Optimizing T cell transduction: a novel transduction device for efficient and scalable gene delivery

Abstract

Background

Viral transduction is a critical step in the manufacturing of genetically modified T cells for immunotherapies, yet conventional transduction methods suffer from low to medium efficiency, high vector consumption, and limited scalability.

Methods

To address these challenges, we introduce the Transduction Boosting Device (TransB), an innovative, automated, and closed-system platform designed to enable efficient and scalable gene delivery and overcome the limitations of conventional transduction methods. TransB improves cell-virus interactions by facilitating proximity between target cells and viral vectors.

Results

TransB demonstrated up to 1-fold decrease in processing time, 3-fold reduction in viral vector consumption, and 0.7-fold increase in transduction efficiency compared to 24—well plate method for donor T cell transduction in studies evaluating its impact on transduction process. Comparison studies transducing T cells from three different donors with Lenti-GFP vectors showed that TransB achieved an average 0.5-fold improvement in transduction efficiencies while maintaining comparable post-transduction cell recovery, viability, growth, and phenotype compared to 24—well plate. Furthermore, TransB delivered consistent performance across two different input cell numbers demonstrating scalability of the process.

Conclusion

These findings suggest that TransB could significantly shorten the transduction time, reduce the transduction cost and improve the transduction efficiency for manufacturing genetically modified T cell therapies. It shows strong potential as a robust, efficient, and scalable platform to enhance T cell therapy manufacturing and help overcome current manufacturing challenges in the field.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12967-025-06836-1.

Introduction

T cell-based immunotherapies, such as CAR—T cell therapy, have revolutionized cancer treatment, demonstrating remarkable efficacy in targeting various hematologic malignancies and solid tumors [1]. A pivotal step in the manufacture of these therapies is the genetic modification of T cells to express therapeutic genes that enhance T cells’ ability to identify and eliminate aberrant cells. The transgene delivery in this process is commonly achieved through viral transduction, employing lentiviruses or retroviruses as delivery vectors [2, 3]. Despite its critical role for T cell-based immunotherapy manufacturing, current viral transduction methods face significant challenges, including low to medium transduction efficiency, high operational costs, and scalability constraints [4, 5].

Conventional transduction methods rely on static incubation in traditional cultureware and involve manual manipulation with open culture systems, introducing process variability and contamination risks. Additionally, these methods suffer from inefficient cell—viral vector interaction, leading to suboptimal transduction efficiency, particularly when dealing with hard-to-transduce cells. These limitations hinder large-scale manufacturing where both consistency and efficiency are paramount, hampering the wide clinical availability of T cell—based therapies [6, 7]. Emerging technologies such as microfluidic platforms have been explored to improve viral transduction efficiency [8–11]. While these systems show promise in research settings, they often face process scalability limitations, restricting their translation to therapeutic manufacturing. Another strategy, spinoculation, which utilizes centrifugal forces, has also shown improved transduction efficiency and been commonly implemented with Sepax C—Pro, a Good Manufacturing Practice (GMP)-compliant platform [12]. However, the process scalability of Sepax C—Pro is limited due to its fixed processing volumes, posing challenges for its application in large—scale workflows.

In this study, we present the Transduction Boosting Device (TransB), a novel viral transduction platform designed to address the limitations of current transduction methods for T cell therapy manufacturing. TransB leverages the high surface area-to-volume (SA: V) ratio of hollow fibres to create an optimized microenvironment that enhances T cell-viral vector interactions and boost transduction efficiency. Our study demonstrated that cells transduced with TransB not only achieved comparable cell recovery rates, viability and growth post transduction, but also significantly enhanced transduction efficiency at various processing cell numbers and across multiple donor T cells compared to those transduced with conventional 24—well plate-based methods. Our findings highlight the potential of TransB as an efficient and scalable viral transduction platform for both T cell therapy research and manufacturing, supporting the growing demand for T cell-based immunotherapies.

Materials and methods

Production of lentiviral vectors

9 × 10⁶ 293T cells (CRL-3216, ATCC, Manassas, Virginia, USA) were seeded in a T-75 flask with 9 ml complete Dulbecco’s Modified Eagle Medium (DMEM) (90% DMEM (Thermo Fisher Scientific, Waltham MA, USA) + 10% Fetal Bovine Serum (FBS) (Hyclone, Logan, UT, USA)) and incubated at 37 °C, 5% CO₂ for 6 h to allow for cell adherence to the surface. 93 µl of FuGENE 6 (Promega, Madison, WI, USA) was added into 1.02 ml DMEM (Thermo Fisher Scientific, Waltham MA, USA) and incubated at room temperature for 5 min. 13.5 µg of pLenti-CMV-GFP-Puro (#17448, Addgene, Watertown, MA, USA), 9 µg of pMDLg/pRRE (#12251, Addgene), 4.8 ug of pCMV-VSV-g (#8454, Addgene) and 3.3 µg of pRSV-Rev (#12253, Addgene) were then added into the diluted FuGENE 6 and incubated for 15–30 min. Culture media from the 293T culture flask was then removed and substituted with 9 ml fresh DMEM (Thermo Fisher Scientific). The mixture of DNA plasmids and FuGENE 6 were then added directly into the culture flask and the cells were incubated at 37 °C, 5% CO₂ for 18–20 h. Full media change was performed, and cells were incubated for another 24 h. The lentivirus contained supernatant was collected and filtered through 0.45 μm PVDF Filter (Sartorius AG, Göttingen. Germany). It was then concentrated 10 times with a hollow fiber (C02-E500-05-S, Repligen) following the manufacturer’s protocol and stored at -80 °C for future use.

