A Biomaterial Platform for T Cell-Specific Gene Delivery

Abstract

Efficient T cell engineering is central to the success of CAR T cell therapy but involves multiple time-consuming manipulations, including T cell isolation, activation, and transduction. These steps add complexity and delay CAR T cell manufacturing, which takes a mean time of 4 weeks. To streamline T cell engineering, we strategically combine two critical engineering solutions - T cell-specific lentiviral vectors and macroporous scaffolds - that enable T cell activation and transduction in a simple, single step. The T cell-specific lentiviral vectors (referred to as STAT virus) target T cells through the display of an anti-CD3 antibody and the CD80 extracellular domain on their surface and provide robust T cell activation. Biocompatible macroporous scaffolds (referred to as Drydux) mediate robust transduction by providing effective interaction between naïve T cells and viral vectors. We show that when unstimulated peripheral blood mononuclear cells (PBMCs) are seeded together with STAT lentivirus on Drydux scaffolds, T cells are activated, selectively transduced, and reprogrammed in a single step. Further, we show that the Drydux platform seeded with PBMCs and STAT lentivirus generates tumor-specific functional CAR T cells. This potent combination of engineered lentivirus and biomaterial scaffold holds promise for an effective, simple, and safe avenue for in vitro and in vivo T cell engineering.

Graphical Abstract

1. Introduction

Adoptive cellular therapies, including Chimeric Antigen Receptor (CAR) T cell therapy, have the potential to revolutionize the treatment of cancer and other diseases. The CAR T cell therapy landscape has rapidly evolved over the last few years, with six CAR T cell products against hematologic malignancies approved since 2017 [1]. However, current CAR T cell products are costly ($350K-$450K) and demonstrate adverse effects, including cytokine release syndrome (seen in 77–93% of leukemia patients as well as 37–93% of lymphoma patients), cell-associated neurotoxicity syndrome (ICANS) (62–67% of leukemia and lymphoma patients) [2] and off-tumor, on-target toxicity (commonly demonstrated as B cell aplasia) [3]. In addition, CAR T cell therapy requires complex manufacturing infrastructure, and take multiple weeks to manufacture [4–6]. There is an unmet need to improve the safety, affordability, and access to these therapies to fully uncover their potential [3,7,8].

Current CAR T cell manufacturing involves T cell isolation, activation, genetic modification, and expansion [9]. Although T cell expansion appears to be dispensable [10], the earlier manufacturing steps are critical to formation of potent CAR T cell products as each step prepares the cell products for the next.

All current clinically approved CAR T cell products utilize viral vectors to attain stable, long-term genetic modification of T cells [11]. Lentiviral vectors enable delivery of large CAR constructs, reliable genetic modification, and stable long-term transgene integration at high transduction efficiencies [12,13]. Commonly used lentiviral vectors are often pseudotyped with Vesicular Stomatitis Virus G (VSVG) glycoprotein, which targets the common low density lipoprotein receptor (LDLR) and enable broad cell tropism [14].

In autologous CAR T cell manufacturing, leukocytes are isolated from a patient’s whole blood. Although only T cell modification is desired, T cells are often isolated alongside a variety of other cells, including macrophages, B cells, dendritic cells, NK cells, tumor cells, and others. The broad cell tropism demonstrated by VSVG-pseudotyped lentiviruses can be problematic in CAR T cell manufacturing due to the possibility of inadvertent tumor cell transduction [15]. To improve the safety of CAR T cell products, many manufacturing protocols include an additional step to select T cell populations before genetic modification. In addition, use of VSVG-pseudotyped viruses necessitates a T cell activation step in order to upregulate LDLR for efficient transgene delivery [16]. Current CAR T cell manufacturing protocols utilize anti-CD3 and anti-CD28 antibodies to achieve broad T cell activation prior to viral-based gene transfer. These additional steps of preselection and pre-activation add complexity, time, and labor to the overall CAR T cell manufacturing process. Dispensing with preselection and requisite activation would significantly speed up CAR T cell manufacturing .

An all-in-one platform that can integrate T cell activation and transduction into a single step and eliminate prior T cell purification needs would be highly desirable. Previously, various groups have reported such multifunctional viral platforms [17–21]. As an example, Dobson and colleagues demonstrated a lentiviral system capable of selective T cell infection along with simultaneous T cell activation owing to the display of an anti-CD3 antibody and the CD80 protein, alongside a mutated VSVG protein (VSVGmut) on the lentiviral surface [18]. In order to use such viral platforms efficiently, transduction-enhancing reagents (e.g. polybrene [22], RetroNectin [23], vectofusin-1 [24]) are often used to bring viral vectors and T cells into close contact to facilitate more efficient gene transfer. However, despite these reagents, transduction often remains modest and requires additional physical force via centrifugation to promote cell-virus colocalization. Simpler, one-step transduction protocols for improved gene transfer are needed to streamline CAR T cell production.

