Biotechnology Advances 60 (2022) 108017 Contents lists available at ScienceDirect Biotechnology Advances journal homepage: www.elsevier.com/locate/biotechadv Research review paper The cell density effect in animal cell-based bioprocessing: Questions, insights and perspectives Jesús Lavado-García a, b, 1, *, Pol Pérez-Rubio a, 1, Laura Cervera a, Francesc Gòdia a a Grup d’Enginyeria Cel⋅lular i Bioprocessos, Escola d’Enginyeria, Universitat Autònoma de Barcelona, Campus de Bellaterra, Cerdanyola del Vallès, 08193 Barcelona, Spain b Novo Nordisk Foundation Center for Biosustainability, Technical University of Denmark, 2800 Kgs. Lyngby, Denmark A R T I C L E I N F O A B S T R A C T Keywords: One of the main challenges in the development of bioprocesses based on cell transient expression is the Cell density effect commonly reported reduction of cell specific productivity at increasing cell densities. This is generally known as Infection the cell density effect (CDE). Many efforts have been devoted to understanding the cell metabolic implications to Transient transfection this phenomenon in an attempt to design operational strategies to overcome it. A comprehensive analysis of the Perfusion main studies regarding the CDE is provided in this work to better define the elements comprising its cause and impact. Then, examples of methodologies and approaches employed to achieve successful transient expression at high cell densities (HCD) are thoroughly reviewed. A critical assessment of the limitations of the reported studies in the understanding of the CDE is presented, covering the leading hypothesis of the molecular implications. The overall analysis of previous work on CDE may offer useful insights for further research into manufacturing of biologics. 1. Introduction weeks (Gutiérrez-Granados et al., 2018). The accessible and straight­ forward implementation of TGE together with its reduced costs and time The emergence of new diseases and viral outbreaks that jeopardize requirements (Geisse and Fux, 2009) have made TGE a valuable tech­ society demand technologies for biomanufacturing of pharmaceuticals nology. More importantly, TGE offers the needed versatility to adapt to to be able to adapt to changing situations and new challenges. The different situations and market needs as production can switch from one versatility of technologies like transient gene expression (TGE) offers a type of product to a different one using the same cellular platform solution for this problematic. TGE is commonly used for lab-scale and varying only the transfected plasmid DNA. Despite having been ach­ preclinical trials batches production of proteins of interest, screening ieved at large volumes and using GMP-compliant bioprocesses (Merten tests and for manufacturing of specific biopharmaceuticals like adeno- et al., 2011; Tuvesson et al., 2008), SGE is still the preferred technology associated viruses (AAVs) (Robert et al., 2017) for the growing gene for large-scale production of biopharmaceuticals in well-established therapy field. Transiently expressing heterologous genes that remain in bioprocesses (Lai et al., 2013). Nevertheless, the rise of personalized episomal form in the host cell provide certain advantages compared to medicine based on individual genetic and epigenetic biomarkers, the process of stable gene expression (SGE). The generation of a stable requiring tailored-made treatments and therapies, has shifted the focus cell line via integration of a gene in the host cell coding for the protein of back to TGE (Ramaswami et al., 2018). The adaptability and versatility interest is an industrious, time-consuming process taking from 3 to 12 for drug variants production have made TGE a promising candidate for months (Bandaranayake and Almo, 2014; Turan et al., 2011) while the the design of cost-effective bioprocesses for individualized treatments. use of TGE can provide high titers of the biological product of interest in Apart from the challenge of the GMP-compliant production of large Abbreviations: AAV, Adeno-associated virus; ATF, Alternating tangential filtration; BEVS, Baculovirus expression vector sytem; CCD, Central composite design; CDE, Cell density effect; CSPR, Cell specific perfusion rate; DMA, Dimethylacetamide; DoE, Design of experiments; HCD, High cell density; HFM, Hollow fiber module; Hpt, Hours post-transfection; LCD, Low cell density; MFA, Metabolic flux analysis; MOI, Multiplicity of infection; MR, Medium replacement; MVA, Modified vaccinia Ankara; SGE, Stable gene expression; TGE, Transient gene expression. * Corresponding author. E-mail address: jlavgar@dtu.dk (J. Lavado-García). 1 These two authors contributed equally to this work. https://doi.org/10.1016/j.biotechadv.2022.108017 Received 15 March 2022; Received in revised form 31 May 2022; Accepted 1 July 2022 Available online 7 July 2022 0734-9750/© 2022 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by- nc-nd/4.0/). J. Lavado-García et al. Biotechnology Advances 60 (2022) 108017 amounts of DNA plasmids required for TGE, one of the main drawbacks process performance at HCD. preventing TGE from being thoroughly exploited at large scale is the so- called cell density effect (CDE). Although the CDE presents subtle nu­ 3.1. Insect cells ances and controversy regarding its definition, it can be generally accepted as a reduction in cell productivity for cultures as transient The two main used insect cell lines are the High Five and Sf9 cells expression is carried out at increasing densities (Backliwal et al., 2008; (Yee et al., 2018) together with the baculovirus expression vector sytem Cervera et al., 2013; Lavado-García et al., 2020b; Le Ru et al., 2010). (BEVS) (Felberbaum, 2015). Optimal infection cell densities for High This effect is also reported in insect cells when working with baculovirus Five batch-cultured in Sf900III medium have been reported at around infection (Bernal et al., 2009; Huynh et al., 2015) and in avian and 2⋅106 cells/mL with a MOI of 2.5 (Puente-Massaguer et al., 2020). When mammalian cells when working with viral infection (Genzel et al., cultured in Insect-XPRESS medium, fed-batch supplementation with 2014). The same observation regarding a decrease in productivity in cysteine increased products titers when infected at 2.5⋅106 cells/mL terms of bioprocess yield is attributed to the CDE as a term encom­ with a MOI of 5 (Patrone et al., 2014). A pseudo-perfusion strategy has passing many different molecular, metabolic, physiological and physi­ permitted fruitful infection process development for High Five cells at cochemical components that are often intertwined while the explanation MOI of 3 in the Insect-XPRESS serum-free medium at 4⋅106 cells/mL of the phenomenon remains unclear. This review aims to thoroughly (Rausch et al., 2013). Regarding Sf9 cells, when batch-cultured in Sf900- analyze the reported bioprocess encounters with the CDE, discuss the II medium, they are correctly infected at cell densities of 1⋅106 cells/mL concerns that it poses for the future of the implementation of TGE with either MOIs of 0.6 or 30 (Bernal et al., 2009; Carinhas et al., 2009). technologies at large scale as well as reviewing successful examples In the same trend Liu et al. found optimal infection cell density in batch overcoming the CDE to enclose the main factors governing it and culture for Sf21 cells in DMEM medium at 1⋅106 cells/mL with a MOI of examine potential solutions for the future. 5 (Liu et al., 2010). Meghrous et al. also reported good infection rates with Sf9 in Ex-Cell 420/405 with a MOI of 5 at 1⋅106 cells/mL 2. The problem defining high cell density (Meghrous et al., 2005). Newly developed insect cell platforms kits such as the ExpiSf9 perform much better having optimal infection cell den­ The first step to better define the CDE is to define what is considered sities with a MOI of 1 at 5⋅106 cells/mL (Kurasawa et al., 2020). Fed- high cell density (HCD) in animal cell technology when performing TGE. batch production of AAV by Sf9 cells in Sf900-II medium infected at a Considering batch culture mode, transient transfection is commonly MOI of 3 allowed maintenance of specific viral yield at cell densities of reported to be carried out at ~2⋅106 cells/mL (Baldi et al., 2005; Cervera 10⋅106 cells/mL if compared to batch cultivation (Joshi et al., 2019). et al., 2015a, 2013; Rajendra et al., 2015c) as cells do not usually grow Moreover, process intensification in Sf9 cells allowed the delaying of the beyond 5–6⋅106 cells/mL. Therefore, the effects related to HCD state are CDE at densities higher than 27⋅106 cells/mL. When cultured in Ex-Cell observed at ~5⋅106 cells/mL, hence these are referred to as HCD 401 medium at 27⋅106 cells/mL and infected at MOI of 1, Sf9 cells dis­ (Bleckmann et al., 2019; Cervera et al., 2013; Gutiérrez-Granados et al., played similar performance to that of a low cell density (LCD) batch 2016; Lavado-García et al., 2020b; Rajendra, 2018). However, advances culture (Zhang et al., 1998). in culture media allow cultivation in batch up to 8–12⋅106 cells/mL (Lavado-García et al., 2020a) starting to observe the effects related to 3.2. Mammalian cells: Chinese hamster ovary (CHO) cells HCD at cell densities of ~10⋅106 cells/mL (Elshereef et al., 2019; Mar­ anga et al., 2005; Nadeau et al., 2000; Stuible et al., 2018). There are Leading animal cell technology research (Walsh, 2018), CHO cells studies that even attribute the term HCD to cells at ~1⋅106 cells/mL are the cell platform with one of the highest reported cell densities, (Durocher et al., 2002) or even state HCD without specifying the actual >200⋅106 cells/mL in a perfusion wave bioreactor (M. F. Clincke et al., cell density for this condition (Park et al., 2015). Also, the cell density 2013a, 2013b; Zhang et al., 2015). Although the raising interest in that can be reached in batch cultivation depends on the specific animal single-use perfusion-based processes, fed-batch remains as the preferred cell platform, varying among mammalian, avian and insect cells. operational strategy for bioprocess development in diverse CHO cell Nevertheless, in any case, the limitations in cell density to successfully lines (Fan et al., 2018), routinely achieving 20⋅106 cells/mL (Huang transfect or infect cell cultures in batch mode are observed when tran­ et al., 2010; Stepper et al., 2020; Wiegmann et al., 2019). Recent fed- sient expression is carried out beyond mid-exponential growth phase. batch intensification strategies involving HCD seeding approaches Likewise, when working in fed-batch and perfusion mode, the expo­ with high-producing stable CHO cell lines led to HCD cultivations over nential growth phase reaches higher densities, delaying the appearance 30⋅106 cells/mL (Yongky et al., 2019). In batch cultivation, CHO cells of the CDE, and in this frame, the term HCD is shifted to ~50⋅106 cells/ have reported to achieve cell densities of ~8⋅106 cells/mL (Kuwae et al., mL and up to ~200⋅106 cells/mL (M. F. Clincke et al., 2013b, 2013a). 2018; Shang et al., 2021). Considering all these aspects, it is rather challenging to precisely define HCD, complicating the comparison of conclusions drawn in the different 3.3. Mammalian cells: Human embryonic kidney (HEK) cells studies reporting working at HCD. Consequently, the term HCD should always be relative to the cell culture medium, the cell line used and the HEK293 cells are the preferred platform for TGE. The restriction that mode of operation, ideally stating what actual cell density the HCD the CDE poses to efficiently achieve transient expression at HCD value is being compared to. together with the advantages of CHO cells on the field of SGE, few studies have been devoted to develop HCD cultivations in HEK293 cells 3. Animal cell platforms for bioprocesses working at HCD (Tan et al., 2021). However, Cervera et al. reported HEK293SF-3F6 cells cultivated in batch to peak at cell densities ranging from 3 to 4.5⋅106 Substantial progresses on the field of cell and culture medium engi­ cells/mL in diverse commercially-available culture media and at 5.4⋅106 neering have broaden the possibilities of developing bioprocesses with a cells/mL upon supplementation with a lipid mixture in FreeStyle293 wide range of animal cell lines (Tripathi and Shrivastava, 2019). medium (Cervera et al., 2013). Likewise, similar cell densities up to Although the increased diversification, the CDE is a common concerning 3.5⋅106 and 5⋅106 cells/mL were obtained in batch culture with a clone factor that influences several TGE-based processes regardless of the cell of HEK293 stably expressing the E2-CD154 protein in 4 commercially- origin and expression system (Dill et al., 2019; Genzel et al., 2014; available culture media (Lorenzo et al., 2019) and for wild type Huynh et al., 2013; Lavado-García et al., 2020a; Wu et al., 2020). This HEK293 and Bax-Bak KO HEK293 cell line grown in HyCell TransFX-H section of the review aims to summarize and discuss the differences of medium (Arena et al., 2019) respectively. HEK293-E6 stably express­ the main animal production platforms in order to address the limits of ing the PYC enzyme and IFNα2b grown in FreeStyle F17 medium and 2 J. Lavado-García et al. Biotechnology Advances 60 (2022) 108017 operated in fed-batch reached cell densities of 11⋅106 cells/mL (Vallée based TGE strategies reporting the observation of the CDE. The principal et al., 2014). Also, fed-batch strategies based upon on-line measure­ reported feature addressing the CDE when using these methods is the ments allowed cell densities of 19⋅106 cells/mL in HEK293-3F6 cells reduction of transfection efficiency when transfecting increasing cell cultured in SFMTransFx-293 medium, although supplemented with 5% densities (Bleckmann et al., 2019; Elshereef et al., 2019; Lavado-García FBS (Martínez-Monge et al., 2018). Using a perfusion approach, et al., 2020a, 2020b). Alternative approaches skipping the DNA:PEI HEK293F cells stably expressing EPO in BalanCD HEK293 medium complexation step have been studied, aiming to avoid the CDE and reached of the highest cell densities reported in this cell line, at 80⋅106 achieve higher cell density successful transfection (Backliwal et al., cells/mL (Schwarz et al., 2020). 