Evaluation of Purification Methods for Minimizing Transgene Expression Background During Viral Manufacturing

Abstract

Gene therapy has emerged as a promising therapeutic avenue, offering targeted treatments for various diseases. Purification of viral vectors presents a pivotal challenge, demanding the removal of impurities while preserving integrity and potency. During manufacturing, producer cells in transfection systems can be transiently transfected or retro-infected by the viral vectors they have just produced—a process referred to as “retro-transduction”—leading them to express the transgenes of interest. This can be a significant source of contamination in the viral solution pool, particularly when the transgenes encode extracellular, secreted proteins, resulting in cytotoxicity and reduced viral potency. Herein, we aimed to evaluate the efficiency of different viral purification systems commonly used in academic and industry settings in removing the transgene background from viral solutions. The efficiency of each system was assessed based on the levels of the secreted transgene Gaussia Luciferase (GLuc), which can be quickly detected in a solution and served as a readout for transgene background contamination in the viral pool during downstream processing. Through a systematic evaluation of purification methods, we identified the most effective approaches for producing pure viral batches with minimal transgene background, all while preserving viral potency and functionality. Our study revealed superior performance of batches that underwent purification via tangential flow filtration, which yielded over 90% reduction in GLuc background and the highest transduction efficiency rates. This work provides significant insights for advancing gene therapy applications that rely on the production of viral vectors encoding secreted transgenes.

INTRODUCTION

In recent years, gene therapy has emerged as one of the most promising therapeutic approaches with the potential to treat a broad spectrum of diseases with high specificity and long-lasting effects.^1,2 Viral vectors such as lentiviruses (LVs) are invaluable tools in gene therapy, serving as competent genetic material delivery platforms, providing efficient, stable, and long-term transgene expression in several target tissues.^3

–6 Despite the progress achieved in the area, viral vector manufacturing capabilities have fallen behind, especially at a research scale.^7,8 Academic facilities are usually restricted to the initial stages of gene therapy clinical development due to scalability challenges, inefficient downstream purification processes, supply chain limitations, and regulatory requirements.

Viral vector manufacturing is a field with many trade secrets and high variability. Unlike conventional small-molecule drugs, biopharmaceutical manufacturing can face challenges at essentially every step. For instance, purification of viral batches is a critical step to remove impurities and contaminants while preserving vector potency and integrity.^9,10 That step is especially crucial for producing viral vectors that encode secreted transgenes. During viral manufacturing, producer cells can be transiently transfected or retro-infected by the viral particles they have just produced and released into the supernatant, leading to the unintended stable transduction of these cells—a process known as “retro-transduction.”¹¹ These mechanisms lead to the expression of genes of interest by the producer cells and, in the case of extracellular proteins, result in the improper secretion of these proteins into the viral pool supernatant.

Transduction of producer cells can lead to the loss of up to 90% of viable viral particles, significantly reducing viral titers and overall production efficiency.^12,13 Undesired transgene contaminants may remain in the viral solution even after downstream processing (DSP) and purification, directly impacting viral integrity and potency. Furthermore, if the resulting viral solution is not adequately purified of transgene background and contaminants, it can negatively impact transduction efficiency, as therapeutic transgenes may be cytotoxic or even induce apoptosis of target cells.¹⁴ This transgene background can also increase the complexity of quality control of viral batches, raising safety concerns, such as variable efficiency and increased risk of immunogenicity, adding time to the manufacturing process. That becomes particularly important in larger production scales for in vivo applications, especially in academic settings where purification capabilities are more limited.

In this study, we aimed to evaluate viral purification methods commonly used in academia and industry, exploring various approaches applicable to different viral production scales. Importantly, the elimination of secreted transgene contaminants from the viral batch was assessed using the secreted biomarker Gaussia Luciferase (GLuc) that can be easily detected in solution.^15,16 Finally, the transduction efficiency of each viral batch was compared to determine whether minimizing transgene background resulted in differences in viral potency and effectiveness. Overall, we evaluated purification systems that facilitated the production of pure viral batches with reduced secreted transgene contaminants, making them particularly suitable for gene therapy applications involving extracellular therapeutic transgenes.

MATERIALS AND METHODS

Lentiviral transient transfection: General reagents and procedures

Human embryonic kidney cell line HEK293T (ATCC CRL-3216) was expanded for 2 days in T175 flasks in Dulbecco's Modified Eagle Medium/Nutrient Mixture F-12 (DMEM/F12) medium with 10% fetal bovine serum (FBS) (Gibco) and 1% antibiotic/antifungal (ThermoFisher). The HEK cell line starting stock always ranged from passages 6–10. Cells were trypsinized using 1× trypsin in Phosphate-Buffered Saline (PBS) and centrifuged at 300 g for 5 min to pellet cells and resuspended in DMEM/F12 complete. Plasmids used for LV transfections were LV-EF1α-RFP, LV-EF1α-GLuc-IRES-RFP, or LV-sGP130-IRES-GLuc, pspax2 (packaging), and pmd2g (envelope) (Addgene), which all contain ampicillin resistance genes. Plasmid maxi-prep stocks were prepared by transforming plasmids in One Shot Stbl3 Chemically Competent E. coli cells (ThermoFisher). Transformed cells were plated on LB agar plates with ampicillin and cultured overnight at 32°C. Colonies were selected to smear on new ampicillin plates and cultured overnight at 32°C. Bacteria was collected from the plates using a pipette tip and transferred to 500 mL liquid Luria-Bertani agar (LB) broth cultures containing 100 μg/mL ampicillin and cultured in a shaking incubator at 225 rpm at 32°C overnight. Plasmid DNA was isolated and purified using the PureLink HiPure Plasmid Maxiprep kit (ThermoFisher) by column purification. DNA was eluted in 500 μL Tris-EDTA Buffer (TE) buffer (10 mM Tris-HCl, 1 mM Ethylenediaminetetraacetic Acid (EDTA), pH 8.0) (ThermoFisher) and quantified. Quantification of DNA concentration was performed using μDrop plates and read using Varioskan LUX (ThermoFisher) plate reader. Preps were ensured to be within a range of 1–2 µg plasmid/mL, minimum. Polyethylenimine (PEIpro, Polyplus-transfection) was used for polyethylenimine in transfections. A reduced serum medium was used to optimize transfection conditions. OptiMEM (ThermoFisher) was prepared without serum or antibiotics and with a pH of 7.2 using sodium bicarbonate. For collection, a rinse step was performed with a 0.001% Pluronic F-68 solution. Supernatant collection after transfection was filtered using 0.45 μM nylon membrane syringe filters (Corning) attached to a 20 mL syringe. For large-scale transfections greater than 100 mL, a 0.22 μM cellulose acetate membrane vacuum filter unit (Corning) was used to filter supernatant.

