GageGFP producing cells were har- concentration of infectious baculovirus particles, expressed vested by centrifugation at 3000×g for 5 min and maintained as plaque forming units (pfu), was measured by visual inspec- at 4 °C until analysis. Intracellular GageGFP content was tion after 3 h staining with a 0.3% w/v Neutral Red solution evaluated by disruption of pelleted cells by means of three (Merck) prepared in DPBS. freeze-thaw cycles (2.5 h frozen at − 20 °C and thawed at 37 °C for 0.5 h). Samples were vortexed for 5 s between Cell growth and baculovirus infection conditions cycles. Lysed pellets were resuspended in 0.5 mL of TMS buffer (50-mM Tris-HCl, 150-mM NaCl, 2-mM MgCl2, Cell growth and infection studies were performed in High pH 8.0), centrifuged at 13,700×g for 20 min and kept at Five cell cultures seeded at 3 × 105 cell/mL in 15 mL and 4 °C. GageGFP fluorescence was measured with a Cary maintained at mid-exponential phase by cell passaging every Eclipse fluorescence spectrophotometer (Agilent 48 h. Several MOI, cell concentrations at infection (CCI), and Technologies, Santa Clara, CA, USA) at RT as follows: times of harvest (TOH) were evaluated and the different pa- λex = 488 nm (5-nm slit), λem = 500–530 nm (10-nm slit). rameter combinations were defined according to a DoE-based Relative fluorescence unit values (RFU) were calculated by approach. Samples were taken every 12–24 h to monitor via- subtracting fluorescence unit values of BV-GageGFP-infected ble cell concentration, cell viability, baculovirus infection, and non-infected cell samples. RFU values were converted intra- and extracellular GageGFP expression, and nanoparticle into GageGFP VLPs as described by Gutiérrez-Granados production. Samples were harvested by centrifugation at and co-authors (Gutiérrez-Granados et al. 2013) and detailed 3000×g for 5 min and supernatants were maintained at 4 °C in the following equation: until analysis. Cell pellets were stored at − 20 °C. GageGFP VLP ðparticles=mLÞ ¼ 1:53 108  RFU: ð1Þ Flow cytometry The Sf900III medium and a 0.1 mg/mL quinine sulfate solution were used as control patterns to normalize RFU The percentage of GageGFP expressing cells was assessed values in samples between different experiments. using a BD FACS Canto II flow cytometer equipped with a 488-nm laser configuration (BD Biosciences). Briefly, 0.3 mL of infected cell cultures was harvested by centrifugation at VLP quantification using nanoparticle tracking 300×g for 5 min and fixed using 4% p-formaldehyde for analysis 10 min. Cells were then centrifuged at 500×g for 5 min to remove p-formaldehyde, resuspended in 0.3-mL ice-cold A NanoSight NS300 (Malvern Panalytical, Malvern, UK) was DPBS (Thermo Fisher Scientific), and maintained at 4 °C used to measure fluorescent and non-fluorescent particles by until further analysis. nanoparticle tracking analysis (NTA). Samples were diluted in 0.22-μm-filtered DPBS and continuously injected into the de- Fluorescence confocal microscopy vice chamber through a pump at an average concentration of 108 particles/mL (20–60 particles/frame). Videos of 60 s from BV-GageGFP–infected High Five cells were observed using a independent triplicate measurements were analyzed with the TCS SP5 confocal microscope (Leica, Wetzlar, Germany). NTA 3.2 software at RT. The nanoparticle number was eval- Cells were stained with 0.05% v/v of CellMask™ and 0.1% uated in the 100–250-nm range. 1572 Appl Microbiol Biotechnol (2020) 104:1569–1582 Nanoparticle quantification using flow virometry productivity responses were transformed to log values using the logarithmic regression to reduce the funnel shape effect of Fluorescent and non-fluorescent nanoparticles were also analyzed the residuals. This transformation compensated the increasing using a CytoFlex LX (Beckman Coulter, Brea, CA, USA) differences in standard deviation associated to higher figures, equipped with a 488-nm blue laser for fluorescent particle detec- a common issue when simultaneously considering values with tion and a 405 nm laser/violet side scatter configuration for im- more than 1000-fold difference (Barbur et al. 1994). In paral- proved nanoparticle resolution. Samples were diluted in 0.22-μm- lel, baculovirus infection and VLP assembly responses were filtered DPBS and triplicate measurements from independent converted using the logistic regression to limit the possible samples were analyzed with the CytExpert 2.3 software at RT. values of the functions to the 0–100% range:   Analytical ultracentrifugation yn Y n ¼ log ð2Þ 100−yn Supernatants of BV-GageGFP–infected High Five cells har- vested at the different conditions of the DoE were where Yn is the converted response and yn is the experimental ultracentrifuged in a double cushion of 5 mL of 25% and response in natural values. 8 mL of 45% (w/v) sucrose (Merck) solution prepared in A Box-Behnken design was used to determine the influ- DPBS and DMEM (Thermo Fisher Scientific), respectively. ence of CCI, MOI, and TOH on each evaluated response. Five milliliters of supernatants from triplicate experiments These variables were screened at a low (− 1), medium (0), were loaded in ultracentrifuge tubes (Beckman Coulter) and and high level (+ 1) as indicated in Table 1. The different filled to the top with sterile DPBS. Sample ultracentrifugation levels were linearly related for CCI and TOH and exponen- was performed in a Beckman Optima L100XP centrifuge tially for MOI, always in an equidistant manner. The results (Beckman Coulter) equipped with a SW-32Ti rotor at obtained for each response were fitted to a second-order poly- 31,000 rpm for 2.5 h at 4 °C. Samples were taken from each nomial by the least squares method (Eq. 3): ultracentrifuge fraction, and pellets were resuspended in 100 μL of DPBS and kept overnight at 4 °C. All samples were Y ¼ β0 þ β1  CCI þ β2  MOI þ β3  TOH þ β4  CCI stored at − 20 °C until their measurement by spectrofluorom- etry and at 4 °C for nanoparticle analysis by flow virometry.  