Embryonic human kidney suspension FreeStyle™ 293-F cells (Thermo Fisher Scientific, United States) were grown in FreeStyle™ 293 expression medium (Gibco, United States) or Dynamis™ supplemented with 8 mM L-glutamine (for STR, Gibco, United States) or GlutaMAX™ (for shake flasks, Gibco, United States). Cells were cultured at 37 °C, 8% CO_2, and at 137 rpm in shake flasks (Thermo Fisher Scientific, United States) using a Minitron shaker incubator (INFORS HT, Switzerland) with an orbit of 5 cm. The adherent NIH/3T3 murine fibroblast target cells (ATCC CRL-1658) were maintained in Dulbecco’s modified Eagle medium high glucose, pyruvate (DMEM; Gibco, Germany), supplemented with 10% fetal bovine serum (FBS; Gibco, United States) at 37°C in a humidified atmosphere at 5% CO₂. Cell number and viability was accessed using a cell counter (anvajo GmbH, Germany).

The MLV-based retroviral transfer vector pLEGFP-N1 (Clontech, United States) harbors the enhanced green fluorescent protein (EGFP) and the neomycin resistance (neoR) genes. The generation of the transfer vector, the packaging construct, the envelope construct in transposon vector backbones and the transposase construct was described previously (Berg et al., 2019).

The SB100X gene was amplified from pCMV-SB100X (Berg et al., 2019) and inserted into pIVTRup (a gift from Ángel Raya (Addgene plasmid #101362; https://n2t.net/addgene:101362; RRID:Addgene_101362). This plasmid served as a template for PCR amplification using primers containing the T7 promoter and polyT tail sequences respectively. The resulting amplicons were subjected to in vitro transcription (IVT) of SB100X-mRNA using HiScribe™ T7 ARCA mRNA Kit (NEB, USA) following the manufacturer’s instructions, respectively. After DNase treatment, the mRNA was purified using the Monarch^® RNA Cleanup Kit (NEB, United States). RNA purity was confirmed using a Tecan Infinite^® and stored in aliquots at −80°C. The final mRNA encompassed the 5′-Cap, the 5′-UTR, the coding sequence of the transposase, the 3′-UTR and the polyA tail (Tschorn et al., 2022).

Packaging cells were generated by co-transfection of 3 × 10⁷ 293-F cells in 20 mL shake flask cultures with 35.6 µg of pSB-gag/pol, 11.9 µg of pSB-env and 2.5 µg of the transposase construct using polyethylenimine transfection reagent (PEI:DNA mass ratio of 3:1; linear PEI, 1 mg/mL, MW 40,000; Polysciences Inc., United States). 9 mL fresh medium was added 3 hours later and a complete medium exchange was performed on the following day. Two days post-transfection, cells were subjected to 4 μg/mL puromycin and 50 μg/mL hygromycin (both InvivoGen, France). Every passage, the concentration of both antibiotics was escalated to a final concentration of 10 μg/mL puromycin and 200 μg/mL hygromycin resulting in a stable VPC bulk population called MuPACK.e within 21 days.

The subsequent transfections with the respective transfer vectors were performed as described in detail (Bauler et al., 2020). For 30 million cells a total amount of 16.5 μg of pDNA/mRNA was co-transfected. A mass ratio of 1:10 (transposase to transfer vector) for the plasmid-based transposase construct and a mass ratio of 1 to 1 for the mRNA encoding the transposase was used (PEI:DNA mass ratio of 2:1 and PEI:mRNA mass ratio of 4:1; linear PEI, 1 mg/mL, MW 40,000; Polysciences Inc., United States). The mRNA-PEI mixes in this case were always prepared in a separate tube and PEI was diluted directly into the mRNA-medium mixture. VPCs stably expressing a transfer vector were subjected to G418 mediated selection pressure 4 days post-transfection at increasing concentrations of G418 ranging from 50 μg/mL to a final concentration of 200 μg/mL. Two weeks post antibiotic selection with G418, all three antibiotics were added at final concentrations of 10 μg/mL puromycin and 200 μg/mL hygromycin and G418. After 3 weeks, for transposition-based transfection and 2 month for classical plasmid-based transfection, and rigorous selection, cells were expanded and cryo-stocks were made and rigorous selection, cells were expanded and cryo-stocks were prepared.