10-cm and 15-cm dish transfections

For small-scale transfections, 10- or 15-cm diameter culture dishes were used (Corning). Cells were seeded at 70,000 cells/cm² one day prior to transfection unless otherwise specified. For 10- and 15-cm dish transfections, PEI:DNA solution was made up in 0.5 and 1 mL solutions, respectively. To do so, PEI was added at a ratio of 2.5 μL PEI per 1 μg of DNA. For 10 cm dish transfections, 52.5 μL PEI was added to OptiMEM for a total volume of 0.25 mL. In a separate tube, 21 μg of plasmid DNA total was added to OptiMEM for a total volume of 0.25 mL. Both tubes were mixed by vortexing, and then PEI solution was added to the DNA solution and mixed by vortexing. Masses of DNA for each plasmid were calculated based on plasmid size (base pair length) and concentration to yield a 1:1:1 molar ratio of the transgene, packaging, and envelope plasmids. Plasmid length was converted to a molar concentration using the New England Bio molarity calculator for DNA. The PEI-DNA solution was incubated at room temperature for 15 min to allow complexes to form. Media was replaced in the 10-cm dish with 10 mL of OptiMEM. After incubation, the PEI-DNA solution was added dropwise to the dish, ensuring that all areas of the dish were covered and swirled gently to mix. The same process was used for 15-cm dish transfections, but the values were scaled based on the increase in cell number and surface area.

The following day, media was replaced with 10 mL fresh OptiMEM, and cultured for an additional 48 h before viral particle harvesting. Supernatant was collected, and the dish was rinsed with 5 mL 0.001% Pluronic solution. Both solutions were filtered through a 0.45-μm syringe filter and stored in single-use aliquots at –80°C.

HYPERFlask® transfections

For large-scale transfections, a HYPERFlask® (Corning), which is a 10-layer flask with 1,720 cm² surface area, was used. One day prior to transfection, 13.8 million cells were seeded in the flask. To do so, 20 mL of cell suspension was added to 500 mL DMEM/F12 complete medium and mixed by stirring in a liter bottle. The solution was then added to the HYPERFlask® by pouring at a 45° angle, according to the manufacturer protocol. Slow pouring was performed to ensure even spreading across all ten layers of the flask and to mitigate foam production. An additional 100 mL DMEM/F12 was added to fill the flask to the top cap thread. The top and bottom two layers of the flasks were checked under the microscope to ensure even cell mixing. The flask was incubated horizontally overnight to allow for cell attachment.

The following day, DNA and PEI solutions were prepared in 10 mL of OptiMEM, and the PEI solution was added to the DNA solution in a 50 mL conical tube. A total of 1,680 μL of PEI was used with a total of 672 μg of DNA. After incubation, the PEI-DNA solution was added to 500 mL OptiMEM in a liter bottle and mixed gently with a 50 mL serological pipette. The HYPERFlask® supernatant was discarded by pouring. The OptiMEM containing transfection complexes was added to the HYPERFlask® by pouring. Cells were cultured for 5 days and the supernatant was collected by pouring. Cells were rinsed with 200 mL of Pluronic solution and collected by pouring.

Lentiviral purification

Sucrose-gradient purification

Viral supernatant was collected from 10 cm dishes and passed through a 0.45-µm filter into a 50 mL conical tube using a 20 mL syringe. A total of 3 mL of 20% sucrose solution was laid in a conical bottom 30 mL ultracentrifuge tube. Very carefully, filtered medium was laid on top of the sucrose solution, and two separate layers were clearly visible. The tube was topped up with fresh DMEM to avoid collapsing during the ultracentrifugation spin. Samples were ultracentrifuged at 75,000 g for 90 min at 4°C. Supernatant was discarded into a bleach pot, and tubes were kept inverted on paper tissue for 10 min. The LV pellet (not visible to the naked eye) was then gently resuspended in 3 mL of sterile PBS.

Lentiviral concentrator

Viral supernatant was collected from HYPERFlask® batches and briefly centrifuged at 500 g for 10 min to remove cells and debris. The clarified supernatant (3 volumes) was combined with 1 volume of Lenti-X concentrator (Takara Bio, Cat #631231). The mixture was incubated at 4°C overnight. Samples were then centrifuged at 1,500 g for 45 min at 4°C. The supernatant was carefully removed, and the pellet was gently resuspended in 3 mL of sterile PBS.