MOI þ β5  CCI  TOH þ β6  MOI  TOH þ β7 The same procedure was applied to characterize the optimal  CCI2 þ β8  MOI2 þ β9  TOH2 þ ε ð3Þ conditions obtained with the DoE and the MCDA approach. where Y is the response variable, β0 is the model intercept Cryogenic transmission electron microscopy term, β1 − β9 are the model coefficients, and ε is the experi- mental error. Fitting of the different equations based on Eq. 3 VLP morphology was assessed using a transmission electron was performed with the R software (R Foundation for microscope equipped for sample cryogenics (cryo-TEM). Statistical Computing, Vienna, Austria). Three-dimensional Briefly, 4 μL of sample was blotted onto EMR Holey carbon plots were constructed to facilitate model interpretation. films on 400-mesh copper grids (Micro to Nano, Wateringweg, The most common DoE approach involves the optimiza- Netherlands) previously subjected to a glow discharge treat- tion of a single response. However, such analysis may not be ment in a PELCO easiGlow™ Discharge Cleaning System enough for the BEVS, since the definition of a production (PELCO, Fresno, CA, USA). Deposited samples were then condition cannot be accurately predicted from an individual plunged into liquid ethane at − 180 °C using a Leica EM GP property. Then, the objective of including more than one re- workstation (Leica) and observed in a JEM-2011 TEM operat- sponse in the study was to determine global conditions inte- ing at 200 kV (Jeol Ltd., Akishima, Tokyo, Japan). grating different key aspects of the BEVS. To this purpose, a multiple-criteria decision analysis (MCDA) based on desir- Multiple optimization of VLP production conditions ability functions was implemented. The different responses were transformed to a dimensionless desirability scale (dn) A response surface methodology (RSM) was applied to assess ranging between 0 and 1: the effect of the main factors influencing the BEVS: CCI (cell/mL), MOI (pfu/cell), and TOH (h) on different re- sponses. The selected DoE responses were baculovirus infec- d n ¼ 0 if yn < minn   tion (% of infected cells), VLP production (fluorescent parti- yn −minn w cle/mL), GageGFP polyprotein assembled in the form of dn ¼ if minn ≤ yn ≤ maxn maxn −minn VLPs (% of GageGFP assembled as VLPs), and VLP produc- tivity (fluorescent particle/mL h). VLP production and VLP d n ¼ 1 if yn > maxn ð4Þ Appl Microbiol Biotechnol (2020) 104:1569–1582 1573 Table 1 Matrix design, independent variable levels, and regression coefficients for the Box-Behnken design. VLP concentration values were quan- tified using nanoparticle tracking analysis Box-Behnken design Variables Responses VLP production VLP productivity Experimental run CCI (106 MOI Infection VLP TOH (h) (108 fluorescent (106 fluorescent cell/mL) (pfu/cell) (%) assembly (%) particle/mL) particle/mL·h) 1 1.5 0.01 84 98.5 345.5 14.4 411.3 2 0.5 0.22 84 99.5 118.1 31.5 140.6 3 1.5 0.22 54 99.2 230.0 17.1 420.0 4 1.5 0.22 54 99.6 214.8 15.2 410.0 5 2.5 5 54 99.3 320.2 13.4 593.0 6 1.5 5 84 99.4 265.3 17.5 315.8 7 1.5 0.01 24 3.3 0.2 98.2 0.9 8 1.5 5 24 95.5 11.6 57.2 48.3 9 2.5 0.01 54 21.7 8.7 14.6 16.1 10 0.5 0.01 54 97.8 34.3 37.7 63.4 11 0.5 0.22 24 56.4 1.9 94.6 7.7 12 2.5 0.22 84 99.5 465.4 13.2 554.1 13 0.5 5 54 99.7 101.8 34.3 188.5 14 1.5 0.22 54 99.4 200.0 13.1 370.0 15 2.5 0.22 24 30.7 1.1 93.9 4.4 Analysis LOF test, p- Function F test, p-value R2 Adjusted R2 Predicted R2 value Infection <0.01 0.1 95.0 90.0 68.7 VLP production <0.01 0.05 99.7 99.1 95.0 VLP assembly <0.01 0.06 94.9 92.1 79.9 VLP productivity <0.01 0.05 99.5 98.7 92.9 Model Infection VLP production VLP assembly VLP productivity Parameters Coefficient p-value Coefficient p-value Coefficient p-value Coefficient p-value Constant 4.7 <0.01 5.4 <0.01 -1.4 <0.01 6.0 <0.01 CCI -0.9 0.03 0.1 >0.1 -0.5 0.04 0.1 >0.1 CCI2 NS >0.1 -0.8 <0.01 NS >0.1 -0.8 <0.01 MOI 2.0 <0.01 1.1 <0.01 -0.5 0.04 1.1 <0.01 MOI2 -1.0 0.08 -0.6 <0.01 NS >0.1 -0.6 <0.01 TOH 2.6 <0.01 2.3 <0.01 -2.0 <0.01 2.0 <0.01 2 TOH -1.8 <0.01 -1.7 <0.01 -1.9 <0.01 -1.5 <0.01 CCI · MOI 1.1 0.06 0.6 <0.01 NS >0.1 0.6 <0.01 CCI · TOH NS >0.1 0.6 <0.01 NS >0.1 0.5 <0.01 MOI · TOH -1.4 0.02 -1.1 <0.01 1.0 <0.01 -1.1 <0.01 NS non-statistically significant where yn is the predicted response of the fitted equation, dn is value, and minn and maxn are the minimum and maximum the desirability response function, w accounts for the weight acceptable values of yn, respectively. The weight (w value) 1574 Appl Microbiol Biotechnol (2020) 104:1569–1582 depended on the relative importance assigned to each individ- ual response. Then, all the different desirability response func- tions under study were combined into a single equation, de- noted as overall desirability (OD): sffiffiffiffiffiffiffiffiffiffi N N OD ¼ ∏ dn ð5Þ 1 where OD is the overall desirability function to be maximized and N is the number of responses studied in the optimization process. Statistical analyses Fig. 1 Cell growth and viability profiles of High Five cells in Sf900III medium in batch culture. Exponentially growing cells were seeded at The experimental design and statistical analyses of the differ- 0.3 × 106 cells/mL in 125-mL flasks. Mean values ± standard deviation ent equations were performed using R software. The quality of of triplicate experiments are represented the regression of the fitted equations was evaluated with the R2 and adjusted R2 coefficients in percentage units. An analysis of variance (ANOVA) F test was used to determine the signif- simplify their industrial application, but these conditions were icance of the equations and the individual coefficients were evaluated as a contrast to the optimal conditions obtained as assessed with a t test. The lack-of-fit (LOF) test was used to further presented. evaluate differences between experimental and pure error of the fitted equations. In all analyses, p