Stable VPCs were seeded at a viable cell density of 2 × 10⁶ cells/mL in 500 mL shake flasks in 100 mL antibiotic-free FreeStyle™ medium or Dynamis™ supplemented with 8 mM GlutaMAX™. After 24 h of production, retroviral vectors were harvested by centrifugation at 100 g for 3 min at RT and made cell-free using a PVDF syringe filter with a 0.45 µm pore size (Carl Roth, Germany). Retroviral vector preparations were frozen at −80°C in 1.8 mL aliquots. For high-density VPC cultivations, the stable VPC was seeded at 4 × 10⁶ cells/mL in 250 mL shake flasks in 50 mL Dynamis™ medium supplemented with 8 mM GlutaMAX™ and MLV-based vectors were harvested after 48, 72 and 96 h.

For larger-scale vector production, cultivation in an STR with a 500 mL working volume (DASGIP^® Parallel Bioreactor System, Eppendorf AG, Cat. 76DG04CCBB) was performed. The STR was equipped with one inclined blade impeller (three blades, 30°angle, 50 mm diameter) and a macro-sparger. Production parameters are shown in Table 1 Prior to inoculation, the stable VPC MuPACK.e.SB-LEGFP.N1_mRNA was thawed and expanded in Dynamis™ medium supplemented with highest concentrations of antibiotics in shake flasks for 2 weeks. When cultures showed viabilities >90% and a density of 2 × 10⁶ cells/mL, cells were centrifuged (300 g, 5 min, RT) and the complete medium was replaced with fresh antibiotic-free Dynamis™ medium. The STR was inoculated with 0.8 × 10⁶ cells/mL and ran at 37°C, pO₂ ≥ 40%, controlled pH 7.0 (deadband ±0.3), and 150 rpm for 10 days. During cultivation thirteen independent viral vector harvests (5 mL) were taken from the cultivation vessel, made cell-free by centrifugation (3,000 g, 10 min at 4°C) and stored at −80°C. The STR was operated in batch mode.

To assess the viral vector titers produced by the established VPCs, 1.0 × 10⁵ adherent target NIH/3T3 murine fibroblasts were seeded in 2 mL per well in six-well dishes 1 day prior to transduction (Nunc, Wiesbaden, Germany). Dilutions of 1:1,000 and 1:10,000 and of retroviral vector samples produced in shake flasks in total volumes of 1 mL were added to target cells. The following day, 1 mL of fresh cultivation medium was added to transduced cells. Three days post-transduction, the percentage of EGFP-positive cells was analyzed using flow cytometry (S3e, Bio Rad, United States; FlowJo BD Biosciences, United States) and used to detect gene transduction efficiencies. Vector titers described as transducing units per mL (TU/mL) were calculated as follows: titer = (F%/(100 × V_mL)) × S × D wherein F% is the percentage of GFP-positive transduced cells, S represents the number of seeded target cells on the day of transduction, D the dilution factor and V_mL the volume of viral vector in mL (Fehse et al., 2004).

For the viral vector titration using vectors produced in STR, adherent target cells were seeded in 48-well dishes at 1 × 10⁴ cells/well in 0.5 mL 1 day prior to transduction. The medium was removed and vector containing supernatant samples in different dilutions of 1:10, 1:100 and 1:1,000, respectively, in a total volume of 0.25 mL were added to the target cells. Three days post-transduction, cells were analyzed employing flow cytometry to determine the percentage of EGFP-positive cells. Vector titers were calculated as described previously employing supernatant dilutions resulting in gene transfer efficiencies between 1.0% and 10.0% EGFP-positive cells (Salmon and Trono, 2007).