Tangential flow filtration

For batches larger than 100 mL (from either multiple dishes or HYPERFlask®), a tangential flow filtration (TFF) system was used to concentrate and purify LV from the supernatant, as previously described.¹⁷ The Minimate benchtop TFF system (Pall) was optimized for LV concentration. Using a 100 kDa Molecular Weight Cut-Off (MWCO) membrane capsule (Pall), the reservoir was connected with tubing to the membrane capsule to the TFF capsule at the feed entry point, and tubing directed the filtrate to a waste container and the retentate back into the reservoir. Pressure gauges were placed upstream and downstream of the capsule to be monitored during the run. A peristaltic pump was connected to the tubing between the reservoir and the capsule to pump the feed into the system. Clamps were placed on the retentate and filtrate tube so that pressure could be altered and filtrate flow could be shut off during collection. The peristaltic pump was set to 70 rpm to achieve a pressure of 20 psi in the upstream pressure gauge, which is recommended by the user manual. The screw clamp on the filtrate tube was adjusted to maintain a pressure of 20 psi in the downstream tubing as well.

The supernatant and PBS rinse solution were added to the reservoir, and the pump was started. Every 20 min, the reservoir levels were checked, and pump speed was adjusted to maintain 20 psi pressure in the upstream and downstream tubing. When the volume in the reservoir reached 50 mL, the pump was paused, the filtrate flow clamp was closed off, and the retentate flow clamp was loosened entirely so there was no pressure in the retentate tubing. The flow direction was then reversed, and the pump speed was increased to 200 rpm to remove any cake layer on the membrane. This step ensures the viral particles are in solution of the retentate. The clamps were then replaced, and the flow direction was returned until the retentate solution approached 25 mL. The filtrate tube was then clamped off and then the three-way valve on the retentate tube was opened, and retentate was collected. An additional 5 mL of PBS was added to remove the hold-up volume of the system, and the last 5 mL of retentate volume was collected. To flush the system, 200 mL of PBS was added to the reservoir, and the system was pumped in reverse for 15 min and then collected. The total processing time on average was about 90 min and was performed at room temperature.

The system was cleaned using a 0.5 M NaOH solution in diH₂O. Briefly, the solution was pumped at 300 rpm with the retentate and filtrate clamps fully loosened for 20 min. The flow was reserved, and fresh solution was pumped for an additional 20 min. All tubing and connectors were removed and rinsed with soap and water and then with 0.5 M NaOH before drying. The membrane cassette was filled with 0.5 M NaOH, closed off with luer caps at all ports, and stored.

Collection, storage, and titration

With final batch products in PBS + Pluronic-F68 solution, 1 mL aliquots of the virus were created and added to cryovial tubes. Product aliquots were preserved by freezing at −80°C for up to 1 year and only thawed once immediately before use. A single aliquot was set aside for (qPCR) Quantitative Polymerase Chain Reaction viral titer quantification so that the titer was determined after one thaw cycle as well. Vector titer was determined by (RT-qPCR) Reverse Transcription qPCR using an LV titration kit (Applied Biological Materials, LV900) on Quant Studio 3 (Thermo Fisher).

Lentiviral transduction

HEK-293T cells (ATCC) were seeded at 25% confluence in 24-well plates and allowed to adhere overnight in DMEM complete (ATCC) with 10% FBS (GIBCO) and 1% antibiotic/antifungal. Lentiviral vectors were added at different multiplicities of infection (MOI) to each well in a total volume of 1 mL using antibiotic-free medium. A total of 20 ng/mL of the transduction enhancer Polybrene (Sigma-Aldrich) was also added to each well. Cells were incubated for 24 h before the medium was changed to complete medium. Post-transduction was initiated at 72 h, and plates were scanned for RFP fluorescence imaging and quantification using the Celigo® Imaging Cytometer (Nexcelom Bioscience, Lawrence, MA). Finally, 300 μL of supernatant was collected from each well for GLuc assessment.

GLuc assay

GLuc’s activity was measured using a bioluminescent flash assay. The substrate, native coelenterazine (CTZ), was reconstituted at 2.7 mg/mL in 200-proof ethanol. Working solutions were made fresh immediately prior to the assay with a 1:1000 dilution in PBS. Twenty microliters of each sample were added to black-walled, clear-bottom 96-well plates and 100 μL of CTZ solution and immediately read using a bioluminescent plate reader (Varioscan, Thermofisher). All samples were read forward and in reverse to account for any time-dependent signal degradation. No more than 12 samples were run at a time. GLuc activity in each sample was calculated by taking the average of the forward and reverse readings and expressed as relative light units.

Statistical analyses

Experimental data were analyzed using GraphPad Prism 9.5.1 for Mac OS X (GraphPad Software, La Jolla, CA). The results are shown as the mean ± standard deviation (SD), and significance is set at p ≤ 0.05. For multiple comparisons, one-way or two-way analysis of variance (ANOVA) with Tukey’s multiple comparisons post-test was used unless otherwise noted.

RESULTS

Process scale-up: Parameter characterization

For effective process scale-up, a universal metric for productivity must be employed. The number of viral particles produced per cell, for example, is a standard metric that can be used when scaling up. This metric is not commonly reported in literature, which makes optimization parameters hard to compare, but it is reported more frequently in suspension-based transfection cultures. In adherent cell-based systems such as ours, the range of viral production per cell can vary substantially,¹⁸ depending on the viral type/serotype, size of transgene, culture vessel, etc. We sought to determine whether cellular production could be maintained with scale.