values of 0.05 and 0.1 were considered statistically significant with 95% and 90% Combining RSM and MCDA to define the optimal VLP confidence, respectively. production conditions A three-variable (CCI, MOI, and TOH) Box-Behnken RSM Results was constructed and the following objective responses were analyzed: baculovirus infection efficiency, VLP production, Definition of the experimental space boundaries VLP assembly, and VLP productivity. This experimental de- sign was selected among different candidates due to its vari- Initially, the growth of High Five cells was characterized to ance distribution and spatial properties (Cervera et al. 2015a; determine an adequate range of viable cell concentration at Gutiérrez-Granados et al. 2016; Puente-Massaguer et al. infection (CCI) for the DoE. Cells maintained a high viability 2018). The resulting matrix consisted in 15 experiments in- (> 90%) until 96 h of culture, coinciding with the peak of cluding a triplicated center point to account for the inherent viable cell concentration. In these conditions, cells reached a variability of the system (Table 1). The different datasets were maximum of 7.3 ± 0.2 × 106 cells/mL and a doubling time (Td) adjusted to the second-order polynomial functions by the least of 18.7 ± 0.5 h (Fig. 1). Considering the extent of the expo- squares method according to Eq. 3. The statistical significance nential phase and the Td, the range for cell concentration at of the different functions and their associated coefficients was infection was restricted to 0.5–2.5 × 106 cells/mL. Some dif- confirmed by ANOVA (Table 1). Each function was used to ficulties were encountered when defining the different levels construct three-dimensional plots for a better comprehension of multiplicity of infection (MOI) and time of harvest (TOH) of the synergies between CCI, MOI, and TOH (Fig. 2). As since very little information on the use of the BEVS with High expected, baculovirus infection progressed more rapidly with Five cells in the Sf900III medium was available. The ranges increasing MOI (Fig. 2a–c). However, the pace of infection for both variables were eventually fixed according to previous decreased when using the same MOI at higher CCI. A similar experience with this system (data not shown) and based on behavior was observed between VLP production and VLP studies of the High Five/BEVS in other cell culture media productivity (Fig. 2d–i). Low MOI required longer TOH to (Sander and Harrysson 2007; Pillay et al. 2009). The MOI achieve higher VLP yields, whereas higher MOI shortened experimental range was fixed between 0.01 and 5 with the this period. In both cases, the best production and productivity aim to screen low and high MOI conditions in the same titers were obtained in the 1.5–2.5 × 106 cell/mL CCI range. DoE. In the case of TOH, the boundaries were defined as TOH also proved to be the most significant variable regarding 24–84 hpi. DoE space boundaries did not involve high CCI the amount of GageGFP monomer assembled in the form of out of the exponential phase or medium exchange in order to VLPs (Fig. 2j–l). Increasing TOH and MOI substantially Appl Microbiol Biotechnol (2020) 104:1569–1582 1575 Fig. 2 Three-dimensional plots of the different functions based on the MOI, and TOH. All the graphs were constructed by representing the Box-Behnken experimental results. a–c Baculovirus infection. d–f VLP effect of two independent variables on a response while maintaining the production. g–i VLP productivity. j–l VLP assembly as a function of CCI, third one at a fixed level decreased the ratio of VLPs to non-assembled monomers, this case, different independent variable combinations resulted regardless the value of CCI. in the optimum condition for each function (Table 2), thus The models were used to predict an optimal combination of making difficult to define a global optimum condition com- CCI, MOI and TOH to maximize each specific response. In bining all the different models at once. Therefore, a multiple- 1576 Appl Microbiol Biotechnol (2020) 104:1569–1582 Table 2 Optimal CCI, MOI, and TOH conditions maximizing each GageGFP levels reached a maximum concentration that sub- response function sequently decreased (Fig. 4b). Analysis of the VLP production Response function CCI (106 cell/mL) MOI (pfu/cell) TOH (h) process of the quantity optimum by flow virometry evidenced that there was a continuous increase in VLP concentration Infection efficiency 0.5 0.2 76.4 (Fig. 4c). Comparison of VLP production in both optima by VLP production 2.0 1.3 73.5 means of nanoparticle tracking analysis (NTA) yielded a 10- VLP assembly 0.5 0.01 24.0 fold and 4.8-fold increase in production and productivity in VLP productivity 2.0 2.3 67.8 the quantity over the quality optimum, respectively. In all cases, the experimental outcome was within the predicted limits (Table 3). Similar ratios of VLPs to total nanoparticles criteria decision analysis (MCDA) based on desirability func- were observed, representing a 9.7 ± 3.1% for the quality and a tions was applied in order to determine a global optimum 9.3 ± 1.1% for the quantity condition. In contrast, analysis of combining the different functions. Every model function was the optima by sucrose cushion ultracentrifugation showed a transformed to a dimensionless scale (Eq. 4) and given a 4.9-fold increase of GageGFP monomer assembled in the weight according to the importance of each response. A higher form of VLPs in the quality optimum (64.6 ± 5.1%). The SN priority was assigned to VLP production, VLP assembly, and and SN – 25% fractions, corresponding to non-assembled VLP productivity (w = 2), while a lower restriction was GageGFP, showed significantly larger levels of fluorescence assigned to infection (w = 1). Two possible response combi- in the quantity optimum (Fig. 4d). However, VLP-containing nations were considered in the MCDA, the first focused on fractions (25–45% and 