To detect the fluorescence intensities of the three VPCs or transduced target cells expressing the EGFP expressing transfer vector, 1 × 10⁶ cells were centrifuged at 100 g for 5 min and the cell pellet was diluted in 1 mL flow cytometry buffer (phosphate-buffered saline (PBS), pH 7.2, 0.5% bovine serum albumin (BSA) and 2 mM EDTA). Prior to flow cytometry, viability was determined using an anvajo cell counter (anvajo GmbH, Germany). A total of 10,000 gated single cells were subsequently analyzed for EGFP expression.

To assess the efficiency of viral vector production, total particle concentration (i.e., physical titer) were quantified using a colorimetric MuLV core p30 antigen ELISA kit (Cell Biolabs, Inc. (Cat. VPK-156, United States)). From each cell-free retroviral particle harvest, one sample was used to detect the total p30 concentration. Samples were thawed from −80°C and diluted 1:10,000 in expression medium and assayed in a 96-well plates in duplicate.

To quantify the viral vector transfer vector RNA (i.e., physical titer), a real-time reverse transcription quantitative PCR (RT-qPCR) was performed. Cell-free and viral vector-containing cell culture supernatant was used for the extraction and purification of vector RNA according to the manufacturer’s instructions (NucleoSpin^® RNA virus kit; Macherey-Nagel, Germany).

A two-step hot start RT-qPCR with sequence-unrelated tagged primers was used to specifically quantify viral vector EGFP mRNA copies (Kawakami et al., 2011). Briefly, an external calibration curve was generated for the EGFP-encoding sequence by amplifying from the pLEGFP-N1 transfer vector template plasmid using the primers: T7-gag/EGFP for 5′- TAA​TAC​GAC​TCA​CTA​TAG​GGA​TGG​TGA​GCA​AGG​GC -3′ and T7-gag/EGFP rev 5′- GCT​AGC​TTC​AGC​TAG​GCA​TCT​TAC​TTG​TAC​AGC​TCG​TCC -3’. 300 ng of the amplicons were in vitro transcribed to RNA for 2 h at 37°C using TranscriptAid T7 High Yield Transcription Kit (ThermoFisher Scientific, United States). Transcribed RNA standards were treated with 10 vol% DNase (30 min, 37°C) followed by 10 vol% EDTA treatment (15 min, 65°C) and purified using an RNA isolation kit (Macherey Nagel, Germany).

Subsequently, a hot start reverse transcription PCR was performed. Here, 1 µL of each EGFP mRNA sample and of each generated RNA standard (ranging from 5.0E-07 ng to 5.0E+00 ng), 0.5 µL of dNTPs, 6.5 µL of nuclease-free water and 0.5 µL MLV EGFP tagged RT primer (rev 5′- GCT​AGC​TTC​AGC​TAG​GCA​TCT​TAC​TTG​TAC​AGC​TCG​TCC​A -3)’ was first incubated at 65°C for 5 min and then at 55 °C for 5 min. For cDNA synthesis, 2 µL of 5X RT buffer (Thermo Fisher Scientific, USA), 1.25 µL of nuclease-free water, and 0.25 µL of Maxima H minus reverse transcriptase (Thermo Fisher Scientific, United States) were added and incubated (30 min, 60°C), before the reaction was terminated (5 min, 85°C). The generated cDNA was diluted to 100 µL.