First, the total yield was quantified for several transgenes in both 10-cm and 15-cm dish transfections. For reference, 0.506 million cells were seeded in 10-cm dish transfections, and 13.8 million cells were seeded in 15-cm dish transfections, so the yield was expected to be approximately 2.75-fold higher. A total of four transfections were performed for each transgene and dish size, and viral titration was determined via qPCR (Fig. 1A and Fig. 2A). As mentioned, titer can be highly variable depending on the transgene, so we used a wide range of coding sequences, testing standard transgene cassettes as well as larger ones with more than one protein expressed in tandem. The total yield across all transgenes averaged 4–5-fold higher in 15- than in 10-cm dish transfections, which was beyond expected, maybe due to the specific plasmids used. Notably, the construct with the largest payload, encoding a secreted anti-cytokine protein (sGP130), exhibited lower overall yields, aligning with earlier findings in the literature.¹⁹

Figure 1.

Viral transfection yields vary in different culture vessel scales. (A) Total infection units (IU) for different lentiviral transgenes in 10 cm (blue) and 15 cm (red) dish transfections (n = 4 per group). (B) Normalized yield (IU/cell) for 10 cm, 15 cm and HYPERFlask® LV transfections (n = 4 for 10 cm and 15 cm, and n = 3 for HYPERFlask®). The transgene used was LV-EF1α-GLuc-IRES-RFP. Bars represent mean ± SD for each group. **p < 0.01, ***p < 0.001, ****p < 0.0001, two-way ANOVA.

Figure 2.

Schematic of the workflow used to evaluate various purification methods. (A) Lentiviral vectors expressing a secreted transgene and a fluorescent tag were used in all the applications. (B) Two distinct transfection scales (10 cm dish and HYPERFlask®) were used to assess different purification methods: sucrose-gradient, LV concentrator, and tangential flow filtration (TFF). (C) Downstream, purification efficiency was determined based on two parameters: viral titration and GLuc background, before and after purification. Finally, viral function and integrity before and after purification were determined via viral transduction using HEK-293T cells. Created with BioRender.com.

When normalized for cell seeding number, the average yield per cell did not significantly increase from 10- to 15-cm dish transfections, despite DNA, PEI, and cell number being scaled proportionally (Fig. 1B). When transfections were further scaled up to multilayer cell culture vessels (HYPERFlask®), the average yield exhibited a 10-fold increase compared with 15-cm dishes and an even greater increase compared with 10-cm dishes.

DSP: Minimizing transgene background

Implementing robust purification processes to remove contaminating particles from viral batches is crucial to ensure quality control of these biopharmaceuticals. We next investigated different purification processes to minimize transgene background products in viral batches. For a more precise determination of purification efficiency, LVs expressing a secreted biomarker (GLuc) and a red fluorescent protein (RFP) tag under the strong constitutive promoter EF1α were employed (Fig. 2A). Different production scales were compared (10 cm and HYPERFlask®) and suitable purification methods were employed for each (Fig. 2B). Downstream, the purification efficiency of each method was evaluated based on GLuc background, and viral transduction analyses were carried out for each different batch, as illustrated in Figure 2C.

Ten-cm dish transfections were carried out using a sucrose-gradient purification protocol, a well-documented method in existing literature.^20,21 Batches were analyzed for viral titration before and after purification, as an ideal protocol should result in minimal viral loss. Figure 3A shows an approximate 1.5-fold reduction in the total number of viral particles postsucrose purification, a significant decrease worthy of consideration. Next, GLuc background was determined before and after purification in 10-cm dishes, as depicted in Figure 3B. A significant reduction in the secreted transgene was observed, suggesting good purification efficiency to eliminate impurities and undesirable background.

Figure 3.

Evaluation of purification efficiency of different methods in minimizing transgene expression background. (A) Sucrose-gradient purification was performed with 10-cm dish transfections and viral titration, and (B) GLuc background was determined before and after purification (n = 6). (C) HYPERFlask® batches were concentrated using a commercially available lentiviral concentrator kit and viral titration, and (D) GLuc background was determined before and after purification (n = 6). (E) Another group of 6 HYPERFlask® batch replicates were purified via tangential flow filtration, and, similarly, viral titration and (F) GLuc background were determined before and after purification. (G) Percentage decrease was determined for viral titration and GLuc background parameters in each group and analyzed for statistical significance. (H) Purification ratio was determined for each group as a proportion between GLuc background and viral titration reduction. Bars represent mean ± SD for each group. *p < 0.05, **p < 0.01, ****p < 0.0001, two-way ANOVA.

HYPERFlask® transfection batches underwent two distinct purification methods: half of them were treated with a commercially available lentiviral concentrator solution, capable of reducing the overall viral suspension volume, enabling the elimination of most residues and debris from the starting supernatant; the other half underwent TFF, a widely used method in bioprocessing and separation technology, that operates on the principle of crossflow filtration, where the feed solution flows tangentially across the surface of a semi-permeable membrane (see Materials and Methods). Both original and concentrated viral solutions were analyzed for final titration and GLuc background. Viral concentration led to a significant loss of viral particles, resulting in an approximately 8-fold decrease in the final count (Fig. 3C). Interestingly, the GLuc background was not significantly reduced, as it would have been expected (Fig. 3D). On the contrary, TFF resulted in the lowest loss of viral vectors while significantly reducing GLuc background by 16-fold (Fig. 3E–F). An optimal purification method should minimize its impact on viral titration while maximizing the reduction in transgene background. Figure 3G depicts the percentage decrease for both parameters, confirming an optimal profile in the HYPERFlask® + TFF group, that is, the lowest impact on titration (less than 15%) and the most significant reduction in GLuc background (over 90%). In addition, when these two parameters were correlated as a purification ratio, the TFF group displayed the highest score, consistent with previous observations (Fig. 3H).