45%) were more enriched in the qual- maximizing product quantity and the second one maximizing ity condition. BV-GageGFP titration of both optima by plaque product quality. To do this, the VLP assembly function was assay resulted in a 2.7-fold increase of infectious particles in excluded from the MCDA in the definition of a global optimal the quantity (13.2 ± 5.1 × 107 pfu/mL) compared with quality condition targeting product quantity. The application of the condition (4.8 ± 1.5 × 107 pfu/mL). MCDA for the product quality optimum resulted in a CCI, An acceptable correlation between indirect VLP quantifi- MOI, and TOH of 1.1 × 10 6 cell/mL, 1 pfu/cell, and cation with spectrofluorometry and direct nanoparticle mea- 33.6 hpi, respectively. The optimum maximizing product surement with NTA was only achieved for the quality opti- quantity was determined as 2.0 × 106 cell/mL for CCI, a mum condition (Eq. 1). On the contrary, GageGFP fluores- MOI of 2.5 pfu/cell, and a TOH of 69.3 hpi. cence levels of the quantity optimum measured by spectroflu- orometry (2448.6 ± 23.8 RFU) and their conversion to VLP Validating the quantity and quality global optima concentration according to Eq. 1 resulted in a 9.6-fold higher value than the one obtained by direct VLP quantification The two global optimal conditions obtained by the combina- using NTA. However, multiplying this fluorescence value by tion of RSM and MCDA were successfully corroborated in a the VLP assembly factor calculated by analytical ultracentri- validation experiment. High Five cell growth was arrested fugation (Table 3) allowed to determine the GageGFP fluores- after BV-GageGFP infection with > 80% of the culture infect- cence exclusively attributed to VLPs (323.2 ± 3.1 RFU) and ed at 24 hpi (Fig. 3a–b). As expected, a higher cell viability discard that of unassembled GageGFP monomer. Introducing was observed for the quality optima (33.6 hpi) compared with the fluorescence value associated to VLPs in Eq. 1 resulted in the quantity condition at TOH (69.3 hpi). Confocal microsco- a better prediction to NTA quantification (4.9 ± 0.5 × 1010 py analysis of infected cells showed successful colocalization VLP/mL). Supernatant characterization of the quality and (yellow) of the GageGFP polyprotein (green) with the cell quantity optima by cryogenic transmission electron microsco- membrane (red), thus indicating that VLP formation and pro- py (cryo-TEM) indicated that VLPs were correctly formed in duction was taking place (Fig. 3c). This process could be both conditions (Fig. 5). Interestingly, the median VLP size tracked in real time using a confocal fluorescence microscope was higher in the quality (181.4 ± 4.4 nm) compared with the with the high-speed acquisition mode (Fig. S1). quantity optimum (164.7 ± 0.3 nm). The presence of EVs and Measurement of fluorescence indicated that the highest baculoviruses was also detected, in agreement with the exis- pace of GageGFP production was achieved with higher tence of a non-fluorescent particle population observed with MOI. Also, the highest fluorescence yield in the supernatant NTA. was achieved with the quantity optimum (2448.6 ± 23.8 RFU) at 69.3 hpi (Fig. 4a). The TOH for the quality optimum con- Comparison with other VLP production conditions dition coincided with the peak of intracellular GageGFP pro- duction, but not for the quantity optimum. In all cases, In addition to the optimization work performed, the quantity GageGFP production in the supernatant progressively in- optimum condition was compared with two conditions used in creased with longer TOH, although the intracellular other works for the production of high VLP titers in order to Appl Microbiol Biotechnol (2020) 104:1569–1582 1577 Fig. 3 Validation experiment of the quality and quantity optima. a Viable tested. Infected cells were measured every 24 h until their optimal TOH. c cell concentration and viability progression after infection with the BV- High Five cells infected with the BV-GageGFP at 2.0 × 106 cell/mL and a GageGFP. The center point of the Box-Behnken DoE (infection control) MOI of 2.5 and observed at 48 hpi. Cell nucleus was stained with Hoechst was also included to account for the robustness of the infection process, (blue) and cell membrane with CellMask™ (red). The mean and standard with cells infected with a MOI of 0.22 at 1.5 × 106 cell/mL and harvested deviation of triplicate experiments are represented at 54 hpi. b BV-GageGFP infection process in the different conditions corroborate the results of the study. The first approach corresponding to 3168.6 ± 100.2 RFU in comparison with consisted in infecting High Five cells at a high cell concentra- the 2528.0 ± 76.3 RFU achieved in the quantity condition tion and MOI. To this purpose, High Five cells were infected (Fig. 6c–d). The same improvement was observed in VLP at 6 × 106 cell/mL with a MOI of 5. The second strategy quantification by NTA, with 4.3 ± 0.2 × 1010 VLP/mL for me- consisted in performing a medium replacement before infec- dium exchange while 3.4 ± 0.3 × 1010 VLP/mL was measured tion using the quantity optimum. This option was considered for the quantity optimum condition. However, baculovirus as a means to value if the added effort of performing a medium infection at the high CCI condition reached a significantly replacement would provide a substantial advantage in VLP lower GageGFP production (66.6 ± 3.3 RFU) and also in concentration. Both conditions were not examined in the terms of VLP production (2.3 ± 0.6 × 108 VLP/mL). DoE in order to avoid the added difficulties required at process The median size of the VLPs produced in these conditions level, but were compared with the quantity optimum as a did not substantially vary from the quantity optimum (168.2 ± contrast since they have been used in the literature. 