T cell preparation for transduction

Donor Peripheral Blood Mononuclear Cells (PBMCs) (STEMCELL technologies, Vancouver, Canada) were thawed and activated with ImmunoCult™ Human CD3/CD28/CD2 T Cell Activator (STEMCELL Technologies) (25 µl/ml of cells with a concentration of 1 × 10⁶ cells/ml) and IL-2 (STEMCELL Technologies) (50 IU/ml) and cultured for 3 days with complete culture medium consisting of RPMI-1640 (Thermo Fisher Scientific), 10% heat inactivated FBS (Hyclone, Logan, UT, USA) and 2 mM L-glutamine (Thermo Fisher Scientific). The activated PBMCs were used for subsequent transduction experiments.

Cell transduction in 24 well plate

On Day 0, donor PBMCs that were 3 days post activation (1.32 × 10⁶ cells unless otherwise specified) were premixed with lentiviral vector at specified Multiplicity Of Infection (MOI). The MOI was defined as virus volume-to-cell volume ratio. 500 µl cell-viral mixture was seeded into each well of 24—well plate and incubated at 37 °C, 5% CO₂ for a specified duration for transduction.

Following transduction, the culture medium was collected and centrifuged at 300 × g for 5 min to remove supernatant on Day 1. The pelleted cells were then reseeded into 24—well plate at 1 × 10⁶ cells/mL and 2 mL/well in complete culture medium and cultured for an additional 3 days before transduction efficiency was assessed on Day 4.

Cell culture in 24 well plate control without virus

On Day 0, donor PBMCs that were 3 days post activation (1.32 × 10⁶ cells unless otherwise specified) were seeded into each well of 24—well plate at 1 × 10⁶ cells /mL and 2 mL/well in complete culture medium and incubated at 37 °C, 5% CO₂ for a specified duration for simulated transduction.

Following simulated transduction, the culture medium was collected and centrifuged at 300 × g for 5 min to remove supernatant on Day 1. The pelleted cells were then reseeded into 24—well plate at 1 × 10⁶ cells /mL and 2 mL/well in complete culture medium and cultured for an additional 3 days before analysis on Day 4.

Cell transduction in TransB

On Day 0, donor PBMCs that were 3 days post activation (1.32 × 10⁶ cells unless otherwise specified) were premixed with viral vector at defined MOI. The MOI was defined as virus volume-to-cell volume ratio. 200 µl (unless otherwise specified) cell-virus mixture was introduced into the intracapillary (IC) space of the hollow fiber of the TransB. The loaded hollow fibre was then incubated at 37 °C, 5% CO₂ in the incubator for a specified duration for transduction. During transduction, the pump system of TransB was positioned outside the incubator while continuously perfusing IL-2-supplemented complete culture medium through the extracapillary (EC) space of the hollow fibre at a flow rate of 0.1 mL/min.

After transduction, cells were harvested by flushing the IC space with 4 mL of complete culture medium at a flow rate of 13 mL/min, while simultaneously flushing the EC space at 6 mL/min for 1 min on Day 1. The harvested medium was then centrifuged at 300 × g for 5 min to remove supernatant. The pelleted cells were seeded into 24—well plate at 1 × 10⁶ cells/mL and 2 mL/well in complete culture medium and cultured for an additional 3 days before transduction efficiency was assessed on Day 4.

Culture analysis

Cell viability and numbers were analysed before transduction on Day 0, after transduction on Day 1 and after cell expansion post transduction on Day 4. Cell phenotype, GFP expression, and Vector Copy Number (VCN) per transduced cell were analysed after cell expansion post transduction on Day 4.

Cell count and live cell recovery analysis

Cell count was performed using Countess Automated Cell Counter (Thermo Fisher Scientific) and live cell recovery rate was calculated based on the formula:

Transduction efficiency and cell phenotype analysis

Cells were stained with Viobility 405/452 Fixable Dye (130-130-404, Miltenyi Biotec, Bergisch Gladbach, Germany) and CD3-APC (130-113-135, Miltenyi Biotec) to assess viability and CD3⁺ T cells. Cells were stained with CD3-VioBlue (130-114-519, Miltenyi Biotec), CD8-VioBlue (130-110-683, Miltenyi Biotec), CD4-APC-Vio770 (130-113-223, Miltenyi Biotec), CCR7-PE (130-120-463, Miltenyi Biotec), CD45RA-PerCP-Vio770 (130-113-368, Miltenyi Biotec) for phenotype analysis. MACSQuant Analyzer 10 (Miltenyi Biotec) was used to acquire the data and MACSQuantify 2.13.3 software (Miltenyi Biotec) was used for data analyses. Transduction efficiency of cell population was calculated based on the formula:

Quantitative real time PCR (qPCR) and VCN per transduced cell

Cells were incubated 3 days post transduction and genomic DNA (gDNA) was extracted from 3 × 10⁶ cells with DNeasy Blood and Tissue Kit (69506, QIAGEN, Hilden, Germany). 50 ng of the gDNA was used per qPCR reaction. eGFP was used as the vector gene target and RPL32 was used as the reference target.

eGFP Forward Primer: 5’– ACGTAAACGGCCACAAGTTC– 3’.

eGFP Reverse Primer: 5’– AAGTCGTGCTGCTTCATGTG– 3’.

RPL32 Forward Primer: 5’– CAAGGAAAGACGAGCTGTAGG– 3’.

RPL32 Reverse Primer: 5’– GGGCAGTTGCATCTTCATATTC– 3’.

iTaq Universal SYBR Green Supermix (1725121, Bio-Rad, Hercules, California, USA) was used for qPCR amplification, and the amplification and quantification was done using the Bio-Rad CFX96 thermocycler (Bio-Rad) with the procedures of: an initial denaturation step at 95 °C for 5 min, followed by 36 cycles of denaturation at 95 °C for 5 s and annealing/extension at 60 °C for 30 s. The VCN per cell and VCN per transduced cell were calculated using the following formulas [13]:

Statistical analysis

Paired student’s two-tailed t-tests were conducted with Microsoft Excel to analyze the statistical significance of difference between the means of two comparison groups. The significance levels were indicated in the plots with “ns” representing no significant difference (p > 0.05) and asterisks * indicating significant difference (p < 0.05). *: p < 0.05; **: p < 0.01; ***:p < 0.001; ****:p < 0.0001.

Results

Development of TransB

The core of the TransB system features a single-use module incorporating a hollow fiber membrane (C02-E500-05-S, Repligen, Waltham, Massachusetts, USA) with a molecular weight cutoff of 500 kD, a coiled silicone gas exchange, and a gas-permeable silicone tubing set. During operation, cells and viral vectors are confined within the IC space of the hollow fiber, which possesses a high SA: V ratio. This design enhances cell-vector contact and thus promotes more efficient transduction (Fig. 1A). Importantly, the flexible modular design of TransB enables easy interchange of hollow fiber component with varying volumes, thereby accommodating different process scales.

Fig. 1.

Schematic of the operation of the TransB. (A) Loading stage: T cells (green) and viral vectors (red) are introduced into the intracapillary (IC) of the hollow fiber membrane. (B) Perfusion/transduction stage: hollow fiber membrane is incubated inside an incubator and culture medium is continuously perfused through the extracapillary (EC). Gas exchange between the hollow fiber and the environment occurs via passive diffusion through gas-permeable silicone tubing, which maintains optimal oxygenation and pH for cell culture. The perfusion loop includes sterile medium and waste bottles, a peristaltic pump for controlled medium flow, and a single-use module that features the hollow fiber component, an gas exchanger and a tubing set. (C) Harvesting stage: Fresh medium is flushed through the IC and EC of the hollow fiber to gently dislodge and push the transduced cells to the outlet of the hollow fiber for collection.

The gas exchanger is located between the hollow fiber and the medium reservoir, enabling efficient oxygen and carbon dioxide delivery through passive diffusion. The tubing set links the side ports of the hollow fiber to a medium reservoir and a waste collection bottle.

A peristaltic pump continuously drives medium through the EC space to provide nutrients and support gas exchange to the cultured cells in the IC. This design maintains a stable culture environment throughout the transduction process without manual intervention (Fig. 1B). The perfusion flow rate and duration are adjustable with the pump, allowing process customization based on cell type, viral vector, or desired transduction kinetics.

Operation of TransB

Prior to use, the single-use module of TransB was assembled aseptically inside a biosafety cabinet and then mounted onto the pumping system. The transduction process with TransB then took place following three sequential stages: loading, perfusion/transduction, and harvesting.

During the loading stage, a premix of T cells and Lenti-GFP vectors, adjusted to match the IC volume of the hollow fiber, was gently introduced into the IC via the inlet port (Fig. 1A). Special care was taken to avoid overfilling or spillage, ensuring efficient viral utilization and containment.

During the perfusion/transduction stage, the inlet and outlet ports of the hollow fiber were sealed, and the side ports were connected to the tubing set to establish a perfusion loop (Fig. 1B). The hollow fiber assembly containing the cell-virus mixture was incubated under standard culture conditions (37 °C, 5% CO₂), while the perfusion tubing remained engaged with a peristaltic pump positioned outside the incubator. Fresh medium was continuously pumped and perfused through the EC at 0.1 mL/min, a flow rate empirically optimized to support sufficient nutrient and gas supply without imposing detrimental shear stress onto the cells.