Viral potency assessment post-DSP

Next, the potency and effectiveness of purified and non-purified viral batches were evaluated and compared via viral transduction. First, HEK-293T cells were infected with increasing MOIs of 10-cm viral batches, and transduction efficiency was determined 72 h later. Cells were assessed for RFP fluorescence and GLuc secretion in the supernatant. Importantly, as we transitioned from discussing transfection to transducing target cells, GLuc secretion was now utilized as a readout for transduction efficiency rather than as a “contaminant” parameter. Figure 4A shows increasing RFP fluorescence with increasing MOIs, with a significant difference between purified and non-purified batches (Fig. 4B). In addition, GLuc secretion in the supernatant followed a similar pattern, with a significantly higher expression in batches that underwent sucrose-gradient purification (Fig. 4C). This observation suggests not only the preservation of viral function postpurification but also a potential enhancement thereof.

Figure 4.

Assessment of viral potency and effectiveness postsucrose-gradient purification. (A) HEK-293T cells were treated with increasing doses of purified or non-purified viral batches from 10-cm transfection batches. Cells were imaged 72 h posttreatment and (B) RFP fluorescence was quantified in each group (n = 3). (C) Cell supernatants were collected 72 h post-transduction and tested for GLuc activity (see Materials and Methods). GLuc secretion was quantified for each group (n = 3). Bars represent mean ± SD for each group. *p < 0.05, ***p < 0.001, ****p < 0.0001, two-way ANOVA.

Finally, viral function was also assessed in purified and non-purified HYPERFlask® batches. Considering the purification ratios previously obtained, only the TFF HYPERFlask® group was tested. Figure 5A shows a marked increase in RFP fluorescence with increasing MOIs, which is substantially more pronounced in the purified group. The quantification of fluorescence in each group confirms that observation (Fig. 5B). Ultimately, GLuc secretion exhibited a parallel trend, with significantly higher expression in groups that underwent TFF purification at MOIs 50 and 100 (Fig. 5C). Noteworthy, the findings here observed mirror the purification efficiency analysis presented in Figure 4, reaffirming TFF as the most effective method for transgene background purification, all while preserving viral potency and integrity.

Figure 5.

Assessment of viral potency and effectiveness post-TFF purification. (A) HEK-293T cells were treated with increasing doses of purified or nonpurified viral batches from HYPERFlask® transfection batches. Cells were imaged 72 h post-treatment and (B) RFP fluorescence was quantified in each group (n = 3). (C) Cell supernatants were collected 72 h post-transduction and tested for GLuc activity (see Materials and Methods). GLuc secretion was quantified for each group (n = 3). Bars represent mean ± SD for each group. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, two-way ANOVA.

DISCUSSION

The translation of promising viral vector-based therapies from bench to bedside faces significant challenges, particularly in the production phase.^18,22,23 At the research scale, where innovative therapies are developed and optimized, the manufacturing process is often hindered by numerous bottlenecks that impede efficiency and scalability.²⁴ This can hinder technological advancements and translation to clinic as in vivo studies require a large number of vectors that are usually beyond the production capabilities of research laboratories. For instance, in vivo experiments in mice typically require lentiviral doses ranging from 1 × 10⁸ to 1 × 10⁹ IU per animal. ^25,26 A standard research lab usually produces between 1 × 10⁶ and 1 × 10⁷ IU/mL in one standard 15 cm-dish batch, with potential further reduction postpurification steps.

Hence, the establishment and evaluation of production systems and purification methods that can be effectively scaled in academia is imperative, as novel gene therapies are developed and the field continues to advance.^27
–29 When examining transfection yields for LVs expressing diverse gene cassettes, it became apparent that longer coding sequences resulted in slightly lower titers, as anticipated.¹⁹ Furthermore, HYPERFlask® contributed to better productivity due to a 10-fold higher cell yield that increases capacity and reduces processing time and incubator space. These multilayer flasks use a gas-permeable film to enable efficient gas exchange, contributing to better productivity and supporting the results here observed.³⁰

Just as there are entire companies focused on the rational design of viral capsids, an industry has sprung up around viral vector production.^{10,18,31
–33} Inefficient manufacturing protocols can result in viral batches with low yields, shorter shelf life and impaired purity, hindering their therapeutic efficacy.^9,18 Currently, there is a shift toward using serum-free media in suspension culture processes, as there are no serum-derived contaminants to eliminate, resulting in significantly higher vector purity from the beginning of DPS.^34,35 While these can be scaled down to smaller shakers in a research laboratory, this type of manufacturing is not typically the gold standard platform in academia; instead, adherent cell cultures are more commonly used, requiring closer attention to DSP and purification methods.

The removal of host cell proteins and transgene contaminants from the viral pool has been a concern in the field for decades. It is known that the presence of therapeutic transgenes in the viral solution can significantly reduce viral titers, either by exerting cytotoxic effects on producer cells, disrupting their metabolism, or interfering with the assembly and functionality of the vector virions, ultimately affecting potency and effectiveness.^13,14 Maunder et al. previously attempted to implement a “Transgene Repression in Vector Production” (TRiP) system to block the translation of one or more transgenes during viral production, thereby inhibiting the release of proteins with cytotoxic or apoptotic effects on the system.¹⁴ Similarly, Blahetek and colleagues explored the use of artificial microRNAs for the transient suppression of toxic transgenes during viral production, successfully achieving effective silencing of the transgene and enhancing vector yields.³⁶ More recently, Banos-Mateos et al. described several strategies to reduce retro-transduction during LV manufacturing by blocking specific receptors on the surface of producer cells, thereby preventing viral interaction and entry.¹³ It is noteworthy that when transgenes are driven by inducible promoters rather than constitutive ones, expression is inherently suppressed until the activator is introduced into the system,³⁷ contributing to a reduction of transgene contaminants in the viral pool even in the presence of retro-transduction.