3.3 nm), being 169.6 ± 3.3 nm for the medium exchange and Baculovirus infection progressed more rapidly after medium 160.5 ± 3.7 nm for the high CCI. Similar levels of total nano- exchange than in the original condition but the drop in cell particles were obtained in the three different conditions, thus viability was not significantly affected (Fig. 6a–b). On the representing a big difference in the ratio of VLP to total nano- other hand, baculovirus infection advanced more slowly in particles (Table 4). Remarkable differences were also observed the high CCI condition, to the point that cell culture was not as regards the amount of infectious baculovirus particles pro- fully infected after 72 hpi. Medium exchange before infection duced. Medium replacement before infection increased the increased GageGFP production by 1.3-fold at 69.3 hpi, levels by 2.1-fold (47.0 ± 9.9 × 107 pfu/mL) compared with 1578 Appl Microbiol Biotechnol (2020) 104:1569–1582 Fig. 4 GageGFP and VLP production in the validation experiment. a–b virometry in the quantity optimum condition. d Fluorescence distribution GageGFP production process in the supernatant and intracellularly. after double sucrose cushion ultracentrifugation of the quality and Samples were taken every 24 h until their respective TOH. c quantity optima. These experiments were performed in triplicate Fluorescent and total nanoparticle production profile measured by flow the 22.7 ± 2.5 × 107 pfu/mL achieved in the quantity optimum of complex multimeric structures, such as enveloped VLPs, condition. However, a much lower baculovirus titer of 0.9 ± have not been explored in detail. To date, the assessment of 0.3 × 107 pfu/mL was obtained in the high CCI condition. VLP production has been performed based on indirect quan- tification methods (Pillay et al. 2009). Moreover, most of the research has been conducted either in low performing or ani- Discussion mal compound containing cell culture media (Krammer et al. 2010; Wilde et al. 2014). Although these works have brought The High Five/BEVS has demonstrated a high versatility to advances in the production of nanoparticles using the High produce a wide variety of recombinant proteins. However, the Five/BEVS, a deeper study would contribute to exploit the capabilities of the High Five cell line as a production platform full potential of such platform. Table 3 Experimental validation of the global optimal conditions and comparison with model predictions. VLP concentration measurements were carried out with NTA. The prediction intervals represent the lower and the upper limits according to the logarithmic or logistic response transformation Response Product quality Product quantity Experimental Model prediction Experimental Model prediction Infection (%) 97.8 ± 2.0 [71.6–99.7] 100 [96.6–100] VLP production (108 fluorescent particle/mL) 39.0 ± 4.5 [15.6–60.2] 391.0 ± 13.9 [337.2–1333.4] VLP assembly (%) 64.6 ± 5.1 [46.4–72.9] 13.2 ± 1.0 [5.7–18.2] VLP productivity (106 fluorescent particle/mL h) 118.1 ± 27.8 [48.3–186.2] 564.2 ± 20.2 [477.6–1884.8] Appl Microbiol Biotechnol (2020) 104:1569–1582 1579 Fig. 5 GageGFP VLP characterization by cryo-TEM. a–b Supernatant visualization of the quality (33.6 hpi) and quantity (69.3 hpi) optima, respectively. The arrows indicate the different specimens: VLP (green), EV (blue), and baculovirus (red) CCI, MOI, and TOH were selected as the critical parame- considered and simultaneously evaluated by means of RSM ters affecting the production of HIV-1 Gag VLPs in the BEVS. and MCDA. TOH and MOI were shown to be particularly For a comprehensive analysis, four different responses were relevant to achieve a high VLP concentration but also in Fig. 6 Comparison of VLP production using the quantity optimum with represented. Medium exchange was performed prior to BV-GageGFP medium exchange and baculovirus infection at a high CCI. a–b Viable infection and consisted in infecting High Five cells at 2.0 × 106 cell/mL cell concentration, cell viability, and BV-GageGFP infection process in with a MOI of 2.5 pfu/cell. As regards the high CCI condition, High Five the different conditions tested. c–d Supernatant and intracellular cells were infected at 6 × 106 cell/mL with a MOI of 5 GageGFP production. The average of triplicate experiments is 1580 Appl Microbiol Biotechnol (2020) 104:1569–1582 Table 4 Comparison of GageGFP VLP production levels of the quantity optimum, medium exchange, and high CCI conditions measured by NTA Condition VLP production Total particle production Ratio fluorescent particles/ VLP size (nm) (108 fluorescent particles/mL) (108 particles/mL) total particles (%) Quantity optimum 335.0 ± 29.0 4260.0 ± 609.0 8 ± 1.7 168.2 ± 3.3 Medium exchange 432.0 ± 22.0 5510.0 ± 232.0 7.8 ± 0.2 169.6 ± 3.3 High CCI 2.3 ± 0.6 5230.0 ± 767.0 0.1 160.5 ± 3.7 maintaining an adequate ratio of GageGFP assembly in the The application of cryo-TEM helped to gain insight into form of VLPs. Pillay and co-workers used a factorial experi- the identification of the morphological features of the VLPs ment to assess the effect of different MOI comprised in the produced (Fig. 5). Cryo-TEM has recently been highlighted as range of 0.1–5 in High Five cells (Pillay et al. 2009). They did powerful methodology for nanoparticle characterization as the not identify any significant difference in Gag yield, but here different specimens can be observed in native conditions we observed that MOI plays a significant role in VLP produc- (Martin et al. 2016; Maldonado et al. 2016; Gallagher et al. tion. Interestingly, including a function for VLP assembly 2018). GageGFP VLPs presented the expected size of 100– based on the measurement of the GageGFP fluorescence dis- 200 nm and morphology consistent with immature HIV-1 tribution in the different fractions of an ultracentrifugation VLPs, comparable with VLPs produced in Sf9 (Nika et al. proved to be highly relevant. These results in combination 2017) and HEK 293 cells (Venereo-Sanchez et al. 2016). with direct nanoparticle quantification by