To perform the qPCR, 4 µL diluted cDNA, 5 µL of 2X QuantiNova SYBR green PCR mix (QIAgen, Germany), and 0.5 µL each of 1 μM primers EGFP qPCR for 5′- CTCGCCGACCACTACC -3′ and EGFP tagged qPCR rev 5′- GCT​AGC​TTC​AGC​TAG​GCA​TC-3′ were mixed. For the real-time quantification, samples were subjected to initial denaturation (5 min, 95°C), before 40 amplification cycles (10 s, 95°C; 20 s, 62°C) were carried out. The melt curve analysis was between 65°C and 90°C. For absolute quantification, a regression curve analysis was formulated by plotting the CT values of ten-fold diluted RNA standards against the log10 number of the RNA molecules (Frensing et al., 2014).

To ensure that detected gene transfer efficiencies were purely a result of vector-mediated transduction and not caused by the unintended generation of RCRs originating from the recombination of complementary vector components, GFP marker rescue assays were performed in triplicate as previously described in detail (Cosset et al., 1995; Berg et al., 2019; van Heuvel et al., 2021). In addition, NIH/3T3 target cells were exposed to vector preparations, expanded and supernatant of transduced cells was collected after five and 12 days. Subsequently, samples were examined using a reverse transcriptase (RT) assay with a detection sensitivity of 10 pg RT per 40 μL sample (Colorimetric reverse transcriptase assay, Roche, Switzerland) following the manufacturer’s instructions. Supernatants of stable VPCs and NIH/3T3 cells exposed to the supernatant of naïve 293-F cells served as positive controls and negative controls, respectively (data not shown).

An unpaired Student’s t-test was used to calculate p values. p values of less than 0.05 were considered statistically significant (*(p ≤ 0.05), **(p ≤ 0.01), ***(p ≤ 0.001), ****(p ≤ 0.0001)). Graphs and statistics were calculated using GraphPad Prism 7 for Windows 10 software (GraphPad Software, Inc. United States).

The stable polyclonal suspension VPC MuPACK.e based on ecotropic MLV was established as described using the expression cassettes illustrated in Figure 1, namely, the packaging and envelope construct pSB-Gag/Pol and pSB-Env together with the transposase encoding plasmid pSB100X followed by selection using puromycin and hygromycin. Upon transfection with the constructs illustrated in Table 2 namely, i) only with pLEGFP-N1, ii) with pSB-LEGFP-N1 and the transposase construct and iii) with pSB-LEGFP-N1 and the mRNA of SB100X, stable cell pools were established in the presence of escalating concentrations of G418. Cell-free supernatants of the resultant VPCs MuPACK.e.LEGFP-N1, MuPACK.e.SB-LEGFP-N1 and MuPACK.e.SB-LEGFP-N1_mRNA were harvested and frozen at −80°C. Thawed harvests were subjected to three independent titration experiments conducted in triplicate using murine NIH/3T3 target cells. Transduced cells were analyzed 3 days later for EGFP expression.

As depicted in Table 3 MuPACK.e.LEGFP-N1 generated mean vector titers ranging from 9.63 × 10⁵ to 2.16 × 10⁶ TU/mL. Viabilities of the VPC at the time of vector harvests varied between 76% and 85%. With viabilities of always >90%, MuPACK.e.SB-LEGFP-N1 revealed slightly higher titers of 2.17 × 10⁶ to 3.21 × 10⁶ TU/mL. MuPACK.e.SB-LEGFP-N1 showed stable productivity over a period of 2 months in the presence as well as absence of selection pressure (data not shown). The VPC MuPACK.e.SB-LEGFP-N1_mRNA showed the by far highest productivity when cells were cultured in FreeStyle™ medium (harvests 1). Titers of 5.12 × 10⁷ TU/mL were detected at VPC viabilities of 80%, respectively.