Biomaterials hold great promise in simplifying T cell engineering [25–27] by providing a favorable niche for cells and suitable architecture and environment for cell attachment, trafficking, and release [28]. We and others have reported biomaterial platforms to provide efficient and robust viral-mediated T cell engineering [25,29,30]. We have demonstrated that dry, macroporous biomaterial scaffolds provide efficient and simple transduction of pre-activated T cells by viral vectors and eliminate the need for both transduction-enhancing agents and sequential centrifugation [30,31].

In this report, we set out to combine two state-of-the-art engineering solutions to address critical challenges associated with T cell transduction. We hypothesized that the potent combination of 1) T cell-specific lentiviral vectors that cause simultaneous T cell activation and transduction (termed STAT lentivirus) through CD3 and CD28 signaling, and 2) macroporous scaffolds (termed Drydux) that enhance viral-mediated transduction of unstimulated T cells could significantly simplify the overall T cell engineering process. In this report, we show that when unstimulated peripheral blood mononuclear cells (PBMCs) are seeded together with UCHT1 (anti-CD3) + CD80 (a ligand for CD28) displaying STAT lentivirus on Drydux scaffolds, T cells are activated and selectively transduced in a single step. Further, we show that the Drydux platform seeded with PBMCs, and STAT lentivirus generates functional, tumor-specific CAR T cells. This combination of engineered virus and biomaterial platform holds promise for a simple, effective, and safe avenue for in vitro and in vivo T cell engineering.

2. Materials and methods

2.1. Synthesis of Drydux scaffolds:

Macroporous alginate scaffolds, termed ‘Drydux’ in this report were prepared using a method previously described [30,32]. Briefly, scaffolds were prepared in three steps. First, ultrapure alginate (Pronova) was dissolved in sterile filtered deionized water to make a 2% weight-by-volume solution. The gelation of alginate was carried out using an equal volume of 0.4% weight-by-volume stock solution of calcium gluconate to form a hydrogel. To cast the gels, 24 well plates (~13 mm diameter/well) were used as molds. Calcium crosslinked alginate gels were cast slowly in the desired well plate. After casting, gels were transferred to a −20°C freezer overnight. Once cryogels were formed, plates were prepared for drying in the lyophilizer for 72 hours. After 72 hours, spongy, macroporous scaffolds are ready for use. Plates were removed from the lyophilizer and stored in a vacuum sealed bag until further use.

2.2. Synthesis of targeting lentivirus

pHIV-ZsGreen transfer plasmid was purchased from Addgene (plasmid #18121) and pHIV-Myc-19BBz-EGFP was generated as reported previously [18]. pMD2.G and psPAX2.1 were gifted by Didier Trono (Addgene plasmids #12259 and #12260). pMD2-VSVGmut was generated as described previously [18] and is available on Addgene (plasmid #182229). To synthesize lentiviral supernatant, 80–90% confluent HEK-293T cells were used for plasmid transfection. For generating targeted lentiviruses, DNA plasmids were diluted in Opti-MEM using plasmid ratios of 5.6:3:3:1 for transfer plasmid (pHIV-ZsGreen or pHIV-Myc-19BBz-EGFP) to packaging plasmid (psPAX2.1) to targeting plasmid (50:50 mix of UCHT1 Fab + CD80) to fusogen plasmid (VSVGmut). For each T-225 flask, plasmid amounts of 42 ug transfer plasmid, 22.5 ug psPAX2.1, 7.5 ug pMD2-VSVGmut, 11.25 ug pMD2-UCHT1 Fab, 11.25 ug pMD2-CD80 were used. TransIT-Lenti Transfection Reagent (Mirus Bio, MIR-6606) was added at a 3:1 mass ratio of Mirus to DNA, and DNA complexes were allowed to form for 10 minutes at room temperature before adding dropwise to HEK-293T cells. Following 48–72 hours post-transfection, HEK-293T viral supernatant was collected. Supernatant was spun down at 1,000 xg for 10 minutes to pellet cell debris. Viral supernatant was then passed through a 0.45 um polyethersulfone syringe filter (Millipore Sigma) prior to ultracentrifugation at 100,000 xg for 90 minutes at 4C to pellet viral particles. Supernatant was then carefully aspirated off and the pellet was resuspended in ice cold serum-free Opti-MEM and kept at 4C overnight prior to aliquoting and long-term storage at −80C.