Following the transduction, media perfusion was halted, and the process transitioned to the harvesting stage. The IC and EC were simultaneously perfused with fresh medium to gently dislodge and push transduced cells towards the outlet of the hollow fiber (Fig. 1C). Optimal flow rates of 13 mL/min for the IC and 6 mL/min for the EC, were empirically established to achieve nearly complete (> 95%) recovery of viable, transduced cells.

TransB achieved comparable live cell recovery without causing adverse effects on T cell viability or cell growth post transduction

To isolate the impact of TransB’s physical environment from that of viral transduction, T cells were firstly incubated without virus in TransB. Two initial seeding cell numbers were tested: Low Seeding (LS): 6.6 × 10⁵ T cells, and High Seeding (HS): 1.32 × 10⁶ T cells. After 18 h of incubation, cells were harvested and analyzed for live cell recovery and viability.

The average live cell recovery rate in the TransB group was 15% lower than that in the 24—well plate control at LS (119.4% in TransB versus 134.6% in 24—well plate), but was similar at HS (125.9% in TransB versus 127.0% in 24—well plate) (Fig. 2A). The difference between the average live cell recovery rates of the two groups were statistically insignificant at both seeding cell numbers (Fig. 2A). Interestingly, while no statistically significant changes in liver cell recovery rates were observed when seeding cell numbers was increased from low to high for both study groups, TransB presented a 5% increase in average cell recovery, whereas 24—well plate showed a 6% decrease (Fig. 2A). TransB thus might potentially carry cell recovery advantages when processing higher cell numbers compared to 24—well plate, though further studies are needed for verification.

Fig. 2.

TransB supported comparable live cell recovery, cell growth, and viability to 24—well plate with and without viral transduction. (A—C) Comparison of donor T cells incubated for 18 h without viral transduction in TransB (labelled as TransB) or 24—well plate (labelled as 24-wp), showing statistically comparable (A) live cell recovery rates, (B) cell expansion, and (C) cell viability across two seeding cell numbers. LS: low seeding (6.6 × 10⁵ cells); HS: high seeding (1.32 × 10⁶ cells). (D—F) Comparison of donor T cells (1.32 × 10⁶ cells) transduced with Lenti-GFP at MOI 0.5 for 18 h in TransB (labelled as TransB) and 24—well plate (labelled as 24-wp) showing comparable (D) live cell recovery rates, (E) cell expansion, and (F) cell viability. For (D—F), a study group incubating T cells (1.32 × 10⁶ cells) in 24-well plate for 18 h without viral exposure was also included for comparison (labelled as 24-wp(Ctrl)). n = 2 for TransB group with HS in (A—C). n > = 4 for all other study groups. Mean values with standard deviations were presented for plots. ns: statistically insignificant difference.

To assess whether exposure to the TransB system affected downstream cell growth, cells harvested after an 18-hour incubation were further cultured for 3 days in IL-2-supplemented medium. The results showed statistically insignificant differences in cell growth or viability at both seeding cell numbers comparing TransB with 24—well plate, suggesting that TransB did not compromise the cells’ proliferative capacity or viability (Fig. 2B and C).

After confirming that exposure to the physical environment of TransB did not adversely affect key cellular attributes, we next evaluated the impact of transduction with TransB on live cell recovery rate, cell growth and viability with MOI 0.5, transduction duration of 18 h and HS benchmarking against 24—well plate. Cells transduced in TransB achieved statistically comparable live cell recovery rate compared to those processed in 24—well plate with or without transduction, despite some differences (on average 114% in TransB with transduction, 120% in 24—well plate with transduction, and 134% in 24—well plate without transduction) (Fig. 2D). 3 days post transduction, cells from the TransB group reached a population that was statistically comparable to those from 24—well plate with or without transduction, despite some differences (on average 3.9 × 10⁶ cells in TransB with transduction, 3.7 × 10⁶ cells in 24—well plate with transduction, and 4.47 × 10⁶ cells in 24—well plate without transduction) (Fig. 2E). Cell viability was between 87% and 89% for all groups with no statistically difference (Fig. 2F). The comparable results between the study groups showed that transduction with TransB can be performed without negatively affecting the cell recovery rate, expansion potential and viability compared to 24—well plate method.

TransB enhanced the transduction efficiency while reducing processing time and viral vector usage

To evaluate the impact of process parameters for lentiviral transduction with the TransB, we conducted a series of experiments transducing donor T cells with Lenti-GFP across varying incubation durations (2, 4, and 18 h), MOIs (0.5, 1, and 2), and cell input numbers (LS: 6.6 × 10⁵ cells and HS: 1.32 × 10⁶ cells) with transduction in 24—well plate as the benchmark. Results were encouraging showing that TransB achieved 0.3- to 0.9-fold improvement in transduction efficiency compared to 24—well plate (Fig. 3A—F), corresponding to 30–90% increase in transduced cell yield.

Fig. 3.