2008; Blackstock et al., 2020; Rajendra, 2018). Although transfection may be achieved at HCD using DNA complexation-based transient 3.4. Avian cells expression systems, transfection efficiency is reduced and remains one of the main challenges for large-scale transient biomanufacturing. Interest in avian cells has risen given their capacity for large-scale manufacturing of vaccines, aiming at replacing the production pro­ 4.2. Electroporation cesses relying on embryonated chicken eggs (Kraus et al., 2011). Batch cultivation of AGE.CR.pIX in CD-U2 medium sustained growth up to Electroporation is also a methodology for transient expression usu­ 14⋅106 cells/mL which was increased to 17⋅106 cells/mL after adapta­ ally employed at small-scale due to its requirements. Since cells need to tion to glutamine-starving conditions (Lohr et al., 2014), revealing be concentrated in a small volume for the use of electroporation cuvettes greater growth capacity compared to CHO or HEK cells. Comparatively, or microfluidic devices, cell densities from 0.5 to 2⋅106 cells/mL are used novel avian cell lines such as the DuckCelt – T17, only achieved 5⋅106 (Bendix et al., 2021; Hur et al., 2020; Zu et al., 2016). This prevent this cells/mL cultured in Optipro SFM (Petiot et al., 2018). Process intensi­ method from being scaled-up and implemented for industrial production fication resulted in HCD cultivations of 50⋅106 cells/mL for AGE1.CR. at large-scale. Some advances have been made recently where the pIX cells in CD-U3 medium (Genzel et al., 2014; Gränicher et al., 2021; implementation of flow electroporation has allowed the successful Vázquez-Ramírez et al., 2018). electroporation of 107 cells/mL cultures in a volume of 2.8 L (Steger et al., 2015). However, the requirement of subsequent intensive MR 4. TGE systems and their differences regarding the CDE steps and the difficulty for further scaling-up have shifted the use of electroporation as an alternative for difficult transfections at small-scale, 4.1. DNA complexation-based transient expression systems focused on single-cell electroporation (Agarwal et al., 2007; Wang et al., 2010). To the best of our knowledge, electroporation has not been Despite the great demand of plasmid DNA required for large-scale thoroughly studied at high working cell densities neither the observa­ transient transfection, and the need to produce large amounts of GMP- tion of the CDE has been reported, since the conditions and requirement compliant plasmids, DNA complexation-based transient expression sys­ for its application do not lead to the appearance of the CDE. tems are the most widely used method for TGE (Gutiérrez-Granados et al., 2018). Calcium phosphate, cationic lipids or cationic polymers are 4.3. Viral transient expression systems the most used transfection reagents. Calcium phosphate presents dis­ advantages when aiming to achieve HCD cultures for large-scale pro­ The use of viral vectors for transient expression is widely extended in duction as it requires culture media containing serum in order to the bioprocessing field. The main disadvantage of these systems is the counterbalance the cytotoxic effects of the Ca3 (PO4)2 particles (Geisse, requirement to constantly rely on virus expansion, titration, storage and 2009). Also, Ca3 (PO4)2 particles promote aggregation, complicating its manipulation, leading to biosafety concerns (Jäger et al., 2015). How­ use with HCD cultures (Park et al., 2006). Lipid-mediated strategies, like ever, the design of modified and attenuated virus to prevent replication the use of Lipofectamine and derived products are an efficient trans­ in case of accidental release have paved the way for the use of viral fection method. They work as a DNA vector introducing it into the cell vector for TGE. Also, the baculovirus expression system is well stab­ by the formation of liposomes. DNA is mixed with cationic liposomes, lished in the bioprocessing field for the transient production of viral resulting in positively charged lipid vesicles containing the DNA of in­ vectors, being recently reviewed by Joshi et al. (Joshi et al., 2021). The terest. These vesicles interact with the negatively charged cell mem­ advantage of the baculovirus system is the possibility of combining it brane and fuse, releasing the DNA into de cytoplasm. However, they with insect cells lines such as Sf9 and HighFive or using it for TGE of present a major bottleneck in the step of plasmid translocation from the recombinant proteins in mammalian cell lines, where its inability for cytosol to the nucleus (Haldankar et al., 2006; Yamano et al., 2010). replication is presented as a biosafety advantage. Both strategies offer a Also, the high price of these molecules prevents this method from being scalable high-yield production although still rely on costly generation of fully exploited at large scale and its use remains at lab-scale and for recombinant baculoviruses (Ames et al., 2007; Joshi et al., 2021; Tsai research purposes (Haldankar et al., 2006; Liu et al., 2008). et al., 2020). Unlike DNA complexation-based TGE strategies, the CDE is The broad research on TGE is reflected in the number of studies using reported as the reduction of cell-specific productivity of viral vectors at cationic polymer-mediated transfection, clearly outnumbering the re­ increasing cell densities when working with viral transient expression ports of any other methodology (Gutiérrez-Granados et al., 2018; Jäger systems. This reduction was observed when using the baculovirus et al., 2015). Cationic polymers, like polyethyleneimine (PEI), work expression system to produce AAVs. Although production was achieved differently. They are positively charged molecules that complex with at higher cell densities and total titers improved compared to a LCD DNA, forming positively charged polyplexes that interact with the cell production, cell-specific productivity was reduced (Joshi et al., 2021; membrane. After endocytosis, DNA:PEI polyplexes leave the endosomal Meghrous et al., 2005; Urabe et al., 2002). Moreover, bioprocesses for compartment and are released to the cytoplasm due to the buffer ca­ viral vector production such as adenoviruses or influenza virus for pacity of PEI, which causes the collapse of intracellular endosomes vaccine development in mammalian cell cultures, which are grown to (Benjaminsen et al., 2013; Freeman et al., 2013). Then, the polyplexes high cell densities in bioreactors before infection, have also reported the need to reach the nucleus. The use of PEI presents clear advantages in same effect on increasing cell densities. These have been currently terms of scalability, cell toxicity and reproducibility compared to the thoroughly reviewed elsewhere (Silva et al., 2021; Tan et al., 2021; other methods (Geisse, 2009; Gutiérrez-Granados et al., 2018; Jäger Watanabe et al., 2021). This effect has been reported in all mammalian, et al., 2015). Both lipid-based and PEI-based strategies rely on analogous avian and insect cells platforms. One of the main common challenges pathways to deliver DNA to the recipient cell (Elouahabi and Ruys­ that they face is the ability to maintain cell-specific productivity or cell- schaert, 2004) and these two methods are the main DNA complexation- specific virus yield when infections are carried out at high cell densities. 3 J. Lavado-García et al. Biotechnology Advances 60 (2022) 108017 Henceforth, this review will focus on the different studies reporting HCD, they proposed that continuous perfusion processes where nutri­ having encountered the CDE since its first mention in the scientific ents are constantly replenished and by-products removed, could offer a literature until the most recent -omics studies designed to uncover the solution and overcome the CDE. In order to test this, Henry et al. carried source of the problem. Finally, the different strategies that have been out adenovirus productions infecting HEK293 cells in continuous reported to alleviate the CDE, achieving successful transient expression perfusion mode using an acoustic filter (Henry et al., 2004). Here, as the at HCD will be analyzed. exponential growth phase reached higher densities, successful infections with no reduction in cell-specific production yield could be achieved up 5. The evolution of the CDE research to 3⋅106 cells/mL at the moment of infection. However, the CDE still appeared, this time at higher densities. They observed that when Back in 1990–1994, Semans & Hu, Banik & Heath and Wohlpart infecting at 6⋅106 cells/mL, specific productivity yields decreased 5-fold. et al. observed that the specific productivity of their antibody produc­ After analyzing metabolite consumption, they hypothesized that the tion of hybridoma cultures in perfusion decreased as cell density CDE was due to a reduction in metabolic activity upon infection and that increased (Banik and Heath, 1994; Craig Seamans and Hu, 1990; the physiological state of the cell was the key parameter for a successful Wohlpart et al., 1990).They hypothesized that it was caused by the lack infection at HCD. Based on these results, they proposed to increase the of nutrients. Following this path, Zeng et al. constructed several kinetic perfusion rate in order to delay the CDE and infect at HCD. However, a and metabolic models to study nutrients availability in a series of works substantial increase in perfusion rates is a solution that poses a challenge in 1995 and 1996 (Zeng, 1995; Zeng and Deckwer, 1995; Zeng, 1996a, to be transferred to industrial scale due to the increase in the economic 1996b). After analyzing consumption rates of glucose and glutamine and cost of the process and due to technical difficulties to maintain cell production rates of lactate, ammonium and the antibody of interest, they viability caused by shear stress in the cell retention device at high concluded that even though consumption rates depended on cell den­ perfusion rates. Furthermore, despite being able to delay the CDE, their sity, there was no factor different from the cell density itself that could results did not investigate its molecular causes. be identified as the cause of the observed effect. They were even skep­ The CDE was also studied by Yuk et al. using HeLaS3 cells in fed- tical about the existence of this effect and hypothesized that this batch and perfusion mode to produce oncolytic adenoviral vectors observation was an artifact of miscalculations and measurement errors. (Yuk et al., 2004). They observed that in fed-batch, despite using me­ They also proposed that if the effect was indeed real, it was being caused dium concentrates in feeding strategies allowing the culture to reach by a component of the culture medium not yet identified. Despite these 5⋅106 cells/mL, no viruses were produced if the cell density at the discrepancies, the CDE has been commonly reported in the bio­ moment of infection was above 2.8⋅106 cells/mL. They reported that all processing field as the reduction in cell productivity at HCD (Fig. 1). attempts to infect at higher densities failed, even though cell viability Although first described in perfusion at high cell densities, the reduction was above 95%. Using continuous perfusion, they successfully achieved in cell-specific productivity as cell density increased has been widely infection at a maximum of 14.8⋅106 cells/mL with a cell-specific virus reported and studied in transient expression systems, as it presented the yield of 3.2⋅104 viral particles/cell. This was a decrease in cell-specific main drawback for scaling up this type of bioprocesses. This effect has virus yield from the perfusion culture infected at 5.2⋅106 cells/mL, been extensively observed in the production of adenoviral vectors by which yielded 4.9⋅104 viral particles/cell. Furthermore, they did not infection using several cell platforms such as HeLaS3, PER.C6, HEK293, increase cell-specific virus yield in the 14.8⋅106 cells/mL perfusion CAP and AGE1.CR (Genzel et al., 2014; Henry et al., 2005; Maranga compared to successful infections in batch and fed-batch, which were et al., 2005; Yuk et al., 2004). Reviewing the bioprocesses and optimi­ reported to provide 3.4⋅104 and 3.7⋅104 viral particles/cell respectively, zation methods for adenoviral production, Nadeau and Kamen observed in agreement with previous reported results using perfusion. They ruled that most of the efforts to increase cell-specific productivity at HCD out the hypothesis of nutrient scarcity as the physiological cause of the encountered the limitation of the CDE. They claimed that the impedi­ CDE and proposed that it was caused by the presence and accumulation ment preventing a successful infection yields beyond ~2⋅106 cells/mL of an unknown inhibitor, as previously proposed by Nadeau and Kamen was due to a nutrient limitation and/or a by-product inhibition (Nadeau (Nadeau and Kamen, 2003). In order to further investigate the role of and Kamen, 2003). Having observed that the optimization attempts nutrient consumption in the CDE, Maranga et al. studied glucose and carried out in fed-batch could not achieve the desired productivity at amino acid metabolism upon infection in PER.C6 cells to produce Fig. 1. Timeline of the cell density effect (CDE) research. Milestones of representative works in the study of the CDE in transient expression systems. 