2.3. Lentiviral titer determination

In order to determine the viral titer for transduction experiments, Jurkat cells were plated in RPMI-1640 complete media at the density of 50,000 cells per 96 well. Viral titering dilutions were made in RPMI-1640 complete media and added to Jurkat T cells in a 50 uL volume. In order to enhance transduction, polybrene or diethylaminoethyl-dextran hydrochloride (Sigma-Aldrich, #D9885) was added to a final concentration of 8 ug/mL. Cells were kept in a humidified 5% CO2 incubator at 37C. Viral transduction was assessed after 48 hr by flow cytometry. Viral titers were calculated as follows:

$$\text{Cell}\hspace{0.17em}\text{number}\hspace{0.17em}*\left[\text{\%}\hspace{0.17em}\text{transduction}\left(\text{FITC+}\right)\right]/\left(\text{virus}\hspace{0.17em}\text{uL}\hspace{0.17em}\text{volume/1000}\right)=\text{Transducing}\hspace{0.17em}\text{units}\hspace{0.17em}(\text{TUs})\hspace{0.17em}\text{per}\hspace{0.17em}\text{mL}\hspace{0.17em}\text{of}\hspace{0.17em}\text{virus}$$

2.4. Drydux mediated T cell transduction

To transduce T cells using Drydux scaffolds, peripheral blood mononuclear cells (PBMCs) were isolated from the buffy coat fractions (Gulf Coast Regional Blood Center) of healthy donors using Lymphoprep density separation (Accurate Chemical and Scientific Corporation). A total of six different donors were used throughout the studies. Three donors were used to test ZsGreen lentiviruses while three different donors were used to test CAR 19 lentiviruses. On the same day, naïve PBMCs were mixed with concentrated lentivirus (ZsGreen encoding or anti-CD19 CAR encoding lentivirus) in ~100 μl to 150 µl of total volume and pipetted onto each dry macroporous scaffold. For all studies, a multiplicity of infection (MOI) = 1 was used unless otherwise noted. Control scaffolds were seeded with naive cells without lentivirus. Following seeding, excess T cell media consisting of 45% Click’s Medium (Irvine Scientific), 45% RPMI 1645 (Gibco), 10% Hyclone FBS (GE Healthcare), 2 mmol/L GlutaMax (Gibco), 100 U/ml penicillin (Gibco), and 100 mg/ml streptomycin (Gibco) was supplemented with 10 ng/ml of IL-7 (PeproTech) and 5 ng/ml of IL-15 (PeproTech) to support T cell proliferation. Seeded scaffolds were cultured for 48–72 hours. After 2–3 days of culture, scaffolds were digested with 0.25M EDTA (calcium chelator) (Gibco), washed twice with excess PBS, and cells were isolated. Isolated cells were analyzed for transduction efficiency using flow cytometry.

2.5. CAR T cell generation using conventional method

To transduce T cells using polybrene and spinoculation, freshly isolated PBMCs were mixed with concentrated lentivirus (anti-CD19 CAR encoding lentivirus) in presence of 8 μg/ml polybrene (sigma). Cell-virus mixture was plated, and plate was centrifuge at 1000g for 90 minutes at 32°C. Next day cells were washed to remove polybrene and cultured for additional 3–4 days and transduction efficiency was determined using flow cytometry.

2.6. T cell activation and proliferation assays

To study activation of human primary T cells following incubation with UCHT1 + CD80 lentivirus and Drydux scaffolds, freshly isolated T cells were used. Cells were mixed with either wild type lentivirus (VSVGwt), UCHT1 only lentivirus, CD80 only lentivirus or UCHT1 + CD80 lentivirus and seeded on the Drydux scaffolds. Cells without the virus were used as a negative control, while cells seeded on an anti-CD3 (Miltenyi Biotec) (1 μg/ml) and anti-CD28 (BD Biosciences) (1 μg/ml) coated plate were used as a positive control. At predetermined time points, cells were isolated from the scaffolds and stained for CD69 or CD25 to assess early-and late-stage T cell activation, respectively.

To study T cell proliferation, freshly isolated T cells were labeled with CellTrace Far Red (CTFR) proliferation dye (Invitrogen). Stained cells were then mixed with wild type lentivirus (VSVGwt), CD80 only lentivirus or UCHT1 + CD80 lentivirus and seeded on the Drydux scaffold. Cells without the virus were used as a negative control, while cells seeded on an anti-CD3 and anti-CD28 coated plate were used as a positive control. Seven days post transduction, cell proliferation (dilution of the CTFR dye) was determined using flow cytometry.

2.7. Selectivity and targeting of virus seeded on Drydux scaffolds

To analyze the ability of UCHT1 + CD80 lentivirus to selectively target CD3+ T cells, T cells were isolated from the buffy coat fraction and percentages of T cells (CD4+ and CD8+ population), monocytes, B cells, and natural killer (NK) cells were determined. 1x10⁶ PBMCs were mixed with ZsGreen lentiviruses at MOI 1 on the same day of isolation and incubated on Drydux scaffolds, 5 days after transduction, T cell fractions were determined using flow cytometry. To further confirm the ability of the UCHT1 + CD80 virus to selectively target CD3+ cells and ability of VSVG wild type (VSVGwt) virus to target broad cell types, Jurkat T cells and Ramos B cells were incubated with UCHT1 + CD80 or VSVGwt lentivirus on Drydux scaffolds. 5 days after incubation, transduction in Jurkat and Ramos cells were analyzed using flow cytometry.