When the VPC was expanded in Dynamis™ (harvests 2 and 3), known as one of the mediums of choice for batch- and fed-batch cultivation of highly efficient mammalian producer cells, to prepare for cultivation at high densities in an STR or in perfusion cultivation, cell viabilities varied between 73% and 90% and vector titer productivities of 2.00 × 10⁷ TU/mL and 3.10 × 10⁶ TU/mL were achieved. For high viable cell density cultivations (VCD), represented in Figure 2, VPCs were cultivated at 50 mL scale and vector particle harvests at VCDs of 9, 10 and 15 × 10⁶ cells/mL were titrated in NIH/3T3 cells. VCDs correlated with functional vector titers ranging from 3.58 × 10⁶ TU/mL at 9 × 10⁶ cells/mL to 3.43 × 10⁷ TU/mL at 15 × 10⁶ cells/mL in NIH/3T3 cells.

These results were supported by the median fluorescence intensities (MFIs) of the three VPCs detected by flow cytometry and represented in Figure 3. The highest expression of EGFP was detected in MuPACK.e.SB-LEGFP-N1_mRNA showing a significantly higher MFI of more than 7,000 compared to the two other VPCs with MFIs between 4,000 and 5,000 (p ≤ 0.0001).

To evaluate the total amount of MLV capsid protein p30 within the three cell-free VPC harvests, a colorimetric ELISA assay was performed, depicted in Table 3. Average p30 concentrations of all three harvests (not shown in Table 3) were for MuPACK.e.LEGFP-N1 8.65 × 10⁴ ng/mL (±2.65 × 10⁴), for MuPACK.e.SB-LEGFP-N1 4.34 × 10⁴ ng/mL (±0.65 × 10⁴) and for MuPACK.e.SB-LEGFP-N1_mRNA 1.23 × 10⁵ ng/mL (±0.58 × 10⁵).

The different functional vector titers were likely to result from different transfer vector transcript amounts available for packaging into the vector particles. Thus, RT-qPCR was performed in duplicate using frozen cell-free samples from all VPC vector harvests and EGFP-specific primers. As illustrated in Figure 4, the mean amount of mRNA detected confirmed the trend observed in functional vector titers obtained from all three VPCs. MuPACK.e.SB-LEGFP-N1_mRNA (1.16 × 10¹⁰ to 2.78 × 10¹⁰ copies/mL) revealed the highest amount of packaged transcripts as compared to MuPACK.e.SB-LEGFP-N1 (6.60 × 10⁹ to 8.96 × 10⁹ copies/mL), while MuPACK.e.LEGFP-N1 yielded the lowest amounts of encapsidated mRNA (1.42 × 10⁹ to 4.16 × 10⁹ copies/mL).

To enable vector production at a larger scale, a cultivation in STR was examined with the most productive stable VPC MuPACK.SB-LEGFP.N1_mRNA. Cells were cultivated for 10 days in 500 mL Dynamis™ medium supplemented with 8 mM L-glutamine in the absence of antibiotics and medium change as described in detail in materials and methods as well as in Table 1. Figure 5 shows the retroviral vector titers detected in NIH/3T3 cells during a 10 days STR procedure. The VPC culture revealed increasing productivity up to day 6. As illustrated in Figure 5A and at the onset of the culture process, the titers were rather low with about 1 × 10⁵ TU/mL correlating with the low VCD of less than 2 × 10⁶ cells/mL. With increasing VCDs up to 6 × 10⁶ cells/mL (Figure 5B) the transduction-competent particle numbers also increased, generating titers of up to 2.81 × 10⁶ TU/mL in NIH/3T3 cells at day 6. From day 6 to day 7, productivity stayed on a small plateau and from day 7 on, the VCDs continuously decreased to 3.72 × 10⁶ cells/mL resulting in declining titers of 2.04 × 10⁶ TU/mL down to 1.4 × 10⁵ TU/mL at the end of the STR process on day 10. A linear scalability was observed until day 6 of production remaining then on a plateau. From day 8 on and as no medium exchange was performed, the vector production rate, as well as vector titer, dropped in correlation with the decline in VCDs and viability, respectively. While osmolality slightly decreased over time from 249 to 212 mOsm, pH values remained considerably stable around 7.0 (+deadband) throughout the process (Figure 5C).