4 J. Lavado-García et al. Biotechnology Advances 60 (2022) 108017 adenovirus encoding HIV-1 p55 gag gene at different cell densities Most of the works regarding TGE at higher densities perform a (Maranga et al., 2005). They infected at increasing consecutive cell concentration step from lower cell densities to be able to achieve densities of 1, 2, 3, 9 and 10⋅106 cells/mL performing the necessary transfection. Shen et al., Sun et al. and Ansorge et al. all reported that to medium replacement (MR) steps to guarantee the availability of nutri­ transfect higher cell densities, at 30, 10 and 5⋅106 cells/mL respectively, ents. They analyzed metabolites consumption upon infection at the a concentration step by centrifugation was needed prior to the trans­ different cell densities and concluded that at cell densities above 10⋅106 fection step (Ansorge et al., 2009; Shen et al., 2013; Sun et al., 2008). All cells/mL there was no nutritional limitation and still specific produc­ these PEI-mediated transfections required this previous concentration tivity was strongly reduced. They reported a total maximum normalized step to successfully transfect at higher cell densities regardless of the cell particle titer of 1.23 ± 0.03 when infecting at 1⋅106 cells/mL and a total line and biologic produced. Ansorge et al. reported production of titer of 1.1 ± 0.2 when infecting at 9⋅106 cells/mL, clearly cell-specific lentivirus in HEK293 cells (Ansorge et al., 2009), Shen et al. reported productivity was being reduced. There was still an unknown cause for production of recombinant tumor necrosis factor receptor fused to an Fc the CDE that prevented HCD cultures to be successfully infected. They domain (TNFR-Fc) in Sf9 insect cells (Shen et al., 2013) and Sun et al. concluded that an unidentified cell physiological trait was responsible reported EPO production in HEK293 EBNA cells (Sun et al., 2008). for the CDE. Regardless of the production platform and product, the implementation Recent advances in cell retention devices for perfusion such as the of a concentration and MR step seemed to alleviate the CDE. alternating tangential flow filtration (ATF) and hollow fiber modules In conclusion, the fact that MR is required to successfully achieve (HFM) have allowed cell culture to achieve higher cell densities, transient expression at high cell densities and that there are reported therefore being able to delay even more the CDE. Genzel et al. reported unidentified limiting physiological elements, suggests that there are two successful adenoviral infection using AGE1.CR and CAP cells at 25 and differential components influencing the CDE: the presence of an inhib­ 27⋅106 cells/mL respectively (Genzel et al., 2014). However, a reduction itory component in the culture medium and a cell physiological condi­ in cell-specific yield, attributed to the CDE was observed at 47 and tion. However, no reported work has directly targeted these molecular 33⋅106 cells/mL respectively for the AGE1.CR and CAP cells perfusion causes. These two components influencing the CDE have been usually cultures. AGE1.CR cells decreased their specific yield from 1708 virions/ tackled in different ways. A common approach to remove possible cell at 25⋅106 cells/mL to 1266 virions/cell when infection was carried inhibitory by-products from the cell culture at the moment of trans­ out at 47⋅106 cells/mL. For CAP cells, 4086 virions/cell at 27⋅106 cells/ fection or infection, is to carry out a MR step (Ikonomou et al., 2004). mL decreased to 1883 virions/cell at 33⋅106 cells/mL. This is usually employed also to renew nutrients and to switch to a new Summing up, the variability of cell densities at which the CDE has culture medium in case the medium used for growth cannot support the been observed depending on the culture mode (batch, fed-batch and process of transfection or infection. There are several examples of this in perfusion), together with the ambiguity of the term “high cell density”, the literature ranging from transfection in mammalian cells (Backliwal contribute to the confusion when it comes to defining the CDE. Advances et al., 2008; Gutiérrez-Granados et al., 2016; Kadlecova et al., 2012; in perfusion technology and use of MR may lead to the successful Powers et al., 2016; Rajendra et al., 2015a, 2011; Ye et al., 2009), implementation of the bioprocess at higher cell densities. The fact that a infection in mammalian cells (Frazzati-Gallina et al., 2001; Garnier MR step allows infection at cell densities that otherwise would not be et al., 1995; Venereo-Sanchez et al., 2017) and infection in insect cells successfully infected contribute to the theory of an unknown inhibitory (Bernal et al., 2009; Elias et al., 2000; Ikonomou et al., 2003; Meghrous compound. However, the molecular causes for a sudden reduction in et al., 2005; Palomares et al., 2005). It can be easily performed by cell-specific productivity are still unknown. centrifugation and resuspension at small scale and by the implementa­ As aforementioned, the CDE has been also reported in transient tion of perfusion-based bioprocesses using cell retention devices at transfection. Cervera et al. reported that PEI-mediated transfection of larger scale. HEK293 cells to produce HIV-1 Gag virus-like particles (VLPs) without The other component, the cell physiological state, has also been re­ MR at 4⋅106 cells/mL was unsuccessful and with a MR step at 4⋅106 ported to be a critical parameter for optimal transient expression pro­ cells/mL, transfection efficiency was significantly reduced compared to cesses. Carpentier et al. and Rajendra et al. both studied the delivery of transfection at 1⋅106 cells/mL upon MR step. Transfection efficiency plasmid DNA in PEI-mediated transient transfections in HEK293-EBNA dropped from ~55% when transfection was carried out at 1⋅106 cells/ and HEK293 and CHO cells respectively. They studied the efficiency mL to ~20% when it was carried out at 4⋅106 cells/mL. (Cervera et al., of plasmid copy delivery at intracellular level as well as the transcription 2013). They considered that this reduction was caused by sub­ efficiency, analyzing the limitations of TGE processes. They concluded populations of the cell culture in different cell cycle phases, more spe­ that the cell physiological state was a influencing factor for successful cifically the number of cells in G2/M phase. To study the effect of cell transient expression (Carpentier et al., 2007; Rajendra et al., 2015c), cycle at different cell densities, dilution and concentration steps were independently of nutrient scarcity or inhibitory compounds in the cul­ carried out to transfect cells at 4⋅106 cells/mL being concentrated from ture medium. 1⋅106 cells/mL and to transfect at 1⋅106 cells/mL being diluted from As an indicator of cell population health, cell viability is usually 4⋅106 cells/mL. They observed that concentrating cells from 1 to 4⋅106 monitored. However, proteomic studies of HCD and transfected cells cells/mL reduced transfection efficiency 20% at 24 h post-transfection reporting an overall disruption of cell homeostasis have shown that (hpt) even though the same percentage of the cell population was at more variables are needed to determine a healthy cell physiological state G2/M phase. Moreover, taking cells at 4⋅106 cells/mL and diluting them (Lavado-García et al., 2020b). Moreover, passage number have also to 1⋅106 cells/mL allowed an increase of 22% in transfection efficiency been reported to influence transfection processes, leading to a decrease at 24 hpt. They concluded that there were different cell physiological in productivity when employing cells coming a from a high passage factors apart from the cell cycle affecting transfection. Similar tests to number (Grieger et al., 2016). select an optimal cell density for PEI-mediated transient transfection in Different approaches combining MR steps and the use of metaboli­ CAP-T cells to produce HIV-1 Gag VLPs were carried out by Gutiérrez- cally active cells have been reported to achieve successful transfections Granados et al. where cell densities from 2 to 8⋅106 cells/mL were tested and infections at HCD and delay the onset of the CDE. The role of cell at the moment of transfection (Gutiérrez-Granados et al., 2016). They metabolism in the CDE is further discussed below. observed that the optimal cell density for CAP-T cells transfection was 3.3⋅106 cells/mL and specific productivity steeply dropped. They 6. The role of energy metabolism in the CDE analyzed a design of experiment response surface and concluded that total VLP titer did not increase with cell density. Therefore, inferring The initial observations associating the CDE to a nutrient-depleted that cell-specific VLP yield decreased with increasing cell density. medium or a metabolic waste by-product have led to several studies 5 J. Lavado-García et al. Biotechnology Advances 60 (2022) 108017 analyzing cell metabolism beyond kinetics and consumption rates to­ upon transfection, including a downregulation in glycosphingolipid wards developing a model linking metabolism to the events of trans­ metabolic processes, AMP signaling, response to external stimuli and fection and infection at high cell densities (Petiot et al., 2015). Henry enzyme binding processes among others. All these changes upon trans­ et al. constructed a metabolic model of HEK293 cells to study intracel­ fection were reported by the authors to overall disrupt homeostasis and lular fluxes of metabolites upon infection in a continuous perfusion suggested that combined with the physiological changes observed in system (Henry et al., 2005). They observed that successful infections increasing cell densities may contribute to prevent a successful trans­ were characterized by a state of high ATP production rate in the cells. fection at HCD, partly explaining the physiological component of the They also reported that cell specific productivity increased with CDE. increasing perfusion rates. However, none of these findings seemed to All reported works analyzing cell metabolism regarding increasing explain the nature of the limitation posed by the CDE, which remained cell densities coincide in the existence of a metabolic shift upon infection unknown. They continued to propose that an unidentified cell physio­ and transient transfection that defines cell productivity. Nevertheless, logical factor was crucial in the appearance of the CDE. Bernal et al. also no causal link allowing to relate specific metabolic changes to the CDE studied the energy metabolism upon baculovirus infection of Sf9 insect were found. cells (Bernal et al., 2009). They carried out metabolic flux analyses (MFA) and observed metabolic downregulations in pathways regarding 7. Achieving successful transient expression at high cell energy metabolism such as glycolysis, TCA and its anaplerotic reactions densities with increasing cell densities. In agreement with Henry et al., they also observed that ATP production rate decreased at HCD. However, this