When such strategies are not feasible, however, the use of effective purification methods is imperative for the proper elimination of contaminants before quality control assays and transduction of target cells can proceed. This ensures that the final viral preparations meet the necessary standards for efficiency and safety. While previous studies have independently addressed transgene silencing and purification techniques—such as Kutner et al., who described common purification methods such as ion exchange chromatography and polyethylene glycol 6000 precipitation,³⁸ as well as a simplified LV production protocol using HYPERFlask® vessels,³⁹ or Kuroda et al., who introduced PEI-mediated transfections for more consistent lentiviral batch production⁴⁰—our focus was on the intersection of these approaches. Specifically, we aimed to determine which commonly used purification methods are most effective in reducing secreted transgene background within the viral pool. One important aspect not addressed in this study, which could be explored in future research, is the percentage of producer cells that were actually retro-transduced compared with those that were only transiently transfected. This could be further investigated by analyzing transgene integration in producer cells at the gene level through quantitative PCR, using specific primers to quantify the target transgene, rather than solely measuring it at the protein level as done in the present study.

When evaluating the purification efficiency of various research-scale methods in reducing transgene background, TFF demonstrated superior performance compared to others. This approach is known to be highly scalable and easily adapted for large-scale productions, being suitable for both laboratory-scale research and industrial manufacturing.^41,42 Furthermore, by allowing for gentle filtration of the sample through a semi-permeable membrane, TFF minimizes product loss, as reflected by the lowest % decrease in viral titrations. Moreover, TFF was effective in removing transgene background contaminants, resulting in batches with the lowest levels of GLuc. TFF is also known to be effective in removing contaminants such as aggregates, host cell proteins, and DNA from samples via size-based separation, resulting in higher purity of the final product⁴³ compared with density-gradient purification techniques, for example, that may not provide as effective removal of contaminants.^44,45 These can directly affect the health and viability of cells that are treated with nonpurified viral batches, which can also be a potential reason for decreased expression and lower transduction efficiency observed in those batches (Fig. 5). Notably, similar MOIs resulted in different transduction efficiencies in nonpurified viral batches (Figs. 4 and 5). This suggests that the different viral manufacturing scales—10 cm and HYPERFlask®, respectively—yield varying numbers of viable, fully mature viral vectors, indicating that manufacturing scale can affect viral integrity, which was not assessed in these studies. This issue could potentially be minimized through size-exclusion chromatography, which can effectively remove empty viral particles.⁴⁶

Although highly efficient, TFF can also present some challenges, such as shear sensitivity of some biomolecules to the shear forces generated during the process, which can result in potential product degradation;⁴⁷ membrane fouling due to accumulation of residues on the membrane surface, leading to reduced filtration efficiency and increased operating costs; limited molecular weight cutoff range due to TFF membrane size limitations.⁴⁸ For viral applications, however, TFF is highly suitable and offers exceptional versatility, which is particularly essential in academic settings. Our assessments further demonstrated that viral integrity was preserved postpurification, with an observed enhancement in viral potency during transduction (Fig. 5), suggesting an efficient separation of genome-devoid (empty) and mutant (incompletely assembled) virions from fully mature and functional vectors.⁴⁹

In conclusion, our study tackles a critical challenge in viral vector manufacturing: the efficient purification of viral pools and the reduction of secreted transgene background using readily available research methods. Moving forward, our report can contribute to paving the way for advancements in gene therapy research,⁶ offering valuable insights for the DSP of viral vectors that encode secreted therapeutic transgenes.

Footnotes

ACKNOWLEDGMENTS

The authors thank Rick Cohen for sharing reagents such as PEI.

AUTHORS’ CONTRIBUTIONS

R.B.: Conceptualization, Investigation, Methodology, Data curation, Formal analysis, Visualization, Writing—original draft. A.B.: Conceptualization, Investigation, Methodology, Data curation, Formal Analysis, Visualization, M.T.: Investigation, Methodology, Data curation, B.P.: Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing—review and editing.

AUTHOR DISCLOSURE STATEMENT

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

FUNDING INFORMATION

The authors declare that financial support was received for the research and publication of this article. Financial support for this project was provided by the National Institutes of Health (NIH) (Grant 1R01EB028782).