TransB enhanced lentiviral transduction efficiency of activated T cells compared to 24-well plate under various process conditions. (A) Transduction efficiency of T cells following transduction for 2, 4, and 18 h (MOI = 0.5). (B) Transduction efficiency ratio between TransB and 24—well plate from (A). (C) Transduction efficiency at different MOIs (MOIs: 0.5, 1, and 2) (duration = 18 h). (D) Transduction efficiency ratio between TransB and 24—well plate from (C). For (A—D), seeding cell number was 1.32 × 10⁶ cells. (E) Transduction efficiency at two seeding numbers: LS (6.6 × 10⁵ cells) and HS (1.32 × 10⁶ cells) at MOI of 0.5 and duration of 18 h. (F) Transduction efficiency ratio between TransB and 24—well plate from (E). TransB: transduction using TransB. 24-wp: transduction using 24—well plate. n = 2 for TransB group with MOI 2 in (C) and (D). n = 4 for all other study groups. Mean values with standard deviations were presented for plots. ns: statistically insignificant difference and p > 0.05. Asterisks indicate statistically significant difference. *: p < 0.05; **: p < 0.01; ***: p < 0.001; ****: p < 0.0001.

In both TransB and 24—well plate systems, transduction efficiency significantly increased with longer incubation durations (Fig. 3A). However, TransB consistently outperformed 24—well plate significantly, delivering a 0.24—0.48 fold improvement in transduction efficiency across all durations (Fig. 3A and B). Notably, TransB achieved an average 14.7% transduction efficiency in 2 h, which is statistically comparable to the average 16.3% efficiency by 24—well plate in 4 h (Fig. 3A), representing 1-fold reduction in processing time.

TransB demonstrated significantly enhanced transduction efficiency at MOI 0.5 and MOI 1, but not MOI 2, achieving an average of 15.7% increase at MOI 0.5, 20.7% increase at MOI 1, and 14.5% increase at MOI 2 specifically (Fig. 3C). It achieved the highest efficiency fold gains at lower MOIs: average of 0.69-fold at MOI 0.5 and MOI 1, and 0.4-fold at MOI 2 (Fig. 3D). Remarkably, transduction at MOI 0.5 in TransB achieved statistically comparable efficiency to that at MOI 2 in 24—well plate (39.4% and 36.1% respectively), representing a 3-fold efficiency improvement with the same amount of virus or a 3-fold reduction in virus consumption by TransB.

Finally, TransB showed statistically significant enhancement to transduction efficiency compared to 24—well plate regardless of the input cell numbers (Fig. 3E). It was interesting to see that the fold increase in transduction efficiency was cell number—dependent with higher fold increase at higher cell number (average 0.54-fold increase at LS and 0.7-fold increase at HS) (Fig. 3F). The advantage of TransB over 24-well plate method may therefore become more pronounced at higher processing cell numbers.

The live cell recovery, cell expansion and viability post transduction under each of the tested transduction condition were comparable between TransB and 24—well plate groups (Supplementary Figure S1).

TransB consistently enhanced transduction efficiency of T cells across multiple donors

To assess the robustness of TransB and related transduction protocol, we evaluated the performance of TransB for transducing T cells from three different donors. Each donor’s cells were activated for 3 days, and then transduced for 18 h at MOI 1, using a seeding cell number of 1.32 × 10⁶ cells. We compared live cell recovery rates, post-transduction cell expansion, viability, phenotype, transduction efficiency and VCN per transduced cell across three experimental groups: (1) transduction in the TransB, (2) transduction in 24—well plate, and (3) incubation in 24—well plate without viral vector exposure.

There was no statistically significant difference in live cell recovery rate or post-transduction expansion between any of the groups (Fig. 4A and B), confirming that the TransB did not induce additional cell stress or cytotoxicity during T cell transduction compared to 24—well plate. Cell viabilities in all study groups were above 80%. Interestingly, cells transduced by TransB exhibited an average of 5% higher viability after 3 days’ culture compared to their counterparts from the other two study groups. This difference was small but statistically significant (Fig. 4C). Cells transduced with the TransB comprised an average of 95% CD3⁺ T lymphocytes, which was about 2% higher than those from 24—well plate transduction or 24—well plate incubation. Again, this difference was small but significantly significant (Fig. 4D).

Fig. 4.