During the last decades several efforts have been made to improve correlated with the fact that growing cells need a higher rate of energy the densities at which cells can be transfected or infected (Petiot et al., production while cell maintenance required lower rates. Moreover, the 2015). Assuming that cell specific productivity would be maintained, CDE was observed before any component of the cell culture medium was increasing cell density would be translated in a boost in volumetric exhausted and before cells reached stationary phase. They concluded productivity (Huang et al., 2010; Xu et al., 2020; Zhang et al., 2015). that accumulation or depletion of a not-measured compound could ac­ Therefore, achieving good transient expression efficiencies at HCD is count for the CDE and hypothesized that the exhaustion of some un­ crucial in order to develop cost-efficient bioprocesses. In this section, identified oligoelement or cofactor crucial for successful infection may some successful examples of transfection and infection at HCD are be behind the limitations and drop in specific productivity imposed by comprehensively reviewed. A rational analysis of all the steps performed the CDE. before and after TGE provides insights on the basis of the problematic. Nevertheless, as aforementioned, the CDE was still reported in ex­ Study cases are divided depending on the strategy performed (Fig. 2) to amples of perfusion where all nutrients are continuously renewed such sum up the common results in the procedures towards understanding the as the works carried out by Henry et al., Fuenmayor et al., Lavado-García CDE. et al. and Yuk et al. among others (Fuenmayor et al., 2019; Henry et al., 2004; Lavado-García et al., 2020a; Yuk et al., 2004). In these works, the 7.1. Cell concentration followed by transfection and subsequent dilution CDE was only delayed due to the ability of the cell cultures to reach higher densities in a continuous perfusion mode of operation, but cell This methodology was first reported by Wurm et al. back in 2004 as a specific productivity still dramatically decreased even though cells were procedure to routinely transfect both CHO and HEK293 cells, which are still in exponential growth phase. Only a few examples of transfection in currently the gold standards in the production of biologics (Derouazi perfusion at HCD are reported in the literature. One of these works de­ et al., 2004; Walsh, 2018). The protocol consists of transfecting cells at scribes the production of lentiviral vectors expressing GFP in HEK293 2⋅106 cells/mL followed by a 1:2 dilution with fresh culture medium 5 h cells at a cell density of 8⋅106 cells/mL by means of PEI transient after DNA:PEI polyplex addition (Fig. 2a). Dilution could not be per­ transfection (Ansorge et al., 2011). This one included the use of an formed immediately after transfection since nearly 80% of polyplex acoustic filter as a cell retention device. Lentiviral titers decreased from entry occurs 90 min post transfection, meaning that an immediate 1010 vg/mL when transfecting at ~5⋅106 cells/mL to 109 vg/mL when dilution could potentially interfere with polyplex uptake (Cervera et al., transfecting at ~8⋅106 cells/mL. Sodium butyrate was added 16 hpt as it 2017; González-Domínguez et al., 2019). is extensively reported to enhance transfection efficiency (Ansorge et al., Nowadays, to the best of our knowledge, there are no described cases 2011; Caron et al., 2015; Cervera et al., 2015a; Chen et al., 2011; performing dilutions after transfection without a previous step of con­ Damiani et al., 2013; Ghorbaniaghdam et al., 2013; Jeon and Gyun, centration to HCD (Fig. 2b). Blaha et al. compared both transfection 2007; Kiela et al., 2007; Li et al., 2022; Parham et al., 1998; Stuible et al., efficiencies (in terms of GFP positive population) and production of an 2018). Likewise, cell cultures were supplemented with punctual ectodomain of a natural killer receptor LLT1 between HEK293T at 2⋅106 concentrated feed. cells/mL and HEK293GnTI− at 2⋅106 and 20⋅106 cells/mL (Blaha et al., Regarding molecular changes in metabolic pathways in increasing 2015). Cell cultures at HCD were obtained via concentration, then cell densities, Lavado-García et al. carried out a quantitative proteomic transfected and finally diluted to 1⋅106 cells/mL. Transfection effi­ study comparing growing HEK293 cells at 2⋅106 cells/mL and 6⋅106 ciencies of 70% were easily achieved following the conventional LCD cells/mL (Lavado-García et al., 2020b). They reported an upregulation approach for the HEK293T cells yet only 45% efficiency was achieved at in mitochondrial content, reflecting an increase in energy demand, HCD for the HEK293GnTI− cells. Although the total number of trans­ coherently with previous HEK293 metabolic studies (Petiot et al., 2015). fected cells was higher in the HCD condition, therefore achieving a Conversely, they observed that growing cells presented a down­ substantial amount of recombinant protein, transfection efficiency was regulation in proteins involved in cell nuclear trafficking of cargo such in fact reduced. Receptor LLT1 production decreased with the HCD as importins and exportins and in lipid biosynthesis enzymes such as transfection protocol in HEK293GnTI− cells if compared to the pro­ hydroxymethylglutaryl-CoA synthase (HMGCS-1), squalene synthase, duction of HEK293T cells at LCD. Thus, they observed a decrease in acyl-CoA desaturase and downregulation in proteins belonging to ni­ transfection efficiency and also an unexplained drop in specific pro­ trogen metabolism such as glutamine synthetase and thymidylate syn­ ductivity between the two cell lines. In such manner, Rajendra et al. thase. They suggested that these metabolic changes may explain the studied the molecular limitations that arise upon transient transfection physiological cell state preventing transfection at these cell densities as at HCD of CHO and HEK293 cells which resulted in a decline in specific lipid membranes may change as cell density increases. The same study productivity using the aforementioned methodology comprising cell compared untransfected growing cells with transiently transfected cells concentration, transfection and dilution (Rajendra et al., 2015c) at 2⋅106 cells/mL. Changes in energy metabolism were also reported (Table 1). They observed that more plasmid DNA addition resulted in an 6 J. Lavado-García et al. Biotechnology Advances 60 (2022) 108017 Fig. 2. Different strategies and methods to achieve successful transfection or infection at high cell densities (HCD). For simplification, each step mentioning transfection can be equally applied to infection a) Cells are transfected at HCD followed by a dilution with fresh medium at 4–5 h post-transfection (hpt). b) Cells are concentrated to HCD carrying out a total medium replacement (MR), transfected and then diluted with fresh medium at 4–5 hpt. c) Cells at HCD are diluted with fresh medium to low cell densities prior to transfection. d) Cells at low cell density are concentrated to HCD carrying out a total MR. e) Cells are grown using a perfusion-based seed train allowing to reach very high cell densities (~40–50⋅106 cells/mL). These cells are then diluted using fresh medium to lower cell density prior to transfection. increase of transfection efficiency up to a limit of 2⋅104 copies of pDNA/ diluted 1:2 with fresh medium after transfection (Table 1). Central cell. Same trend was observed with intracellular mRNA expression. composite design (CCD) showed that the optimal conditions for DNA Levels above this threshold resulted in a prompted reduction in specific and PEI concentrations were 24 μg/mL and 45 μg/mL, respectively, productivity. They also observed limitations in protein translation pro­ significantly higher than the 1–3 μg/mL of DNA and 2–6 μg/mL of PEI cesses. In order to improve this method, statistical optimization strate­ which are routinely used. These lower concentrations allow each indi­ gies like design of experiments (DoE) have been used as a tool to exploit vidual cell to receive approximately 102–103 copies of pDNA which is the cell metabolic state upon the steps of concentration, transfection and reported to be sufficient for recombinant expression (Cervera et al., subsequent dilution. Elshereef et al. optimized the ratio of DNA:PEI, 2013; Cohen et al., 2009; Gutiérrez-Granados et al., 2016). Therefore, obtaining a transfection efficiency of 60% with cells at 18⋅106 cells/mL combining the methodology of cell concentration, transfection and (Elshereef et al., 2019). They concentrated cells from 3⋅106 cells/mL and dilution with heavily boosting plasmid DNA and transfection reagent 7 J. Lavado-García et al. Table 1 Examples of successful transient transfections at high cell density (HCD) and the different methods and approaches employed to achieve it. MR: medium replacement. Cell Line Cell culture medium Recombinant Amount of Transfection Concentration Step/ Dilution Step Addition Temperature Transfected Cell Product titer Reference Protein DNA Reagent MR prior to (After/Prior) of feed shift after Density transfection transfection Expi293F Expi293 IgG1 1 μg/mL ExpiFectamine No Yes (Prior) Yes No 2.5⋅106 cells/mL 1193 mg/L Jain et al., (HEK293) 293 From 4⋅106 to 2017 2⋅106 cells/mL ExpiCHO-S ExpiCHO IgG1 1 μg/mL ExpiFectamine No Yes (Prior) Yes Yes (37→32 ◦ C) 6⋅106 cells/mL 3279 mg/L Jain et al., CHO From 10⋅106 to 2017 6⋅106 cells/mL CHO-S FreeStyle MAX CHO IgG1 1 μg/mL FreeStyleMAX No Yes (Prior) Yes No 1⋅106 cells/mL 47 mg/L Jain et al., CHO From 1.5⋅106 to 2017 1⋅106 cells/mL HEK293S Ex-Cell293 GFP 20 μg/mL Linear PEI Yes Yes (After) No No 20⋅106 cells/mL ~45% of GFP Blaha et al., GnTI− (25KDa) From 1⋅106 to 20⋅106 From 20⋅106 to positive 2015 cells/mL 1⋅106 cells/mL CHO-3E7 BalanCD Transf. CHO GFP 1.3 μg/mL Linear PEI No Yes (Prior) Yes Yes (37→32 ◦ C) 5⋅106 cells/mL ~30% of GFP Stuible (40KDa) From 7⋅106 to positive et al., 2018 5⋅106 cells/mL CHO K1SV CHO-TNX Propietary mAb 3.8 μg/mL Linear PEI Yes No Yes Yes (37→32 ◦ C) 4⋅106 cells/mL 200 mg/L Rajendra GS KO (unclosed) (25KDa) From 1⋅106 to 4⋅106 et al., cells/mL 2015a 8 CHO-K1 CHOMACS CD GFP 24 μg/mL Linear PEI Yes Yes (After) No No 18⋅106 cells/mL ~60% of GFP Elshereef (25KDa) From 3⋅106 to 18⋅106 From 18⋅106 to positive et al., 2019 cells/mL 9⋅106 cells/mL CHO DG44 ProCHO5 GFP 0.1–1 μg/ Linear PEI Yes No No Yes (37→31 ◦ C) 5⋅106 cells/mL ~70% of GFP Rajendra 106 cells (25KDa) From 4⋅106 to 5⋅106 positive et al., cells/mL 2015c HEK293E Ex-Cell293 GFP 0.05–2.5 Linear PEI Yes Yes (After) From No No 20⋅106 cells/mL ~90% of GFP Rajendra μg/ 106 (25KDa) From 2⋅106 to 20⋅106 20⋅106 to 1⋅106 positive et al., cells cells/mL cells/mL 2015c CHO K1SV CHO-TNX Propietary 20 different 3.2 μg/mL Linear PEI Yes No Yes Yes (37→32 ◦ C) 4⋅106 cells/mL ~100 mg/L Schmitt GS KO mAbs (40KDa) From 2⋅106 to 4⋅106 et al., 2020 cells/mL ExpiCHO-S ExpiCHO 20 different 1 μg/mL ExpiFectamine – – Yes Yes (37→32 ◦ C) 6⋅106 cells/mL ~300 mg/L Schmitt mAbs CHO et al., 2020 CHO K1SV CHO-TNX Propietary 20 different 14.5 μg/ Linear PEI Cells were maintained Yes (Prior) Yes Yes (37→32 ◦ C) 20⋅106 cells/mL ~3000 mg/L Schmitt GS KO mAbs mL (40KDa) in a Perfusion From 50⋅106 et al., 2020 Bioreactor at 50⋅106 (Perfusion Biotechnology Advances 60 (2022) 108017 cells/mL Bioreactor) to 20⋅106 cells/mL HEK293SF- HyQSFM4TransFx293 GFP-Lentivirus – Linear PEI All experiments were No Yes No Ranging from ~1010VG/mL for Ansorge 3F6 (25KDa) performed in 4.4⋅106cells/mL 5⋅106 cells/mL and et al., 2011 perfusion operation to 8⋅106 cells/mL ~ 109VG/mL for mode 8⋅106 cells/mL J. Lavado-García et al. Biotechnology Advances 60 (2022) 108017 amount allow the achievement of successful transfection at HCD. MRs are carried out via low-speed centrifugation. If pelleted cell cultures Almost all HCD successful transfections at the range of 20⋅106 cells/ are resuspended in less volume, cell density is artificially increased. mL are achieved using this method (Table 1). Nevertheless, transfection Indeed, concentration is a frequent practice used in transient expression effciciency values that are achieved did not reach their corresponding at HCD (Barnard et al., 2015; Rajendra et al., 2015c, 2015b) (Fig. 2d). LCD reference (Blaha et al., 2015) suggesting that removing the factor of Rajendra et al. developed a proprietary transfection technology in CHO- the presence of an inhibitory component in the cell culture media is not GS KO cells based on concentration at 4⋅106 cells/mL from cultures at sufficient to completely overcome the CDE and that the cell-dependent 1⋅106 cells/mL. Feed and dimethylacetamide (DMA) addition up-scaled physiological factor remains unsolved. Also, the impplementation of production to >200 mg/L in the same process time that initially was this metholody at large scale poses a major drawback. yielding around 50 mg/L (Rajendra et al., 2015a). Interestingly, DMA seems to be present in many of HCD transfection procedures involving 7.2. Dilution followed by transfection cell concentration. DMA has been suggested to increase mRNA levels and when combined with mild hypothermia is capable of maximizing Most HCD transfection protocols rely on this simple yet effective one- titers up to 2-fold (Rajendra et al., 2015c). Cell concentration steps have step procedure to enhance productivity (Fig. 2c) (Jain et al., 2017; also been used to study how to overcome the CDE. Nielsen et al. looked Schmitt et al., 2020). In this sense, many protocols have been developed into the effects of cell concentration and presented a comparative study in an effort to further improve this method, such as the commercially of infected High Five insect cells at densities ranging from 0.5 to 5⋅106 available one reported by Jain et al. (Jain et al., 2017). In this study cells/mL (Table 2). They compared growing High Five cells and infect­ Expi293, ExpiCHO and FreeStyleMAX CHO systems were compared in ing at HCD to infecting at the same density of interest by means of cell an attempt to establish a high-productive TGE platform for HEK293 and concentration from a 0.5⋅106 cells/mL culture (Huynh et al., 2015). CHO cells. Dilutions with fresh culture medium were carried out before Intracellular and extracellular viral DNA were monitored as well as transfecting cultures at HCD. A remarkable 3 g/L IgG1 titer was ach­ β-Gal mRNA and protein expression. A clear decline in all measured ieved in the best condition, which came from an ExpiCHO-S culture at parameters was observed when growing at HCD prior to infection. 