References

Dulak

. Gene therapy. The legacy of Wacław Szybalski. Acta Biochim Pol, 2021; 68(3):359–375; doi: 10.18388/abp.2020_5805

Lundstrom

. Viral vectors in gene therapy. Diseases, 2018; 6(2); doi: 10.3390/diseases6020042

Kotterman

, Chalberg

, Schaffer

. Viral vectors for gene therapy: Translational and clinical outlook. Annu Rev Biomed Eng, 2015; 17:63–89; doi: 10.1146/annurev-bioeng-071813-104938

Colón-Thillet

, Jerome

, Stone

. Optimization of AAV vectors to target persistent viral reservoirs. Virol J, 2021; 18(1):85; doi: 10.1186/s12985-021-01555-7

Poletti

, Mavilio

. Designing lentiviral vectors for gene therapy of genetic diseases. Viruses, 2021; 13(8); doi: 10.3390/v13081526

Bulcha

, Wang

, Ma

, et al. Viral vector platforms within the gene therapy landscape. Signal Transduct Target Ther, 2021; 6(1):53; doi: 10.1038/s41392-021-00487-6

Nishida

, Kodama

, Sengoku

. The gap between development and manufacturing in gene therapy: Strategic options for overcoming traps. Drug Discov Today, 2023; 28(2):103429; doi: 10.1016/j.drudis.2022.103429

Burr

, Erickson

, Bento

, et al. Allometric-like scaling of AAV gene therapy for systemic protein delivery. Mol Ther Methods Clin Dev, 2022; 27:368–379; doi: 10.1016/j.omtm.2022.10.011

Rodrigues

, Carrondo

, Alves

, et al. Purification of retroviral vectors for clinical application: Biological implications and technological challenges. J Biotechnol, 2007; 127(3):520–541; doi: 10.1016/j.jbiotec.2006.07.028

10.

Segura

, Mangion

, Gaillet

, et al. New developments in lentiviral vector design, production and purification. Expert Opin Biol Ther, 2013; 13(7):987–1011; doi: 10.1517/14712598.2013.779249

11.

Ohishi

, Shioda

, Sakuragi

. Retro-transduction by virus pseudotyped with glycoprotein of vesicular stomatitis virus. Virology, 2007; 362(1):131–138; doi: 10.1016/j.virol.2006.12.030

12.

Ferreira

, Cabral

, Coroadinha

. Progress and perspectives in the development of lentiviral vector producer cells. Biotechnol J, 2021; 16(1):e2000017; doi: 10.1002/biot.202000017

13.

Banos-Mateos

, Lopez-Robles

, Yubero

, et al. Abolishing retro-transduction of producer cells in lentiviral vector manufacturing. Viruses, 2024; 16(8); doi: 10.3390/v16081216

14.

Maunder

, Wright

, Kolli

, et al. Enhancing titres of therapeutic viral vectors using the transgene repression in vector production (TRiP) system. Nat Commun, 2017; 8(1):14834; doi: 10.1038/ncomms14834

15.

Wurdinger

, Badr

, Pike

, et al. A secreted luciferase for ex vivo monitoring of in vivo processes. Nat Methods, 2008; 5(2):171–173; doi: 10.1038/nmeth.1177

16.

Tannous

. Gaussia luciferase reporter assay for monitoring biological processes in culture and in vivo . Nat Protoc, 2009; 4(4):582–591; doi: 10.1038/nprot.2009.28

17.

Lock

, Alvira

, Vandenberghe

, et al. Rapid, simple, and versatile manufacturing of recombinant adeno-associated viral vectors at scale. Hum Gene Ther, 2010; 21(10):1259–1271; doi: 10.1089/hum.2010.055

18.

Merten

, Gény-Fiamma

, Douar

. Current issues in adeno-associated viral vector production. Gene Ther, 2005; 12(Suppl 1):S51–S61; doi: 10.1038/sj.gt.3302615

19.

Sweeney

, Vink

. The impact of lentiviral vector genome size and producer cell genomic to gag-pol mRNA ratios on packaging efficiency and titre. Mol Ther Methods Clin Dev, 2021; 21:574–584; doi: 10.1016/j.omtm.2021.04.007

20.

Ipseiz

, Czubala

, Bart

VMT

, et al. Effective in vivo gene modification in mouse tissue-resident peritoneal macrophages by intraperitoneal delivery of lentiviral vectors. Mol Ther Methods Clin Dev, 2020; 16:21–31; doi: 10.1016/j.omtm.2019.10.004

21.

Jiang

, Hua

, Wei

, et al. An optimized method for high-titer lentivirus preparations without ultracentrifugation. Sci Rep, 2015; 5:13875; doi: 10.1038/srep13875

22.

, Ataai

. Production of viral vectors for gene therapy applications. Curr Opin Biotechnol, 2000; 11(2):205–208; doi: 10.1016/s0958-1669(00)00080-x

23.

Cribbs

, Kennedy

, Gregory

, et al. Simplified production and concentration of lentiviral vectors to achieve high transduction in primary human T cells. BMC Biotechnol, 2013; 13(1):98; doi: 10.1186/1472-6750-13-98

24.

McCarron

, Donnelley

, McIntyre

, et al. Challenges of up-scaling lentivirus production and processing. J Biotechnol, 2016; 240:23–30; doi: 10.1016/j.jbiotec.2016.10.016

25.

Carbonaro-Sarracino

, Tarantal

, Lee

CCI

, et al. Dosing and re-administration of lentiviral vector for in vivo gene therapy in rhesus monkeys and ADA-Deficient mice. Mol Ther Methods Clin Dev, 2020; 16:78–93; doi: 10.1016/j.omtm.2019.11.004

26.

van den Brand

, Vermeij

, Waterborg

, et al. Intravenous delivery of HIV-based lentiviral vectors preferentially transduces F4/80+ and Ly-6C+ cells in spleen, important target cells in autoimmune arthritis. PLoS One, 2013; 8(2):e55356; doi: 10.1371/journal.pone.0055356

27.

Manceur

, Kim

, Misic

, et al. Scalable lentiviral vector production using stable HEK293SF producer cell lines. Hum Gene Ther Methods, 2017; 28(6):330–339; doi: 10.1089/hgtb.2017.086

28.