2.8. Assessment of CAR T cell viability and proliferation

Viability and proliferation of transduced cells isolated from Drydux scaffolds were assessed. Cells were isolated from Drydux scaffolds and cultured in T cell media in the presence of cytokines (IL-7 + IL-15). Cells were stained with Trypan Blue to assess the viability and counted manually to determine the total cell number at the indicated days post-transduction.

2.9. Phenotypic assessment of CD19 targeting CAR T cells

T cells incubated with STAT lentivirus and controls on Drydux scaffolds were assessed for their immunophenotypic composition to determine the percentages of stem cell memory (SCM) (CCR7+, CD45RA+), central memory (CM) (CCR7+, CD45RA−), effector memory (EM) (CCR7−CD45RA−) and effector (Eff) (CCR7−CD45RA+) phenotypes. T cells isolated from Drydux scaffold were cultured for two days before assessment of presence memory markers using flow cytometry.

2.10. In vitro cytotoxicity of Drydux generated CAR T cells

To study the ability of CD19-targeting CAR T cells to kill CD19 positive tumor cells, Daudi cells (ATCC, CCL-213) were co-cultured with Drydux generated CAR T cells in vitro 5 days following transduction. Daudi cells were cultured in RPMI-1640 medium. Approximately 500,00 tumor cells were cocultured with CAR T cells at 1:1 and 1:5 effector to target (E:T) ratios with the number of CAR T cells normalized based on transduction efficiency except for in “no virus” controls, which used T cell numbers without any adjustments. On day 5 of co-culture, cells were collected, T cells were stained for CD3, and Daudi cells were stained for CD20. Dead cells were gated out using live/dead stain and percentages of residual tumor cells were assessed.

2.11. Cytokine production analysis by CAR T cells

CD19 targeting CAR-T cells generated using Drydux scaffolds were cocultured with Daudi tumor cells at 1:1 and 1:5 effector to target [E:T] ratio. 24 h following coculture, culture supernatant was collected. Presence of IL-2 and IFN-𝛾 cytokines were quantified by specific ELISA tests using the manufacturer’s protocol (R&D Systems).

2.12. Flow cytometry analysis

All samples were acquired on a BD LSRII using BD FACSDiva software or Cytoflex S flow cytometer, and a minimum of 10,000 events were acquired per sample. Samples were analyzed with FlowJo software (version 10.8.1). Zombie Aqua Dye (Biolegend, cat# NC0498216) was used to gate out dead cells. Monoclonal antibodies for CD3 (APC-Cy7, clone SK7, BD Biosciences, cat# 557832), CD4 (APC-Cy7, clone PRA-T4, BD Biosciences, cat# 561839), CD8 (PerCp-Cy5.5, clone SK1, BD Biosciences, cat# 565310), CCR7 (FITC, clone 150503, R&D systems, FAB197F-100), CD45RA (PE, clone HI100, BD Biosciences, cat# 555489), and CD20 (PE, clone 2H7, BD Biosciences, cat# 555623) were used.

2.13. Statistical analysis

Unpaired one-tailed or two-tailed Student’s t-test with Holm-Sidak correction for multiple comparison was used to compare and perform statistical analysis between two groups. For multiple comparisons, one way ANOVA with Tukey post hoc analysis was used. All the analysis was performed using GraphPad Prism software (version 9.4.1).

3. Results

Wild type VSVG-pseudotyped lentiviral vectors (VSVGwt) demonstrate broad cell tropism. Engineered lentiviral vectors express a mutated VSVG glycoprotein (VSVGmut) possessing two-point mutations (K47Q and R354A) that decouple their ability to interact with LDLR yet preserve their fusogen capacity [33]. When VSVGmut-pseudotyped lentiviral vectors are combined with additional targeting ligands on the viral surface they redirect tropism, thereby enabling cell type-specific transduction (Figure 1A). As previously demonstrated by Dobson et al [18], surface display of UCHT1 enables T cell-specific targeting, but do not efficiently transduce T cells. The combination of anti-CD3ε Fab clone (“UCHT1”) alongside human CD80 protein on the viral surface enables specific binding to CD3 and CD28 present on CD4+ and CD8+ T cells and provides critical signaling (signal 1 + costimulatory signal 2) necessary for robust T cell activation while enabling efficient viral transduction (Figure 1B). This targeted virus, capable of simultaneous T cell activation and transduction, is referred to as STAT lentivirus.

Figure 1: Overview of single step T cell transduction using STAT lentivirus and Drydux scaffold.

A) Schematics showing wild type lentivirus and targeted lentivirus with the ability to simultaneously activate and transduce T cells by displaying VSVGmut alongside targeting moieties like anti-CD3e Fab clone UCHT1 and CD80 protein on their surface. B) T cell specificity engendered by STAT lentivirus through providing simultaneous T cell activation and transduction. C) Single step Drydux-mediated preferential transduction of T cells from freshly isolated, unstimulated PBMCs by STAT lentivirus.