NTA and flow The rod-shaped structure corresponding to the infective form virometry permitted to obtain a more accurate perspective of of the baculovirus was also detected. The presence of EVs was the process. observed alongside VLPs, which are known to be produced as A pronounced decrease in VLP assembly was shown with way of communication in eukaryotic cells (Ratajczak et al. increasing TOH, limiting a favorable combination with the other 2006). EVs were more heterogeneous in size (50–500 nm) three responses. To this purpose, two different global optimiza- and shape compared with VLPs and less electron-dense. tion strategies were examined by MCDA, one considering the These results coincide with the measurements performed by VLP assembly function (therefore with priority on product qual- NTA and flow virometry where a remarkable population of ity) and the other not (therefore with priority on product quantity). non-fluorescent nanoparticles was detected. Future efforts are Remarkable differences were encountered between the two opti- needed to understand their role in insect cell lines since little ma, achieving a 10-fold higher VLP production in the quantity information is available. condition, whereas a 4.9-fold increase in the assembly of A maximum HIV-1 GageGFP VLP concentration of 3.9 ± GageGFP as VLPs was obtained in the quality optimum. Then, 0.1 × 1010 VLP/mL was achieved in the supernatant using the it is evident that High Five cells possess a high capacity to pro- product quantity optimum. Thus, these conditions would be se- duce GageGFP, but there is a remarkable competition between lected in case of designing a VLP manufacturing process. the formation of VLPs and the release of unassembled GageGFP However, if the presence of baculoviruses posed a risk for the monomer. The large amount of GageGFP produced, potentially final VLP application, the quality optimum provides a 2.7-fold overcoming the capacity of the cell line itself to process all the reduction in infectious particles, which would facilitate the down- GageGFP in the form of VLPs, combined to cell membrane stream process. Difficulties were encountered in the comparison disruption could be the reason explaining these results. This could of the VLP titer obtained since most of the studies producing Gag be appreciated even in the quality optimum itself, in which a high VLPs with the insect cell/BEVS used indirect quantification cell viability of 92.4 ± 0.6% was measured at harvest (33.6 hpi) methods based on monomer detection (Buonaguro et al. 2001; (Fig. 3a) and a 64.6 ± 5.1% of the fluorescence in the supernatant Pillay et al. 2009; Visciano et al. 2011). Nevertheless, when the was associated to VLPs, meaning that around the 35% remaining comparison of nanoparticle production was possible, this opti- was due to unassembled GageGFP monomer. These results em- mum condition proved to be superior to Gag VLP production in phasize the need to complement the methods most currently used HEK 293 cells without (Cervera et al. 2013) and with additive for VLP identification or quantification based on monomer de- supplementation (Cervera et al. 2015a) by 14.4 and 1.7-fold, tection with direct nanoparticle assessment techniques, specially respectively. Moreover, these conditions proved to be superior using the High Five cell line (Maranga et al. 2003; Pillay et al. to influenza VLP production using the Sf9 (Thompson et al. 2009; Tagliamonte et al. 2010; Shoja et al. 2013; Monteiro et al. 2015) and High Five/BEVS (Carvalho et al. 2016) by 6.7 and 2016; Fuenmayor et al. 2019). 25-fold, respectively. The potential of the quantity optimum was Appl Microbiol Biotechnol (2020) 104:1569–1582 1581 also evaluated against other baculovirus infection strategies in References this work. Infection at a high CCI using a high MOI resulted in a much lower yield of VLP production probably owing to the Andersson AMC, Schwerdtfeger M, Holst PJ (2018) Virus-like-vaccines against HIV. Vaccines 6:1–17 scarcity of some nutrients in the medium (Table 4). BV produc- Barbur VA, Montgomery DC, Peck EA (1994) Introduction to linear tion was also significantly reduced in such conditions. These regression analysis. Wiley differences in production could be related to the cell density Bernal V, Carinhas N, Yokomizo AY, Carrondo MJT, Alves PM (2009) effect previously described in insect cells with the BEVS Cell density effect in the baculovirus-insect cells system: a quanti- (Bernal et al. 2009; Huynh et al. 2013; Huynh et al. 2014). On tative analysis of energetic metabolism. Biotechnol Bioeng 104: 162–180 the other side, medium replacement before BV infection moder- Bollin F, Dechavanne V, Chevalet L (2011) Design of experiment in CHO ately increased VLP production but duplicated baculovirus pro- and HEK transient transfection condition optimization. Protein Expr duction. This strategy has proven to be an efficient manner to Purif 78:61–68 increase both the yield of recombinant proteins (Caron et al. Bukzem AL, Signini R, dos Santos DM, Lião LM, Ascheri DPR (2016) 1990) and baculoviruses (Carinhas et al. 2009) with the Sf9/ Optimization of carboxymethyl chitosan synthesis using response surface methodology and desirability function. Int J Biol Macromol BEVS cells and also in mammalian cells (Bollin et al. 2011; 85:615–624 Cervera et al. 2015b). In this case, no substantial benefit was Buonaguro L, Buonaguro FM, Tornesello ML, Mantas D, Beth-Giraldo obtained, thus indicating that the added efforts required to per- E, Wagner R, Michelson S, Prevost MC, Wolf H, Giraldo G (2001) form medium replacement might not be advantageous. High efficient production of Pr55gag virus-like particles expressing multiple HIV-1 epitopes, including a gp120 protein derived from an In conclusion, a detailed study of Gag VLP production Ugandan HIV-1 isolate of subtype A. Antivir Res 49:35–47 using the High Five/BEVS was conducted. The application Buonaguro