The TransB transduction system consistently enhanced the transduction efficiency across different donor PBMCs. (A) Live cell recovery rates after the harvest step at the end of transduction. (B) Cell expansion over 3 days post transduction. (C) Cell viability over 3 days post transduction. (D) Percentage of CD3⁺ T lymphocytes 3 days post transduction. (E) CD4⁺ and CD8⁺ cell subpopulation percentage in CD3 + T lymphocytes 3 days post transduction. (F) Memory phenotype subpopulation percentage in CD3⁺ T lymphocytes 3 days post transduction and gating strategy for each subpopulation (T_N: CCR7⁺CD45RA⁺, T_CM: CCR7⁺CD45RA⁻, T_EM: CCR7⁻ CD45RA⁻; T_EMRA: CCR7⁻CD45RA⁺). (G) Transduction efficiency (%GFP⁺) of CD3⁺, CD4⁺ and CD8⁺ T lymphocytes 3 days post transduction. (H) Transduction efficiency ratio between TransB and 24—well plate for CD3⁺, CD4⁺ and CD8⁺ cells from (G). (I) Transduction efficiency (%GFP⁺) of T_N, T_CM, T_EM and T_EMRA 3 days post transduction. (J) Transduction efficiency ratio between TransB and 24—well plate for memory subpopulations from (I). (K) VCN per transduced cell 3 days post transduction. TransB: transduction using TransB. 24-wp: transduction using 24—well plate. 24-wp(Ctrl): cell incubation using 24—well plate without viral vector exposure. n = 3. Average values with standard deviations were presented for plots except (E) and (F). ns: statistically insignificant difference and p > 0.05. Asterisks indicate statistically significant difference. *: p < 0.05; **: p < 0.01; ***: p < 0.001; ****: p < 0.0001.

Given the donor-to-donor variance, CD4⁺, CD8⁺ and memory subpopulations of the cells before and after transduction were compared for cells from individual donors (Fig. 4E and F). In all study groups, a higher initial CD4⁺ subpopulation ratio corresponded to a higher post-processing CD4⁺ ratio, and vice versa for the CD8⁺ population, indicating that the relative CD4⁺ to CD8⁺ ratio was preserved throughout transduction with TransB (Fig. 4E). Phenotypic analysis using CCR7 and CD45RA markers showed all study groups resulted in reduced naïve T cells (T_N: CCR7⁺CD45RA⁺) and terminally differentiated effector memory T cells (T_EMRA: CCR7⁻CD45RA⁺) across donor cells. TransB resulted in lower population percentage of central memory T cells (T_CM: CCR7⁺CD45RA⁻) and higher lower population percentage of effector memory T cells (T_EM: CCR7⁻CD45RA⁻) compared to 24—well plate transduction and 24—well plate incubation for all donor cells. We hypothesize that continuous perfusion of IL-2-supplemented media in the TransB system may have influenced the phenotypic distribution of T cell memory subsets. Further investigation will help elucidate the underlying mechanisms and assess the statistical significance of the difference.

TransB enhanced overall transduction efficiency of CD3⁺ T lymphocytes by an average of 16.7% compared to 24—well plate across all donor cells (Fig. 4G). This improvement was consistent across CD4⁺ and CD8⁺ subpopulations (Fig. 4G). Such differences were statistically significant (Fig. 4G) and corresponded to 0.49-fold improvement for CD3⁺ cells, 0.43-fold improvement for CD4⁺ cells, and 0.63-fold improvement for CD8⁺ cells on average respectively (Fig. 4H). Furthermore, a detailed analysis of the different memory cell phenotypes revealed that TransB enhanced transduction efficiency across all the memory subpopulations significantly except terminally differentiated effector memory T cells (T_EMRA: CCR7⁻CD45RA⁺) (Fig. 4I and J).

Lastly, we assessed VCN per transduced cell to check if it remained within safety parameters (FDA recommended threshold: VCN < 5 [14]). While VCN levels were elevated in the TransB group significantly compared to 24—well plate transduction, they remained well below the safety threshold, supporting the system’s suitability for clinical translation (Fig. 4K).

TransB demonstrates potential for scalable cell transduction

To evaluate the process scalability of TransB, we compared transduction performance between TransB (L), a system incorporating a longer hollow fiber membrane (C04-E500-05-S), and TransB (S), the system incorporating a shorter hollow fiber membrane which was used in previous experiments (Fig. 5A) for donor T cell transduction at MOI 1 and transduction duration of 18 h. TransB (L) has an IC volume of 415 µl, compared to 200 µl in TransB (S). To match the IC volume, 2.73 × 10⁶ cells were used with TransB (L) compared to 1.32 × 10⁶ cells for TransB (S).

Fig. 5.

The TransB supported efficient and consistent T cell transduction efficiency at varying process scales. (A) Different hollow fibres for TransB (L) and TransB (S). (B) Cell recovery rates post transduction. (C) Cell growth tracking post transduction. (D) Cell viability tracking post transduction. (E) Percentage of CD3⁺ T cells 3 days post transduction. (F) CD4⁺ and CD8⁺ cell subpopulation percentage in CD3⁺ T lymphocytes 3 days post transduction. (G) Memory phenotype subpopulation percentage in CD3⁺ T lymphocytes 3 days post transduction (T_N: CCR7⁺CD45RA⁺, T_CM: CCR7⁺CD45RA⁻, T_EM: CCR7⁻CD45RA⁻, T_EMRA: CCR7⁻CD45RA⁺). (H) Transduction efficiency (%GFP⁺) of CD3⁺, CD4⁺ and CD8⁺ T cells respectively 3 days post transduction. (I) Transduction efficiency ratio between TransB (L) and TransB (S) for CD3⁺, CD4⁺ and CD8⁺ cells from (H). (J) Transduction efficiency (%GFP⁺) of T_N, T_CM, T_EM and T_EMRA respectively 3 days post transduction. (K) Transduction efficiency ratio between TransB(L) and TransB(S) for memory subsets from (J). (L) VCN per transduced cell 3 days post transduction. n = 3. Mean values with standard deviations were presented for plots. ns: statistically insignificant difference and p > 0.05.