To facilitate T cell transduction without any additional transduction-enhancing reagents or spinoculation, we prepared 3-dimensional macroporous scaffolds using calcium-crosslinked alginate. The well-connected macroporous architecture and hygroscopic nature of these scaffolds force cell and virus colocalization, leading to static transduction in a simple, single step [29,34]. We prepared dry macroporous scaffolds in 24-well molds with approximately 14 mm diameter and 5 mm height for T cell engineering (Figure S1A). To use these scaffolds, along with STAT lentivirus as a T cell activation and transduction platform, unstimulated PBMCs isolated from the blood of healthy donors were mixed with concentrated STAT lentivirus and seeded onto dry Drydux scaffolds. 48–72 hours after incubation, transduced cells were isolated from the scaffold (Figure 1C, S1B).

To demonstrate the ability of STAT lentivirus to selectively target T cells from PBMC fractions within the scaffold, we examined percent of T cell population before and after transduction. Freshly isolated PBMCs were comprised of ~50% T cells (CD45+CD3+) (with ~ 55% of CD4+ population and ~37% of CD8+ population), 7% B cells (CD45+CD19+), 23% monocytes (CD45+CD14+), and 7% natural killer cells (CD45+CD56+). Seeding PBMCs with STAT lentivirus on Drydux enriched T cell populations to >95%, suggesting T cell-specific activation, proliferation, and selective growth (Figure 2A).

Figure 2: STAT viruses are T cell selective and show simultaneous T cell activation and transduction.

A) Specific targeting and enrichment of T cells by STAT lentivirus seeded on Drydux scaffolds. B) Ability of STAT lentivirus to specifically and selectively target CD3+ T cells. C) Assessment of expression of early activation marker (CD69) 20 hours after T cell incubation with STAT lentivirus or control lentiviruses on Drydux scaffolds (one-way ANOVA with Tukey’s correction). D) Assessment of expression of late activation marker (CD25) 5 days after T cell incubation with STAT lentivirus or control lentiviruses on Drydux scaffolds (one-way ANOVA with Tukey’s correction). E) Proliferation of T cells following incubation with STAT lentivirus or control lentiviruses on Drydux scaffolds. Cells incubated on antibody-coated plates without scaffolds were used as positive control. F) Drydux mediated transduction efficiency of ZsGreen-encoding STAT lentivirus seeded with freshly isolated PBMCs (*p<0.05 by two-tailed unpaired Student’s t-test).

To further confirm that STAT lentivirus specifically targets T cells owing to anti-CD3 display, we exposed Jurkat (CD3+ cells) or Ramos (CD3− cells) cells seeded on Drydux scaffold to either wild type (VSVGwt) or STAT lentivirus encoding the fluorescent protein ZsGreen. Since both Jurkat and Ramos are immortalized cell lines that constitutively express LDLR, Drydux with wild type lentivirus that shows broad cell tropism transduced both cell lines. Difference in the transduction levels following wild type transduction is likely due to varied LDLR density, cell activation state, [14,16,33], presence of restriction factors [35,36] or degree of transcriptional activity in the target cells [37]. On the other hand, STAT lentivirus transduced only CD3+ Jurkat cells (T cells), demonstrating the ability of STAT viruses to selectively target T cells (Figure 2B).

Efficient T cell transduction using wild type viruses requires T cells to be pre-activated, usually by signaling through CD3 and CD28. Since the STAT virus provides these signals to naïve T cells, we evaluated whether T cells co-seeded with STAT lentivirus on Drydux scaffolds demonstrated robust activation and proliferation. Drydux seeded with ZsGreen-encoding wild type virus, which lack T cell-targeting ligands served as a control. Drydux seeded with only cells, but no virus served as a negative control. Plates coated with anti-CD3 and anti-CD28 antibodies were used as a positive control for T cell activation. Targeted viruses displaying only anti-CD3 (UCHT1) or only CD80 (anti-CD28) were used to evaluate the independent effects of displaying activation (signal 1) and co-stimulation (signal 2) inputs on the lentiviral surface. Early activation was assessed by expression of CD69 while late-stage activation was assessed by expression of CD25 on the T cell surface. After 20 hours of cell-virus incubation on the Drydux scaffolds, T cells incubated with STAT lentiviruses showed 40% CD69+ expression in comparison to approximately 47% CD69+ with antibody-coated plates. All control groups (VSVGwt, only CD80, only anti-CD3, or cells-only) demonstrated marginal (5–6%) CD69+ expression, demonstrating the robust T cell activation by combined UCHT1 + CD80 signals displayed on the virus within the Drydux scaffolds. Moreover, approximately 5% CD69+ expression in cells incubated with viruses presenting only anti-CD3 or only CD80 demonstrated the importance of the simultaneous presentation of activation and co-stimulation signals within the Drydux scaffolds (Figure 2C). Assessment of late activation marker CD25 five days after seeding demonstrated that cells incubated with STAT lentivirus showed ~39% activation, somewhat below the standard positive control of plates coated with anti-CD3 and anti-CD28 antibodies (~69%), but above that of the CD25 signal seen with cells incubated with wild type virus (~1%), anti-CD3 only (~12%), or CD80 only (~1.3%) virus, demonstrating marginal activation in the absence of both activation (CD3) and co-stimulation (CD28) signals (Figure 2D).