10⋅106 cells/mL, previously diluted at 6⋅106 cells/mL (Table 1). However, when concentrating early exponential-phase cells to HCDs Considering that high-producer CHO stable cell lines are able to attain β-Gal specific yield plateaued at 0.2 U/cell while an almost-zero value titers of approximately 15 g/L, this reported TGE output is a remarkable resulted in the controls without cell concentration. However, with this achievement (Kunert and Reinhart, 2016). Similarly, Barnard et al. approach, the two different effects contributing to the CDE could not be conducted a series of transfection experiments in which CHO K1SV KS individually studied as the condition employing metabolically young KO and ExpiCHO-S cell lines were transfected under diverse feeding cells also was influenced by the fact of carrying out a complete MR with regimes (Schmitt et al., 2020). Complex dilutions steps were performed fresh culture medium. Nevertheless, considering that MRs provide fresh at HCD prior to transfection at cell densities ranging from 4⋅106 up to nutrients, the authors declared that the CDE could not be fully explained 20⋅106 cells/mL. Cultures were successfully transfected at 20⋅106 cells/ by nutrient limitation nor waste-products accumulation since specific mL, which to date is the highest reported cell density to be transfected productivity was nowhere near restored. Aligned with these conclu­ without any dilution to LCD afterwards. Interestingly, CHO cells were sions, Dill et al. proved that 100% MR with conditioned media coming maintained in a perfusion bioreactor at cell densities up to 50⋅106 cells/ from grown cultures decreased viral specific productivity, yet produc­ mL before being diluted down to 20⋅106 cells/mL and transfected tion was not completely hindered (Dill et al., 2019). Likewise, in the (Fig. 2e). Noticeably, cells were two times washed with proprietary previously mentioned study by Rajendra and colleagues (Rajendra et al., media before dilution. This approach was tested with 20 different 2015c), they also used a cell concentration step to improve their trans­ monoclonal antibodies (mAbs), yielding in the best case >3 g/L in <7 fection efficiencies in CHO DG44 cells (Table 1). days. However, transfection efficiency was not evaluated here. In other Altogether, cell concentration seems to be a rapid and useful tool to attempts such as the one proposed by Stuible et al. a GFP reporter gene simulate HCD conditions but their impact on productivity is substan­ was used to track the transfection efficiency at HCD (Stuible et al., tially reduced if not followed by an additional consecutive dilution step. 2018). CHO cell cultures were directly transfected at 3⋅106 cells/mL and Actually, successful transfections and infections at HCD with just one 5⋅106 cells/mL. Additionally, cultures at 7⋅106 cells/mL were diluted to cell concentration step are found in cell density ranges from 3 to 6⋅106 3 and 5⋅106 cells/mL and transfection efficiency in terms of % GFP cells/mL, indicating the existence of more complex limitations conse­ positive population was examined between these four conditions. quence of increased cellular density (Huynh et al., 2013). Diluted cultures at 3 and 5⋅106 cells/mL coming from 7⋅106 cells/mL exhibited better performance than their counterparts since there was an 7.4. Infections at HCD in perfusion mode increase of 5–10% in GFP positive cell subpopulation, suggesting that the simple procedure of dilution may affect both factors comprising the Continuous manufacturing is becoming the spearhead operational CDE, reducing the presence of an inhibitory compound (Shen et al., strategy due to its capacity to overcome media-related limitations 2010) and contributing to a healthy cell physiological state as the enhancing productivity (Bielser et al., 2018). Perfusion allows for con­ dilution may act as a new cell passage. stant MR in turn diminishing by-product accumulation and permitting To the best of our knowledge, no infections are reported neither metabolic homogeneity among cell culture, thus reducing product diluting from HCD to LCD and infecting at HCD and then diluting to microheterogeneities (Walther et al., 2019). Still, process intensification LCD. does not entirely by-pass the CDE, as previously analyzed, being still a major constrain in both infection and transfection. Regarding infection, 7.3. Cell concentration followed by transient expression perfusion at HCD is extensively used for the manufacturing of viral vectors by infecting cell cultures at low MOIs, greatly extending culture As previously mentioned, MR steps are commonly used before production phase (Grein et al., 2017) (Table 2). Genzel et al. developed a transfection or infection as an easy one-step way to remove metabolic perfusion approach with three different membrane cut-offs for the by-products that may inhibit transfection and infection. For instance, production of H1N1 influenza virus in which glucose and glutamine Cervera et al. reported that repeated MRs coupled with retransfections were used as tracking metabolites to increase or reduce the perfusion greatly improved VLP production in HEK293 cell cultures (Cervera rate (Genzel et al., 2014). Levels were above 10–20 mM and 0.3 mM for et al., 2015b). Moreover, MRs are also exploited to normalize initial cell glucose and glutamine, respectively. Additionally, 50% of spent medium culture conditions at time of transient expression resulting in lower rates was removed before infection. Altogether, cell densities of 47.7⋅106 and of variability between biological replicates (Vis et al., 2020). Generally, 26.9⋅106 cells/mL were achieved in AGE1.CR and CAP cells respectively, 9 J. Lavado-García et al. Table 2 Examples of successful infections at high cell density (HCD) and the different methods and approaches employed to achieve it. MR: medium replacement. Cell Line Cell Recombinant MOI Virus Concentration Step / MR Dilution Addition Temperature Infected Cell Density Product titter Reference culture Product prior to infection Step after of feed shift after medium infection infection High Five Express- β-Gal 10 PFU/cell rAcMNPV No No No No Ranging from 1.2 U/cell for 0.5⋅106 Huynh et al., Five 0.5⋅106cells/mL to cells/mL and ~ 0 U/cell 2015 6⋅106 cells/mL for 6⋅106 cells/mL High Five Express- β-Gal 10 PFU/cell rAcMNPV Yes No No No Ranging from 1.2 U/cell for 0.5⋅106 Huynh et al., Five Ranging from 0.5⋅106 0.5⋅106cells/mL to cells/mL and 0.2 U/cell 2015 cells/mL to 6⋅106 cells/mL 5⋅106 cells/mL for 5⋅106 cells/mL Sf9 SF900II – 0.6 PFU/cell rAcMNPV No No No No Ranging from 172 pfu/cell for 1⋅106 Bernal et al., 1⋅106cells/mL to cells/mL and 2 pfu/cell 2009 3.5⋅106 cells/mL for 3.5⋅106 cells/mL Sf9 SF900II – 30 PFU/cell rAcMNPV No No No No Ranging from 486 pfu/cell for 1⋅106 Bernal et al., 1⋅106cells/mL to cells/mL and 2060 pfu/ 2009 3.5⋅106 cells/mL cell for 3.5⋅106 cells/mL HEK293SF- NSFM13 GFP – Adv5 All experiments were No Yes No Ranging from 1 VG/cell for 2.5⋅106 Henry et al., 3F6 performed in perfusion 2.5⋅106cells/mL to cells/mL and 0.19 for 2005 operation mode 6⋅106 cells/mL 6⋅106 cells/mL (Normalized results) HEK293SF- NSFM13 GFP MOI ranging 10–20 Adv5 All experiments were No Yes No Ranging from 15000 VG/cell for 5⋅105 Henry et al., 3F6 (Authors claim no performed in perfusion 0.5⋅106 cells/mL to cells/mL and 3800 for 2004 10 productivity operation mode 5.8⋅106 cells/mL 5.8⋅106 cells/mL difference) HEK293 Mix of GFP 10 PFU/cell Adv All experiments were No Yes No 11⋅106 cells/mL 5.7⋅1011 VG/mL Liu et al., DMEM (Serotype performed in perfusion 2009 and F-12 not operation mode specified) AGE1.CR CD-U3 – 0.001 PFU/cell Human All experiments were No Yes No Ranging from 25⋅106 1708 Virions/cell for Genzel et al., Influenza performed in perfusion cells/mL to 47⋅106 25⋅106 (Membrane Cut- 2014 H1N1 operation mode using cells/mL Off of 0.5 μm) and 1266 various membrane Cut- Virions/cell for 47⋅106 Offs cells/mL CAP CDM2 – 0.025 PFU/cell Human All experiments were No Yes No Ranging from 27⋅106 4086 Virions/cell for Genzel et al., Influenza performed in perfusion cells/mL to 33⋅106 27⋅106 cells/mL and 2014 H1N1 operation mode using cells/mL 1883 virions/cell for various membrane Cut- 33⋅106 cells/mL Offs MDCK Xeno- – 0.001 PFU/cell Human Comparison of Batch vs No Yes Yes (37→33 ◦ C) Ranging from 10476 Virions/cell for Wu et al., 7.6⋅106 cells/mL 7.6⋅106 cells/mL and Biotechnology Advances 60 (2022) 108017 CDM2 Influenza Perfusion 2021 H1N1 (Batch) to 40⋅106 4077 virions/cell for cells/mL (perfusion) 40⋅106 cells/mL AGE1.CR. CD-U3 – 0.05 PFU/Cell Human Comparison of Batch vs No Yes No Ranging from 100 Virions/cell for Vázquez- pIX Influenza Perfusion 2.7⋅106 cells/mL 2.7⋅106 cells/mL and 38 Ramírez H1N1 (Batch) to 83⋅106 virions/cell for 83⋅106 et al., 2018 cells/mL (perfusion) cells/mL J. Lavado-García et al. Biotechnology Advances 60 (2022) 108017 with no visible CDE considering that viral specific productivity was Funding comparable to that of batch cultures at LCD. Interestingly, the CDE was only noticeable at HCD when using the smaller cut-off membrane The research group is recognized as 2017 SGR 898 by Generalitat de (50KDa). Similarly, Vázquez-Ramírez et al. described the application of Catalunya. a glucose-based perfusion system to the production of MVA virus in CR. pIX cells (Vázquez-Ramírez et al., 2018). Initial perfusion runs reached CRediT authorship contribution statement 83⋅106 cells/mL, though specific productivity was 3-fold reduced compared to LCD. Bioprocess limitations were surpassed by developing Jesús Lavado-García: Conceptualization, Investigation, Writing – a scale-down model in shake flasks. Conventional perfusion was original draft, Writing – review & editing. Pol Pérez-Rubio: Concep­ mimicked with a combination of punctual feed additions and periodic tualization, Investigation, Writing – original draft, Writing – review & full MRs. Importantly, cell density down-scaled from 83⋅106 to 63⋅106 editing. Laura Cervera: Supervision. Francesc Gòdia: Supervision. cells/mL. Overall, specific viral yield was reestablished, consequently boosting volumetric productivity. Following the same trend, Wu et al. optimized perfusion parameters in MDCK cell cultures at densities above Declaration of Competing Interest 40⋅106 cells/mL to manufacture H1N1 influenza by means of infection (Wu et al., 2021). Decreasing cell specific perfusion rate (CSPR) from 60 The authors declare no conflict of interest. to 40 pL/cell/day and switching production temperature from 37 ◦ C to 33 ◦ C resulted in viral specific productivity restoration (Table 2). Other Data availability successful strategies rely on physical parameters such as the aggregation rate (Liu et al., 2009). Liu et al. demonstrated that HEK293 cell aggre­ No data was used for the research described in the article. gates with a mean diameter of 300 μm were effectively infected at densities exceeding 10⋅106 cells/mL while specific productivity was kept References constant, hence increasing adenovirus titers 10-fold, comparing to LCD. To the best of our knowledge, there are no reported cases in the Agarwal, A., Zudans, I., Weber, E.A., Olofsson, J., Orwar, O., Weber, S.G., 