Analysis showed insignificant differences in live cell recovery rates, cell expansion potential, or viability post transduction between TransB (S) and TransB (L) (Fig. 5B—D). Additionally, the distribution of CD3⁺, CD4⁺, CD8⁺, and memory subpopulations (T_N: CCR7⁺CD45RA⁺, T_CM: CCR7⁺CD45RA⁻, T_EM: CCR7⁻CD45RA⁻; T_EMRA: CCR7⁻ CD45RA⁺) remained similar between the two systems across all donor cells (Fig. 5E–G). Importantly, transduction efficiency was statistically comparable between TransB (S) and TransB (L) across all assessed T cell subpopulations (Fig. 5H—K). The two systems also resulted in statistically comparable VCN per transduced cell (Fig. 5L).

Discussion

TransB was developed as an advanced automated and closed cell transduction system with three key objectives: (1) achieving high transduction efficiency at both research and manufacturing scales, (2) minimizing viral vector consumption, and (3) shortening transduction time. The system was designed to help reduce T cell therapy manufacturing costs, shorten turnaround time and improve overall yield. In this study, we successfully developed a functional prototype and verified its performance.

Unlike conventional transduction platforms such as culture plates or bags, which often suffer from inefficient cell-viral vector interactions and consequently low to moderate transduction efficiency, TransB enhances these interactions by spatially co-locating cells and viral vectors. It also maintains a consistent cell-friendly transduction environment through continuous perfusion. This design helps improve transduction efficiency while minimizing cellular stress.

Our results demonstrate that TransB enabled efficient Lenti-GFP transduction of T cells without compromising cell recovery, viability, or growth. TransB substantially enhanced transduction efficiency across all tested transduction durations (2, 4 and 18 h), MOIs (0.5, 1 and 2), and input cell numbers (6.6 × 10⁵ cells and 1.32 × 10⁶ cells), achieving 0.3- to 0.9-fold efficiency gains over 24—well plate controls, which corresponded to a 30–90% increase in transduced cell yield. Notably, it reached equivalent transduction efficiency in half of the time (2 instead of 4 h) and reduced viral vector consumption by 3-fold (MOI 0.5 instead of 2) in some investigation studies, offering a compelling advantage for rapid and cost‑effective manufacturing.

The enhanced transduction efficiency with TransB was consistent across T cells from multiple donors, achieving an average of 0.5-fold improvement compared to 24—well plate. TransB also preserved subpopulation percentages of CD4⁺, CD8⁺ and memory cell types (T_N: CCR7⁺CD45RA⁺, T_CM: CCR7⁺CD45RA⁻, T_EM: CCR7⁻ CD45RA⁻, T_EMRA: CCR7⁻CD45RA⁺), and maintained safe VCN thresholds. This robustness to donor variability supports the system’s potential for broad clinical applications.

The observed improvements in transduction efficiency in TransB are likely attributable to the use of a hollow fiber component with a high SA: V ratio and continuous perfusion, which together enhance cell-viral vector interactions and maintain a stable transduction environment compared to 24—well plate method.

TransB also demonstrated consistent transduction efficiency enhancement across different cell input numbers (1.32 × 10⁶ and 2.73 × 10⁶ cells). This scalability is enabled by the system’ modular design, which allows for straightforward replacement of hollow fiber components with varying processing capacities, while maintaining critical process parameters such as flow dynamics, shear stress, and nutrient exchange.

The combination of enhanced process efficiency and scalability positions TransB as a promising gene delivery platform for both T cell therapy research and manufacturing. In the manufacturing setting, TransB’s closed-system design can help minimize contamination risks, its automation can improve process reproducibility, and its scalable architecture can enable seamless process transfer across different manufacturing scales.

While the field has seen considerable development effort toward more effective viral vectors or novel chemical enhancers, the use of these strategies in T cell therapy manufacturing is often constrained by challenges related to GMP—grade supply and concerns over biological safety or variability. In contrast, TransB employs a purely mechanical approach that introduces fewer changes and risks to the transduction process, simplifying implementation and integration into existing manufacturing workflows. Moreover, TransB can be used in conjunction with chemical enhancers to further boost transduction efficiency when appropriate. Importantly, TransB avoids the centrifugation-induced cell stress associated with spinoculation, potentially preserving cell health more effectively.

It should be noted that this study was not designed to optimize the transduction process using TransB. Instead, it aimed to investigate the impact of key process parameters on the transduction outcomes with TransB, and to provide an initial assessment of TransB’s potential to enhance transduction compared to conventional transduction methods. While the results highlight the promise of TransB as a gene delivery platform for manufacturing genetically modified cell therapies—offering potential benefits in scalability, efficiency, and cost-effectiveness—further process optimization tailored to specific T cell therapy products will be essential to fully unlock these advantages.