L, Tagliamonte M, Visciano ML (2013) Chemokine receptor of RSM and MCDA allowed to go a step further and capture interactions with virus-like particles. In: Methods in molecular biol- the synergies between the influential factors affecting the VLP ogy. Humana Press, Totowa, pp 57–66 production process. The combination of advanced statistical Carinhas N, Bernal V, Yokomizo AY, Carrondo MJT, Oliveira R, Alves PM (2009) Baculovirus production for gene therapy: the role of cell methods with a variety of nanoparticle quantification and density, multiplicity of infection and medium exchange. Appl characterization techniques allowed to obtain a more compre- Microbiol Biotechnol 81:1041–1049 hensive picture of the process and to determine a high- Caron AW, Archambault J, Massie B (1990) High-level recombinant yielding and reproducible condition for VLP production. protein production in bioreactors using the baculovirus-insect cell The approach here proposed is valuable for studies dealing expression system. Biotechnol Bioeng 36:1133–1140 Carvalho SB, Freire JM, Moleirinho MG, Monteiro F, Gaspar D, with the production of other complex nanoparticles. Future Castanho MARB, Carrondo MJT, Alves PM, Bernardes GJL, efforts in transferring this process into bioreactor scale should Peixoto C (2016) Bioorthogonal strategy for bioprocessing of allow to define a robust platform for VLP production. specific-site-functionalized enveloped influenza-virus-like particles. Bioconjug Chem 27:2386–2399 Acknowledgments The authors would like to thank Dr. Paula Alves Cervera L, Gutiérrez-Granados S, Martínez M, Blanco J, Gòdia F, Segura (Instituto de Biologia Experimental e Tecnológica, Oeiras, Portugal) MM (2013) Generation of HIV-1 Gag VLPs by transient transfec- and Dr. Nick Berrow (Institute for Research in Biomedicine, Barcelona, tion of HEK 293 suspension cell cultures using an optimized Spain) for providing the High Five and Sf9 cell lines, respectively. Marta animal-derived component free medium. J Biotechnol 166:152–165 Martínez-Calle developed the BV-GageGFP. Ángel Calvache and Jorge Cervera L, Fuenmayor J, González-Domínguez I, Gutiérrez-Granados S, Fomaro (Beckman Coulter) facilitated the access to the CytoFlex LX flow Segura MM, Gòdia F (2015a) Selection and optimization of trans- cytometer. The support of Martí de Cabo and Mónica Roldán from Servei fection enhancer additives for increased virus-like particle produc- de Microscòpia of Universitat Autònoma de Barcelona (UAB) with the tion in HEK293 suspension cell cultures. Appl Microbiol cryo-TEM and confocal microscopy, respectively, and Paolo Saccardo Biotechnol 99:9935–9949 from Plataforma de Producción de Proteínas, CIBER de Bioingeniería, Cervera L, Gutiérrez-Granados S, Berrow NS, Segura MM, Gòdia F Biomateriales y Nanomedicina (CIBER-BBN) of UAB with baculovirus (2015b) Extended gene expression by medium exchange and repeat- titration are appreciated. The research group is recognized as 2017 SGR ed transient transfection for recombinant protein production en- 898 by Generalitat de Catalunya. hancement. Biotechnol Bioeng 112:934–946 Deo VK, Kato T, Park EY (2015) Chimeric virus-like particles made using GAG and M1 capsid proteins providing dual drug delivery Funding information Eduard Puente-Massaguer is a recipient of an FPU and vaccination platform. Mol Pharm 12:839–845 grant from Ministerio de Educación, Cultura y Deporte of Spain (FPU15/ Fernandes F, Teixeira AP, Carinhas N, Carrondo MJT, Alves PM (2013) 03577). Insect cells as a production platform of complex virus-like particles. Expert Rev Vaccines 12:225–236 Fuenmayor J, Cervera L, Gòdia F, Kamen A (2019) Extended gene ex- Compliance with ethical standards pression for Gag VLP production achieved at bioreactor scale. J Chem Technol Biotechnol 94:302–308 Conflict of interest The authors declare that they have no competing Gallagher J, McCraw D, Torian U, Gulati N, Myers M, Conlon M, Harris interests. A (2018) Characterization of hemagglutinin antigens on influenza virus and within vaccines using electron microscopy. Vaccines 6:1– Ethical approval This article does not contain any studies with human 21 participants performed by any of the authors. 1582 Appl Microbiol Biotechnol (2020) 104:1569–1582 Gómez-Sebastián S, López-Vidal J, Escribano JM (2014) Significant Pillay S, Meyers A, Williamson AL, Rybicki EP (2009) Optimization of productivity improvement of the baculovirus expression vector sys- chimeric HIV-1 virus-like particle production in a baculovirus-insect tem by engineering a novel expression cassette. PLoS One 9:1–10 cell expression system. Biotechnol Prog 25:1153–1160 Gutiérrez-Granados S, Cervera L, Gòdia F, Carrillo J, Segura MM (2013) Pinzi S, Leiva D, Arzamendi G, Gandia LM, Dorado MP (2011) Multiple Development and validation of a quantitation assay for fluorescently response optimization of vegetable oils fatty acid composition to tagged HIV-1 virus-like particles. J Virol Methods 193:85–95 improve biodiesel physical properties. Bioresour Technol 102: Gutiérrez-Granados S, Cervera L, Segura M d l M, Wölfel J, Gòdia F 7280–7288 (2016) Optimized production of HIV-1 virus-like particles by tran- Puente-Massaguer E, Lecina M, Gòdia F (2018) Nanoscale characteriza- sient transfection in CAP-T cells. Appl Microbiol Biotechnol 100: tion coupled to multi-parametric optimization of Hi5 cell transient 3935–3947 gene expression. Appl Microbiol Biotechnol 102:10495–10510 Haynes JR, Dokken L, Wiley JA, Cawthon AG, Bigger J, Harmsen AG, Puente-Massaguer E, Badiella L, Gutiérrez-Granados S, Cervera L, Richardson C (2009) Influenza-pseudotyped Gag virus-like particle Gòdia F (2019) A statistical approach to improve compound screen- vaccines provide broad protection against highly pathogenic avian ing in cell culture media. Eng Life Sci 19:1–13 influenza challenge. Vaccine 27:530–541 Puente-Massaguer E, Lecina M, Gòdia F (2020) Application of advanced