2007. Effect of literature involving successful transfections at HCD in perfusion-based cell size and shape on single-cell electroporation. Anal. Chem. 79, 3589–3596. https://doi.org/10.1021/AC062049E. processes. Ames, R.S., Kost, T.A., Condreay, J.P., 2007. BacMam technology and its application to drug discovery. Expert Opin. Drug Discovery 2, 1669–1681. https://doi.org/ 8. Current state and perspectives 10.1517/17460441.2.12.1669. Ansorge, S., Lanthier, S., Transfiguracion, J., Durocher, Y., Henry, O., Kamen, A., 2009. Development of a scalable process for high-yield lentiviral vector production by Overcoming the CDE is still one of the main challenges in the field of transient transfection of HEK293 suspension cultures. J. Gene Med. 11, 868–876. bioprocess development. Many efforts have been made towards https://doi.org/10.1002/jgm.1370. analyzing the metabolic causes leading to the reduction of cell specific Ansorge, S., Lanthier, S., Transfiguracion, J., Henry, O., Kamen, A., 2011. Monitoring lentiviral vector production kinetics using online permittivity measurements. productivity at HCD. All metabolic studies regarding transfection and Biochem. Eng. J. 54, 16–25. https://doi.org/10.1016/j.bej.2011.01.002. infection at HCD have determined that nutrient availability in the cell Arena, T.A., Chou, B., Harms, P.D., Wong, A.W., 2019. An anti-apoptotic HEK293 cell culture medium influences the final yield of the process but is not the line provides a robust and high titer platform for transient protein expression in bioreactors. MAbs 11, 977–986. https://doi.org/10.1080/19420862.2019.1598230. cause of the CDE. The presence of an inhibitory compound or by-product Backliwal, G., Hildinger, M., Hasija, V., Wurm, F.M., 2008. High-density transfection was still a plausible explanation as implementing sequential MR steps with HEK-293 cells allows doubling of transient titers and removes need for a priori allowed the achievement of better transfection and infection processes DNA complex formation with PEI. Biotechnol. Bioeng. 99, 721–727. https://doi.org/ 10.1002/BIT.21596. at HCD. However, all reported works of transfections and infections at Baldi, L., Muller, N., Picasso, S., Jacquet, R., Girard, P., Huy, P.T., Derow, E., Wurm, F. HCD suggest the existence of two independent factors influencing the M., 2005. Transient gene expression in suspension HEK-293 cells: application to CDE. Apart from this inhibitory compound that could be dealt with MR large-scale protein production. Biotechnol. Prog. 21, 148–153. https://doi.org/ 10.1021/BP049830X. steps, there is also a cell physiological element hitherto unknown. This is Bandaranayake, A.D., Almo, S.C., 2014. Recent advances in mammalian protein tackled by developing strategies involving both dilution and cell con­ production. FEBS Lett. 588, 253–260. https://doi.org/10.1016/j. centrations that somewhat alter cell physicochemical and physiological febslet.2013.11.035. properties subsequently allowing transfection or infections at HCD. Banik, G.G., Heath, C.A., 1994. An investigation of cell density effects on hybridoma metabolism in a homogeneous perfusion reactor. Bioprocess Biosyst. Eng. 6, Moreover, perfusion operations stand as the go-to alternative for HCD 229–237. cultivations. Their capacity of providing metabolic stability as well as Barnard, G.C., Hougland, M.D., Rajendra, Y., 2015. High-throughput mAb expression continuous fresh feeding seem to be the keys to delay the CDE. All in all, and purification platform based on transient CHO. Biotechnol. Prog. 31, 239–247. https://doi.org/10.1002/btpr.2012. current strategies do not attack the root of the problem but serve as a Bendix, M.B., Brint, E., Forde, P.F., 2021. Electroporation of suspension cell lines - a doorway to temporarily avoid it. The high variability of scenarios where proposed assay set for optimizations. Bioelectrochemistry 142. https://doi.org/ the CDE can be observed contributes to the difficulty of precise molec­ 10.1016/J.BIOELECHEM.2021.107891. Benjaminsen, R.V., Mattebjerg, M.A., Henriksen, J.R., Moghimi, S.M., Andresen, T.L., ular characterization. This should be tackled considering that there are 2013. The possible “proton sponge ” effect of polyethylenimine (PEI) does not two factors that contribute to the CDE and studying each of them indi­ include change in lysosomal pH. Mol. Ther. 21, 149–157. https://doi.org/10.1038/ vidually, thoroughly avoiding experimental designs that mix both mt.2012.185. Bernal, V., Carinhas, N., Yokomizo, A.Y., Carrondo, M.J., Alves, P.M., 2009. Cell density problematic points to successfully characterize each of them. The cur­ effect in the baculovirus-insect cells system: a quantitative analysis of energetic rent state-of-the-art molecular insight suggests that energy metabolism metabolism. Biotechnol. Bioeng. 104, 162–180. https://doi.org/10.1002/bit.22364. is not the main driving force for the CDE. The most recent studies point Bielser, J.M., Wolf, M., Souquet, J., Broly, H., Morbidelli, M., 2018. Perfusion mammalian cell culture for recombinant protein manufacturing – a critical review. to the direction of lipid biosynthesis homeostasis and a disruption in Biotechnol. Adv. https://doi.org/10.1016/j.biotechadv.2018.04.011. membrane renewal pathways. Blackstock, D.J., Goh, A., Shetty, S., Fabozzi, G., Yang, R., Ivleva, V.B., Schwartz, R., To date, the specific molecular causes of the CDE are still unclear. Horwitz, J., 2020. Comprehensive flow cytometry analysis of PEI-based transfections This is the main endeavor to be undertaken for the future of TGE bio­ for virus-like particle production. Research 2020, 1–10. https://doi.org/10.34133/ 2020/1387402. processes working at HCD. Blaha, J., Pachl, P., Novak, P., Vanek, O., 2015. Expression and purification of soluble and stable ectodomain of natural killer cell receptor LLT1 through high-density Author contributions transfection of suspension adapted HEK293S GnTI (− ) cells. Protein Expr. Purif. 109, 7–13. https://doi.org/10.1016/j.pep.2015.01.006. Bleckmann, M., Schü Rig, M., Endres, M., Samuels, A., Gebauer, D., Konisch, N., van den Both co-first authors have agreed to list their name first in their CV. Heuvelid, J., 2019. Identifying Parameters to Improve the Reproducibility of 11 J. Lavado-García et al. Biotechnology Advances 60 (2022) 108017 Transient Gene Expression in High Five cells. https://doi.org/10.1371/journal. Fuenmayor, J., Cervera, L., Gòdia, F., Kamen, A., 2019. Extended gene expression for Gag pone.0217878. VLP production achieved at bioreactor scale. J. Chem. Technol. Biotechnol. 94, Carinhas, N., Bernal, V., Yokomizo, A.Y., Carrondo, M.J.T., Oliveira, R., Alves, P.M., 302–308. https://doi.org/10.1002/jctb.5777. 2009. Baculovirus production for gene therapy: the role of cell density, multiplicity Garnier, L., Ravallec, M., Blanchard, P., Chaabihi, H., Bossy, J.P., Devauchelle, G., of infection and medium exchange. Appl. Microbiol. Biotechnol. 81, 1041–1049. Jestin, A., Cerutti, M., 1995. Incorporation of pseudorabies virus gD into human https://doi.org/10.1007/s00253-008-1727-4. immunodeficiency virus type 1 Gag particles produced in baculovirus-infected cells. Caron, A.L., Picanço-Castro, V., Ansorge, S., Covas, D.T., Kamen, A., Swiech, K., 2015. J. Virol. 69, 4060–4068. https://doi.org/10.1128/jvi.69.7.4060-4068.1995. Production of lentiviral vectors encoding recombinant factor VIII expression in Geisse, S., 2009. Reflections on more than 10 years of TGE approaches. Protein Expr. serum-free suspension cultures. Braz. Arch. Biol. Technol. 58, 923–928. https://doi. Purif. 64, 99–107. https://doi.org/10.1016/j.pep.2008.10.017. org/10.1590/S1516-89132015060367. Geisse, S., Fux, C., 2009. Recombinant protein production by transient gene transfer into Carpentier, E., Paris, S., Kamen, A.A., Durocher, Y., 2007. Limiting factors governing mammalian cells. Methods Enzymol. 463, 223–238. https://doi.org/10.1016/s0076- protein expression following polyethylenimine-mediated gene transfer in HEK293- 6879 (09)63015-9. EBNA1 cells. J. Biotechnol. 128, 268–280. https://doi.org/10.1016/j. Genzel, Y., Vogel, T., Buck, J., Behrendt, I., Ramirez, D.V., Schiedner, G., Jordan, I., jbiotec.2006.10.014. Reichl, U., 2014. High cell density cultivations by alternating tangential flow (ATF) Cervera, L., Gutiérrez-Granados, S., Martínez, M., Blanco, J., Gòdia, F., Segura, M.M., perfusion for influenza A virus production using suspension cells. Vaccine 32, 2013. Generation of HIV-1 Gag VLPs by transient transfection of HEK 293 suspension 2770–2781. https://doi.org/10.1016/j.vaccine.2014.02.016. cell cultures using an optimized animal-derived component free medium. Ghorbaniaghdam, A., Henry, O., Jolicoeur, M., 2013. A kinetic-metabolic model based J. Biotechnol. 166, 152–165. https://doi.org/10.1016/j.jbiotec.2013.05.001. on cell energetic state: study of CHO cell behavior under Na-butyrate stimulation. Cervera, L., Fuenmayor, J., González-Domínguez, I., Gutiérrez-Granados, S., Segura, M. Bioprocess Biosyst. Eng. 36, 469–487. https://doi.org/10.1007/S00449-012-0804- M., Gòdia, F., 2015a. Selection and optimization of transfection enhancer additives 3/FIGURES/10. for increased virus-like particle production in HEK293 suspension cell cultures. Appl. González-Domínguez, I., Grimaldi, N., Cervera, L., Ventosa, N., Gòdia, F., 2019. Impact Microbiol. Biotechnol. 99, 9935–9949. https://doi.org/10.1007/s00253-015-6842- of physicochemical properties of DNA/PEI complexes on transient transfection of 4. mammalian cells. New Biotechnol. 49, 88–97. https://doi.org/10.1016/j. Cervera, L., Gutiérrez-Granados, S., Berrow, N.S., Segura, M.M., Gòdia, F., 2015b. nbt.2018.09.005. Extended gene expression by medium exchange and repeated transient transfection Gränicher, G., Tapia, F., Behrendt, I., Jordan, I., Genzel, Y., Reichl, U., 2021. Production for recombinant protein production enhancement. Biotechnol. Bioeng. 112, of modified vaccinia ankara virus by intensified cell cultures: a comparison of 934–946. https://doi.org/10.1002/bit.25503. platform technologies for viral vector production. Biotechnol. J. 16 https://doi.org/ Cervera, L., Gonzalez-Dominguez, I., Segura, M.M., Godia, F., 2017. Intracellular 10.1002/biot.202000024. characterization of Gag VLP production by transient transfection of HEK 293 cells. Grein, T.A., Weidner, T., Czermak, P., 2017. Concepts for the production of viruses and Biotechnol. Bioeng. 114, 2507–2517. https://doi.org/10.1002/bit.26367. viral vectors in cell cultures. In: New Insights Into Cell Culture Technology.. https:// Chen, F., Kou, T., Fan, L., Zhou, Y., Ye, Z., Zhao, L., Tan, W.S., 2011. The combined effect doi.org/10.5772/66903. of sodium butyrate and low culture temperature on the production, sialylation, and Grieger, J.C., Soltys, S.M., Samulski, R.J., 2016. Production of recombinant adeno- biological activity of an antibody produced in CHO cells. Biotechnol. Bioprocess Eng. associated virus vectors using suspension HEK293 cells and continuous harvest of 1157–1165. https://doi.org/10.1007/S12257-011-0069-8, 2011 16:6 16. vector from the culture media for GMP FIX and FLT1 clinical vector. Mol. Ther. 24, Clincke, M.F., Mölleryd, C., Samani, P.K., Lindskog, E., Fäldt, E., Walsh, K., Chotteau, V., 287–297. https://doi.org/10.1038/mt.2015.187. 2013a. Very high density of Chinese hamster ovary cells in perfusion by alternating Gutiérrez-Granados, S., Cervera, L., Segura, M.E.L., Wölfel, J., Gòdia, F., 2016. tangential flow or tangential flow filtration in WAVE BioreactorTM-part II: Optimized production of HIV-1 virus-like particles by transient transfection in CAP-T applications for antibody production and cryopreservation. Biotechnol. Prog. 29, cells. Appl. Microbiol. Biotechnol. 100, 3935–3947. https://doi.org/10.1007/ 768–777. https://doi.org/10.1002/btpr.1703. s00253-015-7213-x. Clincke, M.F., Molleryd, C., Zhang, Y., Lindskog, E., Walsh, K., Chotteau, V., 2013b. Very Gutiérrez-Granados, S., Cervera, L., Kamen, A.A., Gòdia, F., 2018. Advancements in high density of CHO cells in perfusion by ATF or TFF in WAVE bioreactor. Part I. mammalian cell transient gene expression (TGE) technology for accelerated effect of the cell density on the process. Biotechnol. Prog. 29, 754–767. https://doi. production of biologics. Crit. Rev. Biotechnol. 38, 918–940. https://doi.org/ org/10.1002/btpr.1704. 10.1080/07388551.2017.1419459. Cohen, R.N., van der Aa, M.A.E.M., Macaraeg, N., Lee, A.P., Szoka, F.C., 2009. Haldankar, R., Li, D., Saremi, Z., Baikalov, C., Deshpande, R., 2006. Serum-free Quantification of plasmid DNA copies in the nucleus after lipoplex and polyplex suspension large-scale transient transfection of CHO cells in WAVE bioreactors. Mol. transfection. J. Control. Release 135, 166–174. https://doi.org/10.1016/j. Biotechnol. 