Naïve T cell proliferation within scaffolds following incubation with STAT lentivirus or wild type virus was evaluated using the qualitative CellTrace Far Red dye cell proliferation as well as quantitative cell counting. Cells exposed to STAT lentivirus within scaffolds, as well those seeded onto coated plates, demonstrated robust proliferation, whereas cells incubated with wild type virus did not show a similar degree of proliferation (Figure 2E, S2), indicating that STAT lentivirus licenses T cells for both activation and cell division within the Drydux scaffold.

We next assessed the ability of STAT lentivirus to transduce naïve T cells within dry Drydux scaffolds. We seeded scaffolds with freshly isolated, but unstimulated, PBMCs and ZsGreen-encoding STAT lentivirus. Wild type virus seeded scaffolds were used as a baseline transduction control. Scaffolds with cells-only, but without viruses, were used as negative controls. Dry Drydux scaffolds showed better transduction than pre-wetted scaffolds (Figure S3A), in agreement with our previous data suggesting the importance of the scaffold’s hygroscopic nature in transduction potential [36]. Cells seeded on Drydux scaffolds along with STAT lentivirus demonstrated approximately 20% transduction with little donor-to-donor variability (Figure 2F) and scaffold-mediated transduction was robustly maintained over 9 days in culture (Figure S3B).

Motivated by the ability of Drydux scaffolds and STAT lentivirus to promote selective T cell activation, proliferation, and transduction, we generated clinically relevant, tumor-specific anti-CD19 CAR T cells in a single step using Drydux scaffolds and STAT virus. STAT lentivirus encoding anti-CD19 CAR transgene (referred to as STAT:CD19 CAR) was used. Assessment of early T cell activation following STAT:CD19 CAR virus and naïve T cell incubation on Drydux scaffold demonstrated 40% CD69 expression, compared to 55% expression shown with positive control T cells activated using anti-CD3 and anti-CD28 antibody-coated plates, in good agreement with our previous experiments (Figure S4A). To assess transduction, STAT:CD19 CAR virus and freshly isolated PBMCs were seeded onto Drydux scaffolds. Anti-CD19 CAR-encoding wild type virus (referred to as wild type:CD19 CAR) seeded scaffolds and scaffolds with unstimulated cells but no virus (referred to as control) were used as controls. At day 5 after transduction, cells incubated with STAT:CD19 CAR virus on the Drydux scaffold demonstrated ~7% transduction with minimal donor-related variability. Cells seeded with wild type:CD19 CAR virus on scaffolds showed very limited CD69 activation and negligible transduction (<2%) (Figure 3A, S4B). The transduction efficiency was largely similar to traditional, “gold-standard” transduction with polybrene and spinoculation (~6–10%) highlighting the effectiveness of Drydux scaffolds in generating anti-CD19 CAR T cells in a single step (Figure 3B). The percent transduction was somewhat low throughout these studies, likely owed to the larger size of the construct, which incorporates both an anti-CD19 CAR and GFP. Increasing the MOI significantly improved the transduction efficiency (Figure S4C), suggesting the potential of the Drydux system to be incorporated within existing T cell engineering practices. Almost 95% of cells were recovered from the scaffolds and showed >80% viability (Figure S5A). Additionally, transduced cells showed robust 9-day and 70-fold cell proliferation in culture relative to wild type or cells-only control, suggesting optimal T cell activation by STAT:CD19 CAR virus on the scaffolds (Figure 3C). We found robust proliferation of Drydux/STAT-transduced T cells (Figure S5B) to clinically relevant doses [38], very similar to CAR T cells transduced using conventional methods. Five days after transduction, cells exposed to STAT:CD19 CAR virus showed >95% T cell enrichment, indicating selective outgrowth of T cells due to the combined UCHT1 + CD80 input signals (Figure S5C).

Figure 3: Single step tumor-specific CAR T cell generation using Drydux scaffolds and STAT lentivirus.