Hermida-Matsumoto L, Resh MD (2000) Localization of human immu- quantification techniques in nanoparticle-based vaccine develop- nodeficiency virus type 1 Gag and Env at the plasma membrane by ment with the Sf9 cell baculovirus expression system. Vaccine. confocal imaging. J Virol 74:8670–8679 https://doi.org/10.1016/j.vaccine.2019.11.087 Honary S, Ebrahimi P, Hadianamrei R (2014) Optimization of particle Pushko P, Tretyakova I, Hidajat R, Zsak A, Chrzastek K, Tumpey TM, size and encapsulation efficiency of vancomycin nanoparticles by Kapczynski DR (2017) Virus-like particles displaying H5, H7, H9 hem- response surface methodology. Pharm Dev Technol 19:987–998 agglutinins and N1 neuraminidase elicit protective immunity to heterol- ogous avian influenza viruses in chickens. Virology 501:176–182 Huynh HT, Tran TTB, Chan LCL, Nielsen LK, Reid S (2013) Decline in Ratajczak J, Wysoczynski M, Hayek F, Janowska-Wieczorek A, baculovirus-expressed recombinant protein production with increasing Ratajczak MZ (2006) Membrane-derived microvesicles: important cell density is strongly correlated to impairment of virus replication and and underappreciated mediators of cell-to-cell communication. mRNA expression. Appl Microbiol Biotechnol 97:5245–5257 Leukemia 20:1487–1495 Huynh HT, Tran TTB, Chan LCL, Nielsen LK, Reid S (2014) Effect of Sander L, Harrysson A (2007) Using cell size kinetics to determine optimal the peak cell density of recombinant AcMNPV-infected Hi5 cells on harvest time for Spodoptera frugiperda and Trichoplusia ni BTI-TN- baculovirus yields. Appl Microbiol Biotechnol 99:1687–1700 5B1-4 cells infected with a baculovirus expression vector system ex- Krammer F, Schinko T, Palmberger D, Tauer C, Messner P, Grabherr R pressing enhanced green fluorescent protein. Cytotechnology 54:35–48 (2010) Trichoplusia ni cells (High Five™) are highly efficient for Senger T, Schädlich L, Gissmann L, Müller M (2009) Enhanced the production of influenza A virus-like particles: a comparison of papillomavirus-like particle production in insect cells. Virology two insect cell lines as production platforms for influenza vaccines. 388:344–353 Mol Biotechnol 45:226–234 Shoja Z, Tagliamonte M, Jalilvand S, Mokhtari-Azad T, Hamkar R, Maldonado J, Cao S, Zhang W, Mansky L (2016) Distinct morphology of Shahmahmoodi S, Rezaei F, Tornesello M, Buonaguro FM, human T-cell leukemia virus type 1-like particles. Viruses 8:1–11 Buonaguro L, Nategh R (2013) Development of a stable insect cell Maranga L, Brazao TF, Carrondo MJT (2003) Virus-like particle produc- line constitutively expressing rotavirus VP2. Virus Res 172:66–74 tion at low multiplicities of infection with the baculovirus insect cell Tagliamonte M, Visciano ML, Tornesello ML, De Stradis A, Buonaguro system. Biotechnol Bioeng 84:245–253 FM, Buonaguro L (2010) Constitutive expression of HIV-VLPs in Martin JL, Cao S, Maldonado JO, Zhang W, Mansky LM (2016) Distinct stably transfected insect cell line for efficient delivery system. particle morphologies revealed through comparative parallel analy- Vaccine 28:6417–6424 ses of retrovirus-like particles. J Virol 90:8074–8084 Thompson CM, Petiot E, Mullick A, Aucoin MG, Henry O, Kamen AA Mohsen MO, Zha L, Cabral-Miranda G, Bachmann MF (2017) Major (2015) Critical assessment of influenza VLP production in Sf9 and findings and recent advances in virus–like particle (VLP)-based HEK293 expression systems. BMC Biotechnol 15:1–12 vaccines. Semin Immunol 34:123–132 Venereo-Sanchez A, Gilbert R, Simoneau M, Caron A, Chahal P, Chen Monteiro F, Bernal V, Chaillet M, Berger I, Alves PM (2016) Targeted W, Ansorge S, Li X, Henry O, Kamen A (2016) Hemagglutinin and supplementation design for improved production and quality of neuraminidase containing virus-like particles produced in HEK-293 enveloped viral particles in insect cell-baculovirus expression sys- suspension culture: an effective influenza vaccine candidate. tem. J Biotechnol 233:34–41 Vaccine 34:3371–3380 Nika L, Wallner J, Palmberger D, Koczka K, Vorauer-Uhl K, Grabherr R Vidigal J, Fernandes B, Dias MM, Patrone M, Roldão A, Carrondo MJT, (2017) Expression of full-length HER2 protein in Sf9 insect cells Alves PM, Teixeira AP (2018) RMCE-based insect cell platform to and its presentation on the surface of budded virus-like particles. produce membrane proteins captured on HIV-1 Gag virus-like par- Protein Expr Purif 136:27–38 ticles. Appl Microbiol Biotechnol 102:655–666 Nika L, Cuadrado-Castano S, Arunkumar GA, Grünwald-Gruber C, Visciano ML, Diomede L, Tagliamonte M, Tornesello ML, Asti V, McMahon M, Koczka K, García-Sastre A, Krammer F, Grabherr Bomsel M, Buonaguro FM, Lopalco L, Buonaguro L (2011) R (2019) An HER2-displaying virus-like particle vaccine protects Generation of HIV-1 virus-like particles expressing different HIV- from challenge with mammary carcinoma cells in a mouse model. 1 glycoproteins. Vaccine 29:4903–4912 5 Vaccines 7:1–19 Wilde M, Klausberger M, Palmberger D, Ernst W, Grabherr R (2014) Pastor AR, González-Domínguez G, Díaz-Salinas MA, Ramírez OT, Tnao38, High Five and Sf9-evaluation of host-virus interactions in Palomares LA (2019) Defining the multiplicity and time of infection three different insect cell lines: baculovirus production and recom- for the production of Zaire Ebola virus-like particles in the insect binant protein expression. Biotechnol Lett 36:743–749 cell-baculovirus expression system. Vaccine 37:6962–6969 Young KR, Ross TM (2006) Elicitation of immunity to HIV type 1 Gag is Pereira Aguilar P, González-Domínguez I, Schneider TA, Gòdia F, determined by Gag structure. AIDS Res Hum Retrovir 22:99–108 Cervera L, Jungbauer A (2019) At-line multi-angle light scattering detector for faster process development in enveloped virus-like par- Publisher’s note Springer Nature remains neutral with regard to jurisdic- ticle purification. J Sep Sci 42:2640–2649 tional claims in published maps and institutional affiliations.