34, 191–199. https://doi.org/10.1385/MB:34:2:191. jconrel.2008.12.016. Henry, O., Dormond, E., Perrier, M., Kamen, A., 2004. Insights into adenoviral vector Craig Seamans, T., Hu, W.-S., 1990. Kinetics of growth and antibody production by a production kinetics in acoustic filter-based perfusion cultures. Biotechnol. Bioeng. hybridoma cell line in a perfusion culture. J. Ferment. Bioeng. 70, 241–245. https:// 86, 765–774. https://doi.org/10.1002/BIT.20074. doi.org/10.1016/0922-338X (90)90056-3. Henry, O., Perrier, M., Kamen, A., 2005. Metabolic flux analysis of HEK-293 cells in Damiani, R., Almeida, B.E., Oliveira, J.E., Bartolini, P., Ribela, M.T.C.P., 2013. perfusion cultures for the production of adenoviral vectors. Metab. Eng. 7, 467–476. Enhancement of human thyrotropin synthesis by sodium butyrate addition to serum- https://doi.org/10.1016/J.YMBEN.2005.08.002. free cho cell culture. Appl. Biochem. Biotechnol. 171, 1658–1672. https://doi.org/ Huang, Y.M., Hu, W.W., Rustandi, E., Chang, K., Yusuf-Makagiansar, H., Ryll, T., 2010. 10.1007/S12010-013-0467-9/TABLES/7. Maximizing productivity of CHO cell-based fed-batch culture using chemically Derouazi, M., Girard, P., Van Tilborgh, F., Iglesias, K., Muller, N., Bertschinger, M., defined media conditions and typical manufacturing equipment. Biotechnol. Prog. Wurm, F.M., 2004. Serum-free large-scale transient transfection of CHO cells. 26, 1400–1410. https://doi.org/10.1002/btpr.436. Biotechnol. Bioeng. 87, 537–545. https://doi.org/10.1002/bit.20161. Hur, J., Park, I., Lim, K.M., Doh, J., Cho, S.G., Chung, A.J., 2020. Microfluidic cell Dill, V., Ehret, J., Zimmer, A., Beer, M., Eschbaumer, M., 2019. Cell density effects in stretching for highly effective gene delivery into hard-to-transfect primary cells. ACS different cell culture media and their impact on the propagation of foot-and-mouth Nano 14, 15094–15106. https://doi.org/10.1021/ACSNANO.0C05169. disease virus. Viruses 11. https://doi.org/10.3390/v11060511. Huynh, H.T., Tran, T.T.B., Chan, L.C.L., Nielsen, L.K., Reid, S., 2013. Decline in Durocher, Y., Perret, S., Kamen, A., 2002. High-level and high-throughput recombinant baculovirus-expressed recombinant protein production with increasing cell density protein production by transient transfection of suspension-growing human 293- is strongly correlated to impairment of virus replication and mRNA expression. Appl. EBNA1 cells. Nucleic Acids Res. 30, e9. https://doi.org/10.1093/NAR/30.2.E9. Microbiol. Biotechnol. 97, 5245–5257. https://doi.org/10.1007/s00253-013-4835- Elias, C., Zeiser, A., Bedard, C., Kamen, A., 2000. Enhanced growth of Sf-9 cells to a 8. maximum density of 5.2 x 10 (7) cells per mL and production of beta-galactosidase at Huynh, H.T., Tran, T.T., Chan, L.C., Nielsen, L.K., Reid, S., 2015. Effect of the peak cell high cell density by fed batch culture. Biotechnol. Bioeng. 68 (4), 381–388. https:// density of recombinant AcMNPV-infected Hi5 cells on baculovirus yields. Appl. doi.org/10.1002/(sici)1097-0290(20000520)68:4<381::aid-bit3>3.0.co;2-d. Microbiol. Biotechnol. 99, 1687–1700. https://doi.org/10.1007/s00253-014-6260- Elouahabi, A., Ruysschaert, J.-M., 2004. Formation and Intracellular Trafficking of z. Lipoplexes and Polyplexes. https://doi.org/10.1016/j.ymthe.2004.12.006. Ikonomou, L., Schneider, Y.J., Agathos, S.N., 2003. Insect cell culture for industrial Elshereef, A.A., Jochums, A., Lavrentieva, A., Stuckenberg, L., Scheper, T., Solle, D., production of recombinant proteins. Appl. Microbiol. Biotechnol. 62, 1–20. https:// 2019. High cell density transient transfection of CHO cells for TGF-β1 expression. doi.org/10.1007/S00253-003-1223-9. Eng. Life Sci. 19, 730–740. https://doi.org/10.1002/elsc.201800174. Ikonomou, L., Bastin, G., Schneider, Y.J., Agathos, S.N., 2004. Effect of partial medium Fan, Y., Ley, D., Andersen, M.R., 2018. Fed-batch CHO cell culture for lab-scale antibody replacement on cell growth and protein production for the high-fiveTM insect cell production. Biotechnol. Bioeng. 1674, 147–161. https://doi.org/10.1007/978-1- line. Cytotechnology 44, 67–76. https://doi.org/10.1023/B: 4939-7312-5. CYTO.0000043413.53044.fa. Felberbaum, R.S., 2015. The baculovirus expression vector system: a commercial Jäger, V., Büssow, K., Schirrmann, T., 2015. Transient Recombinant Protein Expression manufacturing platform for viral vaccines and gene therapy vectors. Biotechnol. J. in Mammalian Cells, pp. 27–64. https://doi.org/10.1007/978-3-319-10320-4_2. 10, 702–714. https://doi.org/10.1002/biot.201400438. Jain, N.K., Barkowski-Clark, S., Altman, R., Johnson, K., Sun, F., Zmuda, J., Liu, C.Y., Frazzati-Gallina, N.M., Paoli, R.L., Mourão-Fuches, R.M., Jorge, S.A.C., Pereira, C.A., Kita, A., Schulz, R., Neill, A., Ballinger, R., Patel, R., Liu, J., Mpanda, A., Huta, B., 2001. Higher production of rabies virus in serum-free medium cell cultures on Chiou, H., Voegtli, W., Panavas, T., 2017. A high density CHO-S transient microcarriers. J. Biotechnol. 92, 67–72. https://doi.org/10.1016/S0168-1656 (01) transfection system: comparison of ExpiCHO and Expi293. Protein Expr. Purif. 134, 00362-5. 38–46. https://doi.org/10.1016/j.pep.2017.03.018. Freeman, E.C., Weiland, L.M., Meng, W.S., 2013. Modeling the proton sponge Jeon, M.K., Gyun, M.L., 2007. Correlation between enhancing effect of sodium butyrate hypothesis: examining proton sponge effectiveness for enhancing intracellular gene on specific productivity and mRNA transcription level in recombinant Chinese delivery through multiscale modeling. J. Biomater. Sci. Polym. Ed. 24, 398–416. hamster ovary cells producing antibody. J. Microbiol. Biotechnol. 17, 1036–1040. https://doi.org/10.1080/09205063.2012.690282. 12 J. Lavado-García et al. Biotechnology Advances 60 (2022) 108017 Joshi, P.R.H., Cervera, L., Ahmed, I., Kondratov, O., Zolotukhin, S., Schrag, J., Chahal, P. Nadeau, I., Kamen, A., 2003. Production of adenovirus vector for gene therapy. S., Kamen, A.A., 2019. Achieving high-yield production of functional AAV5 gene Biotechnol. Adv. 20, 475–489. https://doi.org/10.1016/S0734-9750 (02)00030-7. delivery vectors via fedbatch in an insect cell-one Baculovirus system. Mol. Ther. 13, Nadeau, I., Jacob, D., Perrier, M., Kamen, A., 2000. 293SF metabolic flux analysis during 279–289. https://doi.org/10.1016/j.omtm.2019.02.003. cell growth and infection with an adenoviral vector. Biotechnol. Prog. 16, 872–884. Joshi, P.R.H., Venereo-Sanchez, A., Chahal, P.S., Kamen, A.A., 2021. Advancements in https://doi.org/10.1021/BP000098L. molecular design and bioprocessing of recombinant adeno-associated virus gene Palomares, L.A., Estrada-Moncada, S., R., O.T., 2005. Principles and applications of the delivery vectors using the insect-cell baculovirus expression platform. Biotechnol. J. insect cell-Baculovirus expression vector system. In: Cell Culture Technology for 16, 2000021. https://doi.org/10.1002/BIOT.202000021. Pharmaceutical and Cell-Based Therapies, pp. 645–710. https://doi.org/10.1201/ Kadlecova, Z., Nallet, S., Hacker, D.L., Baldi, L., Klok, H.A., Wurm, F.M., 2012. Poly 9780849351068-21. (ethyleneimine)-mediated large-scale transient gene expression: influence of Parham, J.H., Iannone, M.A., Overton, L.K., Hutchins, J.T., 1998. Optimization of molecular weight, polydispersity and N-propionyl groups. Macromol. Biosci. 12, Transient Gene Expression in Mammalian Cells and Potential for Scale-up Using 628–636. https://doi.org/10.1002/MABI.201100404. Flow Electroporation. Springer. Kiela, P.R., Kuscuoglu, N., Midura, A.J., Midura-Kiela, M.T., Larmonier, C.B., Lipko, M., Park, J.Y., Lim, B.P., Lee, K., Kim, Y.G., Jo, E.C., 2006. Scalable production of adeno- Ghishan, F.K., 2007. Molecular mechanism of rat NHE3 gene promoter regulation by associated virus type 2 vectors via suspension transfection. Biotechnol. Bioeng. 94, sodium butyrate. Am. J. Phys. Cell Physiol. 293, 64–74. https://doi.org/10.1152/ 416–430. https://doi.org/10.1002/BIT.20776. ajpcell.00277.2006.-Sodium. Park, E., Cho, H.B., Takimoto, K., 2015. Effective gene delivery into adipose-derived stem Kraus, B., von Fircks, S., Feigl, S., Koch, S.M., Fleischanderl, D., Terler, K., Dersch- cells: transfection of cells in suspension with the use of a nuclear localization signal Pourmojib, M., Konetschny, C., Grillberger, L., Reiter, M., 2011. Avian cell line - peptide–conjugated polyethylenimine. Cytotherapy 17, 536–542. https://doi.org/ technology for large scale vaccine production. BMC Proc. 5, 8–10. https://doi.org/ 10.1016/J.JCYT.2014.11.008. 10.1186/1753-6561-5-s8-p52. Patrone, M., Carinhas, N., Sousa, M.Q., Peixoto, C., Ciferri, C., Carfì, A., Alves, P.M., Kunert, R., Reinhart, D., 2016. Advances in recombinant antibody manufacturing. Appl. 2014. Enhanced expression of full-length human cytomegalovirus fusion protein in Microbiol. Biotechnol. 100, 3451–3461. https://doi.org/10.1007/s00253-016-7388- non-swelling baculovirus-infected cells with a minimal fed-batch strategy. PLoS One 9. 9, 1–13. https://doi.org/10.1371/journal.pone.0090753. Kurasawa, J.H., Park, A., Sowers, C.R., Halpin, R.A., Tovchigrechko, A., Dobson, C.L., Petiot, E., Cuperlovic-Culf, M., Shen, C.F., Kamen, A., 2015. Influence of HEK293 Schmelzer, A.E., Gao, C., Wilson, S.D., Ikeda, Y., 2020. Chemically defined, high- metabolism on the production of viral vectors and vaccine. Vaccine 33, 5974–5981. density insect cell-based expression system for scalable AAV vector production. Mol. https://doi.org/10.1016/j.vaccine.2015.05.097. Ther. 19, 330–340. https://doi.org/10.1016/j.omtm.2020.09.018. Petiot, E., Proust, A., Traversier, A., Durous, L., Dappozze, F., Gras, M., Guillard, C., Kuwae, S., Miyakawa, I., Doi, T., 2018. Development of a chemically defined platform Balloul, J.M., Rosa-Calatrava, M., 2018. Influenza viruses production: evaluation of a fed-batch culture media for monoclonal antibody-producing CHO cell lines with novel avian cell line DuckCelt®-T17. Vaccine 36, 3101–3111. https://doi.org/ optimized choline content. Cytotechnology 70, 939–948. https://doi.org/10.1007/ 10.1016/j.vaccine.2017.03.102. s10616-017-0185-1. Powers, A.D., Piras, B.A., Clark, R.K., Lockey, T.D., Meagher, M.M., 2016. Development Lai, T., Yang, Y., Ng, S.K., 2013. Advances in mammalian cell line development and optimization of AAV hFIX particles by transient transfection in an iCELLis (®) technologies for recombinant protein production. Pharmaceuticals. https://doi.org/ fixed-bed bioreactor. Hum. Gene Ther. Meth. 27, 112–121. https://doi.org/ 10.3390/ph6050579. 10.1089/HGTB.2016.021. Lavado-García, J., Cervera, L., Gòdia, F., 2020a. An alternative perfusion approach for Puente-Massaguer, E., Lecina, M., Gòdia, F., 2020. Integrating nanoparticle the intensification of virus-like particle production in HEK293 cultures. Front. quantification and statistical design of experiments for efficient HIV-1 virus-like Bioeng. Biotechnol. 8, 617. https://doi.org/10.3389/FBIOE.2020.00617. particle production in high five cells. Appl. Microbiol. Biotechnol. 104, 1569–1582. Lavado-García, J., Jorge, I., Cervera, L., Vázquez, J., Gòdia, F., 2020b. Multiplexed https://doi.org/10.1007/s00253-019-10319-x. quantitative proteomic analysis of HEK293 provides insights into molecular changes Rajendra, Y., 2018. PEI-mediated transient gene expression in CHO cells. Methods Mol. associated with the cell density effect, transient transfection, and virus-like particle Biol. 1850, 33–42. https://doi.org/10.1007/978-1-4939-8730-6_3. production. J. Proteome Res. 19, 1085–1099. https://doi.org/10.1021/acs. Rajendra, Y., Kiseljak, D., Baldi, L., Hacker, D.L., Wurm, F.M., 2011. A simple high- jproteome.9b00601. yielding process for transient gene expression in CHO cells. J. Biotechnol. 153, Le Ru, A., Jacob, D., Transfiguracion, J., Ansorge, S., Henry, O., Kamen, A.A., 2010. 22–26. https://doi.org/10.1016/J.JBIOTEC.2011.03.001. Scalable production of influenza virus in HEK-293 cells for efficient vaccine Rajendra, Y., Hougland, M.D., Alam, R., Morehead, T.A., Barnard, G.C., 2015a. A high manufacturing. Vaccine 28, 3661–3671. https://doi.org/10.1016/j. cell density transient transfection system for therapeutic protein expression based on vaccine.2010.03.029. a CHO GS-knockout cell line: process development and product quality assessment. Li, W. Feng, Fan, Z. Lin, Wang, X. Yin, Lin, Y., Wang, T.Y., 2022. Combination of sodium Biotechnol. Bioeng. 112, 977–986. https://doi.org/10.1002/BIT.25514. butyrate and decitabine promotes transgene expression in CHO cells via apoptosis Rajendra, Y., Hougland, M.D., Schmitt, M.G., Barnard, G.C., 2015b. Transcriptional and inhibition. New Biotechnol. 69, 8–17. https://doi.org/10.1016/J.NBT.2022.02.004. post-transcriptional targeting for enhanced transient gene expression in CHO cells. Liu, C., Dalby, B., Chen, W., Kilzer, J.M., Chiou, H.C., 2008. Transient transfection factors Biotechnol. Lett. 37, 2379–2386. https://doi.org/10.1007/s10529-015-1938-6. for high-level recombinant protein production in suspension cultured mammalian Rajendra, Y., Kiseljak, D., Baldi, L., Wurm, F.M., Hacker, D.L., 2015c. Transcriptional and cells. Mol. Biotechnol. 