Dormond

, Meneses-Acosta

, Jacob

, et al. An efficient and scalable process for helper-dependent adenoviral vector production using polyethylenimine-adenofection. Biotechnol Bioeng, 2009; 102(3):800–810; doi: 10.1002/bit.22113

29.

Teryek

, Jadhav

, Bento

, et al. High-Throughput production of microcapsules for human bone marrow derived mesenchymal stem cell biomanufacturing in a vertical-wheel bioreactor. Biotechnol Bioproc E, 2023; 28(4):528–544; doi: 10.1007/s12257-023-0020-9

30.

Kearney

, Voorhees

, Williams

, et al. Corning HYPERFlask(®) for viral amplification and production of diagnostic reagents. J Virol Methods, 2017; 242:9–13; doi: 10.1016/j.jviromet.2016.12.011

31.

Wang

, Fu

, Park

, et al. Decoding cellular mechanism of recombinant adeno-associated virus (rAAV) and engineering host-cell factories toward intensified viral vector manufacturing. Biotechnol Adv, 2024; 71:108322; doi: 10.1016/j.biotechadv.2024.108322

32.

Sanber

, Knight

, Stephen

, et al. Construction of stable packaging cell lines for clinical lentiviral vector production. Sci Rep, 2015; 5(1):9021; doi: 10.1038/srep09021

33.

Merten

. State-of-the-art of the production of retroviral vectors. J Gene Med, 2004; 6(Suppl 1):S105–S24; doi: 10.1002/jgm.499

34.

Segura

, Garnier

, Durocher

, et al. Production of lentiviral vectors by large-scale transient transfection of suspension cultures and affinity chromatography purification. Biotechnol Bioeng, 2007; 98(4):789–799; doi: 10.1002/bit.21467

35.

Perry

, Rayat

. Lentiviral vector bioprocessing. Viruses, 2021; 13(2); doi: 10.3390/v13020268

36.

Blahetek

, Mayer

, Zuber

, et al. Suppression of toxic transgene expression by optimized artificial miRNAs increases AAV vector yields in HEK-293 cells. Mol Ther Methods Clin Dev, 2024; 32(3):101280; doi: 10.1016/j.omtm.2024.101280

37.

Jadav

, Truong

. Creation of a synthesis-friendly inflammation-inducible promoter suitable for cell therapy. Integr Biol (Camb), 2021; 13(9):230–236; doi: 10.1093/intbio/zyab015

38.

Kutner

, Zhang

, Reiser

. Production, concentration and titration of pseudotyped HIV-1-based lentiviral vectors. Nat Protoc, 2009; 4(4):495–505; doi: 10.1038/nprot.2009.22

39.

Kutner

, Puthli

, Marino

, et al. Simplified production and concentration of HIV-1-based lentiviral vectors using HYPERFlask vessels and anion exchange membrane chromatography. BMC Biotechnol, 2009; 9:10; doi: 10.1186/1472-6750-9-10

40.

Kuroda

, Kutner

, Bazan

, et al. Simplified lentivirus vector production in protein-free media using polyethylenimine-mediated transfection. J Virol Methods, 2009; 157(2):113–121; doi: 10.1016/j.jviromet.2008.11.021

41.

Musumeci

, Leonardi

, Bonaccorso

, et al. Tangential flow filtration technique: An overview on nanomedicine applications. Pharm Nanotechnol, 2018; 6(1):48–60; doi: 10.2174/2211738506666180306160921

42.

Tehrani

, Bharadwaj

, Leblond Chain

, et al. Purification processes of polymeric nanoparticles: How to improve their clinical translation? J Control Release, 2023; 360:591–612; doi: 10.1016/j.jconrel.2023.06.038

43.

van Reis

, Gadam

, Frautschy

, et al. High performance tangential flow filtration. Biotechnol Bioeng, 1997; 56(1):71–82; doi: 10.1002/(sici)1097-0290(19971005)56:1<71::Aid-bit8>3.0.Co;2-s

44.

James

, Cooney

, Agopsowicz

, et al. Novel high-throughput approach for purification of infectious virions. Sci Rep, 2016; 6(1):36826; doi: 10.1038/srep36826

45.

Strobel

, Miller

, Rist

, et al. Comparative analysis of cesium chloride- and iodixanol-based purification of recombinant adeno-associated viral vectors for preclinical applications. Hum Gene Ther Methods, 2015; 26(4):147–157; doi: 10.1089/hgtb.2015.051

46.

Zhao

, Keating

, Dolman

, et al. Characterization of complete particles (VSV-G/SIN-GFP) and empty particles (VSV-G/EMPTY) in human immunodeficiency virus type 1-based lentiviral products for gene therapy: Potential applications for improvement of product quality and safety. Hum Gene Ther, 2008; 19(5):475–486; doi: 10.1089/hum.2007.119

47.

Wang

, Godfrey

, Ravikrishnan

, et al. Shear contributions to cell culture performance and product recovery in ATF and TFF perfusion systems. J Biotechnol, 2017; 246:52–60; doi: 10.1016/j.jbiotec.2017.01.020

48.

Wang

, Godfrey

, Radoniqi

, et al. Larger pore size hollow fiber membranes as a solution to the product retention issue in filtration-based perfusion bioreactors. Biotechnol J, 2019; 14(2):e1800137; doi: 10.1002/biot.201800137

49.

Gao

, Li

, Zhong

, et al. Empty virions in AAV8 vector preparations reduce transduction efficiency and may cause total viral particle dose-limiting side-effects. Mol Ther Methods Clin Dev, 2014; 1(9):20139; doi: 10.1038/mtm.2013.9