A) Transduction potential of Drydux scaffolds seeded with freshly isolated PBMCs and STAT:CD19 CAR virus (*p<0.05 by two-tailed unpaired Student’s t-test). B) Comparison of transduction efficiencies of conventional polybrene + spinoculation method vs scaffold method (two-tailed unpaired Student’s t-test). C) Proliferation assessment of transduced cells in culture over time. D) Assessment of CD4+ and CD8+ population frequencies at 5 days post-transduction (*p<0.05 by two-tailed unpaired Student’s t-test). E) Phenotype of CD4+ T cells at 5 days post-transduction (SCM: Stem Cell Memory, CM: Central Memory, EM: Effector Memory, Eff: Effector)(*p<0.05 by two-tailed unpaired Student’s t-test). F) Phenotype of CD8+ T cells at 5 days post-transduction (SCM: Stem Cell Memory, CM: Central Memory, EM: Effector Memory, Eff: Effector)(*p<0.05 by two-tailed unpaired Student’s t-test). G) Percentage of residual tumor cells at 5 days following co-culture of anti-CD19 CAR T cells generated using Drydux scaffolds and STAT:CD19 CAR virus with CD19+ tumor cells (Daudi) at 1:1 and 1:5 effector:target (E:T) ratios (one-way ANOVA with Tukey’s correction). H) Quantification of IL-2 (one-way ANOVA with Tukey’s correction) and I) IFN-𝛾 (one-way ANOVA with Tukey’s correction) release by anti-CD19 CAR T cells assessed by ELISA after 24 hours of co-culture with CD19+ Daudi tumor cells.

To demonstrate that T cell transduction was specific, transduction was assessed in both CD3+ and CD3-cells. While the STAT:CD19 CAR virus demonstrated transduction only in the CD3+ population, wild type:CD19 CAR virus showed no such preference (Figure S5D). To assess T cell phenotypic composition, T cells transduced on Drydux scaffolds using STAT:CD19 were stained for CD4, CD8, and memory markers (CCR7, CD45RA). On day 5 post transduction, T cells demonstrated ~42% CD4+ population and ~55% CD8+ population frequency (Figure 3D). Among the CD4+ cells within the STAT lentivirus-exposed group, almost 55% of cells showed a less differentiated stem-like memory (SCM) phenotype followed by 34% of the cells exhibiting a central memory (CM) phenotype. Less than 6% of the population demonstrated highly differentiated effector memory (EM) and effector (Eff) phenotypes (Figure 3E). Among the CD8+ cells within the STAT lentivirus-exposed group, 70% of the population showed a stem-like (SCM) phenotype capable of proliferation and differentiation while 27% of the population showed a highly cytotoxic effector (Eff) phenotype essential for the antitumor function of CAR T cells (Figure 3F).

We then looked at in vitro functionality of the CAR T cells by coculturing scaffold generated anti-CD19 CAR T cells with CD19+ tumor lymphoma cells (Daudi) at 1:1 and 1:5 effector to target (E:T) ratios. T cells incubated with wild type CD19 CAR virus and scaffolds seeded with cells (but without virus) were used as controls. On day 5 of co-culture, the presence of residual Daudi cells was determined by flow cytometry. While control T cells and wild type CD19 CAR T cells showed >80% of residual tumor cells remaining at both E:T ratios, CAR T cells generated by STAT:CD19 CAR virus completely eliminated Daudi tumor cells, suggesting highly functional Drydux-generated anti-CD19 CAR T cells (Figure 3G). Since CAR T cells rapidly secrete cytokines when activated in presence of target antigen, we looked at the secretion of IL-2 and IFN-γ 24 hours following co-culture with CD19+ Daudi tumor cells. CAR T cells generated from STAT lentivirus within Drydux scaffolds, but not wild type virus nor control T cells, demonstrated robust release of cytokines, indicating their robust functionality (Figure 3H, 3I).

We assessed the phenotype and cytotoxicity of CAR T cells generated with Drydux / STAT and compared these to conventional methods used in the clinics. Seven days after transduction, scaffold-generated CAR T cells showed similar CD4 and CD8 composition as conventionally generated CAR T cells (Figure S6A), as well as similar immunophenotypic composition within each subgroup (Figure S6B, S6C). Additionally, when tumor killing ability of these cells was tested against Daudi cells in a co-culture experiment at 1:5 E:T ratio, CAR T cells generated using either method demonstrated efficient tumor killing abilities (Figure S6D).

Taken together, the integration of macroporous Drydux scaffolds with STAT lentivirus provides T cell specificity via simultaneous T cell activation and transduction. Incorporation of such platforms into the current CAR T cell manufacturing pipeline could significantly simplify the ex vivo T cell manipulation process. Furthermore, the future use of these scaffolds as an implantation system could allow safe and efficacious CAR T cell generation and delivery in vivo, allowing improved T cell persistence and preventing undesired cell differentiation [15] while significantly reducing the time and cost associated with CAR T manufacturing.

4. Discussion

Engineering therapeutic T cells currently involves multiple labor-intensive steps, including T cell selection, activation, and transduction. These steps create complexity and limit widespread access to this revolutionary therapy. In this work, we demonstrate that biomaterial-mediated transduction in combination with T cell-targeting lentivirus enable selective T cell activation and transduction in a single step. This protocol replaces sequential preselection, pre-activation, and transduction. In addition, this protocol eliminates the need for transduction-enhancing reagents and centrifugation. We expect that the Drydux scaffold + STAT lentivirus platform could be readily incorporated into current manufacturing protocols, improving production of therapeutic cells for treatment of cancer.