39, 141–153. https://doi.org/10.1007/S12033-008-9051-X. post-transcriptional limitations of high-yielding, PEI-mediated transient transfection Liu, H., Liu, X.M., Li, S.C., Wu, B.C., Ye, L.L., Wang, Q.W., Chen, Z.L., 2009. A high-yield with CHO and HEK-293E cells. American Institute of Chemical Engineers Biotechnol. and scaleable adenovirus vector production process based on high density perfusion Prog 31, 541–549. https://doi.org/10.1002/btpr.2064. culture of HEK 293 cells as suspended aggregates. J. Biosci. Bioeng. 107, 524–529. Ramaswami, R., Bayer, R., Galea, S., 2018. Precision medicine from a public health https://doi.org/10.1016/j.jbiosc.2009.01.004. perspective. Annu. Rev. Public Health 39, 153–168. https://doi.org/10.1146/ Liu, Y.K., Yang, C.J., Liu, C.L., Shen, C.R., Shiau, L.D., 2010. Using a fed-batch culture ANNUREV-PUBLHEALTH-040617-014158. strategy to enhance rAAV production in the baculovirus/insect cell system. J. Biosci. Rausch, M., Pörtner, R., Knäblein, J., 2013. Increase of protein yield in high five cells in a Bioeng. 110, 187–193. https://doi.org/10.1016/j.jbiosc.2010.02.004. single-use perfusion bioreactor by medium replacement. Chemie-Ingenieur-Technik Lohr, V., Hädicke, O., Genzel, Y., Jordan, I., Büntemeyer, H., Klamt, S., Reichl, U., 2014. 85, 111–117. https://doi.org/10.1002/cite.201200121. The avian cell line AGE1.CR.pIX characterized by metabolic flux analysis. BMC Robert, M.A., Chahal, P.S., Audy, A., Kamen, A., Gilbert, R., Gaillet, B., 2017. Biotechnol. 14, 1–14. https://doi.org/10.1186/1472-6750-14-72. Manufacturing of recombinant adeno-associated viruses using mammalian Lorenzo, E., Méndez, L., Rodríguez, E., Gonzalez, N., Cabrera, G., Pérez, C., Pimentel, R., expression platforms. Biotechnol. J. 12 https://doi.org/10.1002/biot.201600193. Sordo, Y., Molto, M.P., Sardina, T., Rodríguez-Mallon, A., Estrada, M.P., 2019. Schmitt, M.G., White, R.N., Barnard, G.C., 2020. Development of a high cell density Plasticity of the HEK-293 cells, related to the culture media, as platform to produce a transient CHO platform yielding mAb titers greater than 2 g/L in only 7 days. subunit vaccine against classical swine fever virus. AMB Express 9. https://doi.org/ Biotechnol. Prog. 36 https://doi.org/10.1002/btpr.3047. 10.1186/s13568-019-0864-8. Schwarz, H., Zhang, Y., Zhan, C., Malm, M., Field, R., Turner, R., Sellick, C., Varley, P., Maranga, L., Auniņš, J.G., Zhou, W., 2005. Characterization of changes in PER.C6TM Rockberg, J., Chotteau, V., 2020. Small-scale bioreactor supports high density cellular metabolism during growth and propagation of a replication-deficient HEK293 cell perfusion culture for the production of recombinant erythropoietin. adenovirus vector. Biotechnol. Bioeng. 90, 645–655. https://doi.org/10.1002/ J. Biotechnol. 309, 44–52. https://doi.org/10.1016/j.jbiotec.2019.12.017. BIT.20455. Shang, M., Kwon, T., Hamel, J.F.P., Lim, C.T., Khoo, B.L., Han, J., 2021. Investigating the Martínez-Monge, I., Comas, P., Triquell, J., Lecina, M., Casablancas, A., Cairó, J.J., 2018. influence of physiologically relevant hydrostatic pressure on CHO cell batch culture. A new strategy for fed-batch process control of HEK293 cell cultures based on alkali Sci. Rep. 11, 1–9. https://doi.org/10.1038/s41598-020-80576-8. buffer addition monitoring: comparison with O.U.R. dynamic method. Appl. Shen, C.F., Voyer, R., Tom, R., Kamen, A., 2010. Reassessing culture media and critical Microbiol. Biotechnol. 102, 10469–10483. https://doi.org/10.1007/s00253-018- metabolites that affect adenovirus production. Biotechnol. Prog. 26, 200–207. 9388-4. https://doi.org/10.1002/btpr.286. Meghrous, J., Aucoin, M.G., Jacob, D., Chahal, P.S., Arcand, N., Kamen, A.A., 2005. Shen, X., Hacker, D.L., Baldi, L., Wurm, F.M., 2013. Virus-free transient protein Production of recombinant adeno-associated viral vectors using a baculovirus/insect production in Sf9 cells. J. Biotechnol. 171, 61–70. https://doi.org/10.1016/J. cell suspension culture system: from shake flasks to a 20-L bioreactor. Biotechnol. JBIOTEC.2013.11.018. Prog. 21, 154–160. https://doi.org/10.1021/BP049802E. Silva, C.A.T., Kamen, A.A., Henry, O., 2021. Recent advances and current challenges in Merten, O.W., Charrier, S., Laroudie, N., Fauchille, S., Dugue, C., Jenny, C., Audit, M., process intensification of cell culture-based influenza virus vaccine manufacturing. Zanta-Boussif, M.A., Chautard, H., Radrizzani, M., Vallanti, G., Naldini, L., Noguiez- Can. J. Chem. Eng. 99, 2525–2535. https://doi.org/10.1002/CJCE.24197. Hellin, P., Galy, A., 2011. Large-scale manufacture and characterization of a Steger, K., Brady, J., Wang, W., Duskin, M., Donato, K., Peshwa, M., 2015. CHO-S lentiviral vector produced for clinical ex vivo gene therapy application. Hum. Gene antibody titers >1 gram/liter using flow electroporation-mediated transient gene Ther. 22, 343–356. https://doi.org/10.1089/hum.2010.060. expression followed by rapid migration to high-yield stable cell lines. J. Biomol. Screen. 20, 545–551. https://doi.org/10.1177/1087057114563494. 13 J. Lavado-García et al. Biotechnology Advances 60 (2022) 108017 Stepper, L., Filser, F.A., Fischer, S., Schaub, J., Gorr, I., Voges, R., 2020. Pre-stage Wiegmann, V., Giaka, M., Bernal Martinez, C., Baganz, F., 2019. Towards the perfusion and ultra-high seeding cell density in CHO fed-batch culture: a case study development of automated fed-batch cell culture processes at microscale. for process intensification guided by systems biotechnology. Bioprocess Biosyst. Eng. Biotechniques 67 (5), 238–241. https://doi.org/10.2144/btn-2019-0063. 43, 1431–1443. https://doi.org/10.1007/s00449-020-02337-1. Wohlpart, D., Kirwan, D., Gainer, J., 1990. Effects of cell density and glucose and Stuible, M., Burlacu, A., Perret, S., Brochu, D., Paul-Roc, B., Baardsnes, J., Loignon, M., glutamine levels on the respiration rates of hybridoma cells. Biotechnol. Bioeng. 36, Grazzini, E., Durocher, Y., 2018. Optimization of a high-cell-density 630–635. https://doi.org/10.1002/BIT.260360611. polyethylenimine transfection method for rapid protein production in CHO-EBNA1 Wu, Y., Jia, H., Lai, H., Liu, X., Tan, W.S., 2020. Highly efficient production of an cells. J. Biotechnol. 281, 39–47. https://doi.org/10.1016/j.jbiotec.2018.06.307. influenza H9N2 vaccine using MDCK suspension cells. Bioresour. Bioprocess. 7 Sun, X., Hui, C.H., Peen, E.G., Yap, M.G.S., 2008. High-density transient gene expression https://doi.org/10.1186/s40643-020-00352-4. in suspension-adapted 293 EBNA1 cells. Biotechnol. Bioeng. 99, 108–116. https:// Wu, Y., Bissinger, T., Genzel, Y., Liu, X., Reichl, U., Tan, W.S., 2021. High cell density doi.org/10.1002/BIT.21537. perfusion process for high yield of influenza a virus production using MDCK Tan, E., Chin, C.S.H., Lim, Z.F.S., Ng, S.K., 2021. HEK293 cell line as a platform to suspension cells. Appl. Microbiol. Biotechnol. 105, 1421–1434. https://doi.org/ produce recombinant proteins and viral vectors. Front. Bioeng. Biotechnol. 9 10.1007/s00253-020-11050-8. https://doi.org/10.3389/FBIOE.2021.796991. Xu, J., Rehmann, M.S., Xu, M., Zheng, S., Hill, C., He, Q., Borys, M.C., Li, Z.J., 2020. Tripathi, N.K., Shrivastava, A., 2019. Recent developments in bioprocessing of Development of an intensified fed-batch production platform with doubled titers recombinant proteins: expression hosts and process development. Front. Bioeng. using N-1 perfusion seed for cell culture manufacturing. Bioresour. Bioprocess. 7 Biotechnol. 7 https://doi.org/10.3389/fbioe.2019.00420. https://doi.org/10.1186/s40643-020-00304-y. Tsai, C.H., Wei, S.C., Lo, H.R., Chao, Y.C., 2020. Baculovirus as versatile vectors for Yamano, S., Dai, J., Moursi, A.M., 2010. Comparison of transfection efficiency of protein display and biotechnological applications. Curr. Issues Mol. Biol. 34, nonviral gene transfer reagents. Mol. Biotechnol. 46, 287–300. https://doi.org/ 231–255. https://doi.org/10.21775/CIMB.034.231. 10.1007/S12033-010-9302-5. Turan, S., Galla, M., Ernst, E., Qiao, J., Voelkel, C., Schiedlmeier, B., Zehe, C., Bode, J., Ye, J., Kober, V., Tellers, M., Naji, Z., Salmon, P., Markusen, J.F., 2009. High-level 2011. Recombinase-mediated cassette exchange (RMCE): traditional concepts and protein expression in scalable CHO transient transfection. Biotechnol. Bioeng. 103, current challenges. J. Mol. Biol. https://doi.org/10.1016/j.jmb.2011.01.004. 542–551. https://doi.org/10.1002/BIT.22265. Tuvesson, O., Uhe, C., Rozkov, A., Lullau, E., 2008. Development of a generic transient Yee, C.M., Zak, A.J., Hill, B.D., Wen, F., 2018. The coming age of insect cells for transfection process at 100 L scale. Cytotechnology 56, 123–136. https://doi.org/ manufacturing and development of protein therapeutics. Ind. Eng. Chem. Res. 57, 10.1007/s10616-008-9135-2. 10061–10070. https://doi.org/10.1021/acs.iecr.8b00985. Urabe, M., Ding, C., Kotin, R.M., 2002. Insect cells as a factory to produce adeno- Yongky, A., Xu, J., Tian, J., Oliveira, C., Zhao, J., McFarland, K., Borys, M.C., Li, Z.J., associated virus type 2 vectors. Hum. Gene Ther. 13, 1935–1943. https://doi.org/ 2019. Process intensification in fed-batch production bioreactors using non- 10.1089/10430340260355347. perfusion seed cultures. MAbs 11, 1502–1514. https://doi.org/10.1080/ Vallée, C., Durocher, Y., Henry, O., 2014. Exploiting the metabolism of PYC expressing 19420862.2019.1652075. HEK293 cells in fed-batch cultures. J. Biotechnol. 169, 63–70. https://doi.org/ Yuk, I.H.Y., Olsen, M.M., Geyer, S., Forestell, S.P., 2004. Perfusion cultures of human 10.1016/j.jbiotec.2013.11.002. tumor cells: a scalable production platform for oncolytic adenoviral vectors. Vázquez-Ramírez, D., Genzel, Y., Jordan, I., Sandig, V., Reichl, U., 2018. High-cell- Biotechnol. Bioeng. 86, 637–642. https://doi.org/10.1002/BIT.20158. density cultivations to increase MVA virus production. Vaccine 36, 3124–3133. Zeng, A.-P., 1995. A kinetic model for product formation of microbial and mammalian https://doi.org/10.1016/j.vaccine.2017.10.112. cells. Biotechnol. Bioeng. 46, 314–324. https://doi.org/10.1002/BIT.260460404. Venereo-Sanchez, A., Simoneau, M., Lanthier, S., Chahal, P., Bourget, L., Ansorge, S., Zeng, A.P., 1996a. Quantitative assessment of cell density effect on the metabolism and Gilbert, R., Henry, O., Kamen, A., 2017. Process intensification for high yield antibody production rate of hybridoma cells at high cell density. J. Biotechnol. 45, production of influenza H1N1 Gag virus-like particles using an inducible HEK-293 243–251. https://doi.org/10.1016/0168-1656 (95)00172-7. stable cell line. Vaccine 35, 4220–4228. https://doi.org/10.1016/j. Zeng, A.P., 1996b. Mathematical modeling and analysis of monoclonal antibody vaccine.2017.06.024. production by hybridoma cells. Biotechnol. Bioeng. 50, 238–247. https://doi.org/ Vis, M.A.M., Ito, K., Hofmann, S., 2020. Impact of culture medium on cellular 10.1002/(SICI)1097-0290 (19960505)50:3<238::AID-BIT2>3.0.CO;2-H. interactions in in vitro co-culture systems. Front. Bioeng. Biotechnol. 8, 1–8. https:// Zeng, A.-P., Deckwer, W.-D., 1995. Mathematical modeling and analysis of glucose and doi.org/10.3389/fbioe.2020.00911. glutamine utilization and regulation in cultures of continuous mammalian cells. Walsh, G., 2018. Biopharmaceutical benchmarks 2018. Nat. Publ. Group 36. https://doi. Biotechnol. Bioeng. 47, 334–346. https://doi.org/10.1002/BIT.260470308. org/10.1038/nbt.4305. Zhang, J., Collins, A., Chen, M., Knyazev, I., Gentz, R., 1998. High-density perfusion Walther, J., Lu, J., Hollenbach, M., Yu, M., Hwang, C., McLarty, J., Brower, K., 2019. culture of insect cells with a biosep ultrasonic filter. Biotechnol. Bioeng. 59, Perfusion cell culture decreases process and product heterogeneity in a head-to-head 351–359. https://doi.org/10.1002/(SICI)1097-0290 (19980805)59:3<351::AID- comparison with fed-batch. Biotechnol. J. 14 https://doi.org/10.1002/ BIT11>3.0.CO;2-H. biot.201700733. Zhang, Y., Stobbe, P., Silvander, C.O., Chotteau, V., 2015. Very high cell density Wang, M., Orwar, O., Olofsson, J., Weber, S.G., 2010. Single-cell electroporation. Anal. perfusion of CHO cells anchored in a non-woven matrix-based bioreactor. Bioanal. Chem. 397, 3235–3248. https://doi.org/10.1007/S00216-010-3744-2. J. Biotechnol. 213, 28–41. https://doi.org/10.1016/j.jbiotec.2015.07.006. Watanabe, M., Nishikawaji, Y., Kawakami, H., Kosai, K.I., 2021. Adenovirus biology, Zu, Y., Huang, S., Lu, Y., Liu, X., Wang, S., 2016. Size specific transfection to mammalian recombinant adenovirus, and adenovirus usage in gene therapy. Viruses 13. https:// cells by micropillar Array electroporation. Sci. Rep. 6 https://doi.org/10.1038/ doi.org/10.3390/V13122502. SREP38661. 14