Future development

System and protocol customization for diverse cell types and viral vectors

While this study demonstrated significant improvement for Lenti-GFP transduction of donor T cells using TransB compared to 24—well plate, future work will extend its application to other immune cell types, such as gamma delta T cells, NK cells, and iPSCs, as well as other viral vectors, including adenoviruses, AAVs, and retroviruses. Further studies examining how vector properties, such as viral tropism and size of the inserted transgene, could influence transduction efficiency and cell functions with TransB will inform the development of more precise and effective transduction strategies for cell therapy manufacturing.

To achieve optimal performance, customization of both hollow fiber configurations and transduction protocols will be essential. Optimization of the hollow fiber membranes must be tailored to specific cell and viral vector types. Key parameters include material selection, pore size and molecular weight cutoff, all of which play important roles in regulating nutrient exchange and retention of cells and viral vectors. Fiber geometry such as diameter, wall thickness, and length, can also be adjusted to enhance molecule diffusion, reduce shear stress to cells and modulate cell-virus interaction. In parallel, flow rates and transduction protocols will need to be fine-tuned to maintain uniform perfusion and preserve cell viability across various cell types and material configurations.

Enhancement of perfusion and nutrient delivery

Optimization of perfusion parameters is critical for maximizing cell viability, transduction efficiency, and overall process yields. Future iterations of the TransB system may incorporate dynamic perfusion rates tailored to the metabolic needs of specific cell types. This could ensure consistent and efficient nutrient delivery during extended incubations, preventing undesired nutrient depletion and metabolic waste accumulation. The ability to tightly control nutrient and waste levels will be especially valuable when high cell numbers are processed for clinical manufacturing.

Real-time monitoring and adaptive process control

To further streamline the manufacturing process, the integration of real-time monitoring and automation tools represents a key area of development. Embedding sensors for critical parameters—such as cell density, viral vector concentration, pH, and dissolved oxygen—will enable continuous feedback and dynamic process adjustments. Additionally, the incorporation of machine learning algorithms to analyze real-time data and predict optimal transduction conditions could significantly enhance system performance and process consistency. This data-driven approach will support robust, scalable, and reproducible manufacturing of gene-modified cell therapies.

Integration with other manufacturing systems

As the demand for CAR-T and other cell-based therapies increases, seamless integration of TransB with downstream processes, such as cell expansion and formulation, will be vital. The closed-system design and modular scalability of TransB make it an ideal candidate for incorporation into fully automated, end-to-end manufacturing platforms, facilitating streamlined, contamination-free workflows from gene delivery to final product formulation.

Regulatory considerations

To support clinical translation and commercialization, future efforts will focus on aligning TransB with GMP requirements. This includes the standardization of operating protocols, intensive validation of system performance, and implementation of rigorous quality control measures. Achieving compliance with regulatory standards will be crucial for the broader adoption of TransB in clinical manufacturing.

Conclusion

TransB has been developed as an innovative solution to enhance transduction efficiency and scalability, reduce viral vector consumption, and shorten transduction time, addressing key bottlenecks in the current manufacturing of genetically modified cell therapies. Our studies demonstrate that TransB can significantly outperform conventional 24—well plate based method in lentiviral transduction of donor T cells, without compromising critical cellular attributes such as cell recovery, expansion, viability, or phenotype. Notably, TransB maintained consistent superior performance across a range of cell input numbers and multiple donor sources, underscoring its robustness and flexibility.

With its ability to improve process efficiency and scalability, reduce reagent consumption, shorten processing time and enable automated, closed-system manufacturing, TransB holds strong potential as a next-generation transduction platform poised to play a pivotal role in the advancement of immune cell therapy research and manufacturing.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

TransB

Transduction Boosting Device

GMP

Good Manufacturing Practice

SAV

Surface Area‑to‑Volume

DMEM

Dulbecco’s Modified Eagle Medium

FBS

Fatal Bovine Serum

PBMCs

Peripheral Blood Mononuclear Cells

MOI

Multiplicity of Infection

IC

Intracapillary

EC

Extracapillary

VCN

Vector Copy Number

qPCR

Quantitative real time PCR

gDNA

Genomic DNA

LS

Low Seeding

HS

High Seeding

T_N

Naïve T Cells

T_CM

Central Memory T Cells

T_EM

Effector Memory T Cells

T_EMRA

Terminally Differentiated Effector Memory T Cells

Author contributions

LKZ and TDN contributed equally to this work. LKZ and TDN performed the experiment. LD conceptualized the study. All authors performed data analysis and wrote the manuscript.

Funding

This work is funded by core fund from Bioprocessing Technology Institute (BTI), Agency for Science, Technology and Research (A*STAR), 20 Biopolis Way, #06 − 01 Centros, Singapore 138668, Republic of Singapore, and Singapore Therapeutics Development Review (STDR) Pre-Pilot Stream 2 grant (grant number: H24H0a0010).

Data availability

Data are available upon reasonable request. The authors declare that the data supporting the findings of this study are available upon reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.