Lentiviruses with broad tropism pose a significant safety concern for in vitro and in vivo CAR T cell manufacturing. Recently a case of accidental tumor cell transduction during in vitro CAR T manufacturing lead to treatment resistance [37]. Patients receiving lentiviral-based cell products could inadvertently test false positive for HIV and therefore establishing monitoring protocols that differentiate between bona fide HIV infection and lentiviral gene therapy are essential [39]. Additionally, use of lentiviruses in vivo is hampered by the risk of significant off-target transduction [25,40]. Modular, cell-specific viruses would improve the safety associated with in vivo CAR T cell manufacturing. In this work, we show that STAT lentivirus could specifically target and activate T cells among a mixed PBMC population consisting of B cells, NK cells, and monocytes leading to T cell enrichment post-transduction. Unlike wild-type virus, targeted STAT lentivirus demonstrated T cell-specificity. The ability to selectively target T cells using an engineered lentivirus mitigates the need to preselect the T cell population prior to transduction and improves the overall safety during the transduction process.

It is notable that the STAT lentivirus provides co-stimulatory signals to naïve T cells for robust activation and proliferation. Having both signal 1 (anti-CD3) and signal 2 (CD80) signaling is critical, as viruses only displaying one of these signals were insufficient to induce T cell activation and proliferation.

The Drydux biomaterial platform not only mediates transduction, but could also serve as a nurturing niche to improve T cell function. T cell-specific transduction affords the opportunity to add the entire PBMC mixture (including B cells and dendritic cells) to the scaffold, creating a local environment more similar to a lymph node. In addition, cytokines or other cell signals could be incorporated into the scaffold to promote cell proliferation and reduce cell differentiation [39,40]. This approach could be used in future animal studies to test the feasibility of this platforms as a single step in vivo CAR T cell generation. To further extend the applicability of these platforms beyond CAR T cells, a broad range of receptors and targeting ligands could be employed within these scaffolds for cell-specific single-step transduction.

Previously we reported that Drydux mediates efficient viral transduction of gamma retroviruses. The macroporous Drydux scaffolds promote efficient cell-virus collisions within constriction points during hygroscopy-driven fluid flow. For lentivirus, much like gamma retroviruses, close contact between target cells and viral vectors is crucial and simple co-culture of cells and virus is insufficient due to the short distance viral particles can travel in the solution. In agreement with these previous results, we saw robust transduction with both ZsGreen-encoding as well as CD19 CAR-encoding STAT lentiviruses with little donor-to-donor variability. On the other hand, minimal transduction was observed when cells were incubated with a STAT lentivirus in a solution without a scaffold, highlighting the importance of a scaffold in cell-virus colocalization in the transduction process. In addition to efficient transduction, engineered cells demonstrated activation and proliferation as well as tumor-killing abilities against Daudi cells. Although our transduction efficiencies are comparable to the efficiencies demonstrated in many preclinical studies and clinical trials [41,42], future studies can focus on improvements in transduction efficiency through additional viral engineering.

Overall, this streamlined manufacturing process is highly suitable to fast-track CAR T cell manufacturing ex vivo (and potentially in vivo) within an established setup and could enable expansion to areas with limited medical access by improving overall costs and simplifying manufacturing, improving safety and persistence of adoptively transferred autologous CD19-targeted T cells in patients with relapsed or chemotherapy refractory B-cell malignancies.

5. Conclusion

The combination of biomaterial-driven cell transduction and engineered T cell-specific lentivirus in our present work provided an opportunity to produce CAR T cells from naive PBMCs in a single step, eliminating the need for pre-activation, use of transduction-enhancing reagents, and centrifugation. Combined use of STAT lentivirus with macroporous Drydux scaffolds showed robust T cell activation, proliferation, CAR T cell generation (transduction), and excellent in vitro functionality against tumor cells. Additionally, the ability of viral vectors to specifically target T cells within scaffolds and the scaffold’s potential for implantation demonstrate the feasibility of such approaches for future in vivo applications to overcome safety concerns associated with in vivo T cell reprogramming. Although significant preclinical toxicology and biocompatibility studies need to be done to bring these technologies to clinical trials, these technologies, together, have the potential to bypass many time-consuming, costly, and labor-intensive steps in current CAR T cell manufacturing.

Statement of Significance:

Manufacturing T cell therapies involves lengthy and labor-intensive steps, including T cell selection, activation, and transduction. These steps add complexity to current CAR T cell manufacturing protocols and limit widespread patient access to this revolutionary therapy. In this work, we demonstrate the combination of engineered virus and biomaterial platform that, together, enables selective T cell activation and transduction in a single step, eliminating multistep T cell engineering protocols and significantly simplifying the manufacturing process.