Human T cell engineering via serial delivery of mRNA encapsulated within lipid nanoparticles Authors: Apoorva Ramamurthy, Giulia Scimonelli, William L Murphy, Krishanu Saha Published: 2026 ABSTRACT AbstractHuman T cells modified with nucleic acids constitute a powerful and emerging therapeutic modality for cancer, autoimmune disorders, and aging-related diseases. However, delivering nucleic acids, such as mRNA encoding synthetic receptors, transcription factors, cytokines, or genome editors to T cells can be challenging, as nucleic acids can unintentionally reduce viable T cell yield and function, particularly when delivered sequentially ex vivo. To address this challenge, we evaluated the efficiency of serial non-viral delivery of synthetic mRNA encapsulated within lipid nanoparticles (LNPs) incorporating cationic lipid, 1,2-dioleoyl-3-trimethylammonium-propane (DOTAP), and fusogenic helper lipid, 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE). DOTAP/DOPE-containing LNPs delivered mRNA more efficiently than a clinically benchmarked 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC)-containing LNP formulation, resulting in significantly higher protein expression in primary human T cells. Furthermore, DOTAP/DOPE-LNP delivery resulted in 100% higher yield of live cells compared to electroporation, an advantage that compounded over serial rounds, with LNP-treated cultures maintaining substantially more viable cells through all four transfections. Additionally, serial transfection of mRNAs encoding a therapeutically relevant chimeric antigen receptor (CAR) payload produced functional CAR T cells. These results demonstrate that LNPs can be a viable platform for serial and iterative T cell engineering using multiple mRNA payloads, providing an alternative to electroporation for the engineering of therapeutic T cells.Keywords: MT: Delivery Strategies, T cells, LNP, CAR T cell engineering, electroporation, mRNA delivery, non viral gene delivery, cell therapy manufacturing, DOTAP, DOPE Introduction Engineered immune cell therapies have emerged as a transformative treatment for a variety of diseases, including cancer,1 autoimmune conditions,2,3 and neurodegenerative diseases.4 These therapies can involve modifying cells with a synthetic receptor, such as a chimeric antigen receptor (CAR), to target a specific antigen, thereby enhancing their ability to recognize and destroy diseased cells.5 Several adaptive immune cells have been explored in this context, including natural killer T cells,6 gamma-delta T cells,7 induced pluripotent stem cell-derived T cells,8 and alpha-beta T cells.9 Among engineered T cell products, alpha-beta CAR T cells, in particular, have achieved success in hematological malignancies like B cell lymphomas and acute lymphoblastic leukemia.1,10 Existing FDA-approved CAR T therapies are autologous and therefore require a patient’s T cells to be engineered ex vivo and then reinfused to elicit a therapeutic response.11 However, they present a host of manufacturing challenges due to their customized nature, high manufacturing failure rates, and long cell expansion times.12,13,14 Furthermore, engineered immune cells in applications in oncology must overcome tumor evasion mechanisms such as antigen loss, immune checkpoint-mediated suppression, and major histocompatibility complex (MHC) downregulation. They also face functional defects, such as exhaustion and impaired cytokine production, driving the need for more sophisticated strategies to achieve effective and safe cell therapies.15,16 These challenges have spurred research to develop more potent cell therapies through cost-effective, scalable manufacturing methods. Allogeneic cell therapies use starting material from healthy donors, bypassing potential donor-cell dysfunction caused by the donor’s tumor microenvironment, age, or prior treatment.17,18 There are currently no FDA-approved allogeneic T cell therapies. Current manufacturing techniques for allogeneic CAR T cells require multiple transfections to target and knockout key contributors to graft vs. host disease (GvHD) like MHC-I, MHC-II, and CD52, as well as resistance to immunosuppressive drugs given during treatment.17,19,20,21 For example, in a recent phase 1 allogeneic CAR T clinical trial where cells were sequentially transduced and electroporated, manufacturing failures occurred in up to 26% of cases, with 60% of these failures due to insufficient cell doses for treatment.22 This high failure rate was partly attributed to the toxicity associated with sequential viral transduction and electroporation.22 Current methods for non-viral delivery of nucleic acids can produce unwanted outcomes in biomanufacturing. Electroporation remains the predominant non-viral method for delivering nucleic acids but introduces cellular stress and mitochondrial membrane damage, reducing cell yields, altering protein and gene expression profiles, and resulting in aberrant cytokine expression.23,24,25,26 Additionally, for workflows that deliver genome editing machinery, some outcomes involve multiple double-stranded DNA breaks (DSBs), which could produce genotoxicity and other cellular stress.27,28 Resolution of multiple DSBs can result in unintended chromosomal abnormalities such as translocations, heterozygosity, and long deletions.27,29,30,31 Ultimately, the cellular stress and genotoxicity associated with electroporation could compromise cell viability and drug product safety. Lipid nanoparticles (LNPs) have been implemented as an alternative to electroporation-mediated mRNA delivery in several studies. Vavassori et al. found that although editing efficiency via LNPs at all tested doses of mRNA was reduced compared to electroporation, the yield of LNP-engineered edited hematopoietic cells was greater due to the toxicity of electroporation.32 Existing LNP platforms for T cells use ionizable lipids, cholesterol, and helper lipids to encapsulate negatively charged mRNA through electrostatic interactions.33,34,35,36,37 Newer formulations have introduced synthesized lipids and antibody-functionalized LNPs to specifically target T cells.33,34,38 However, these LNPs have not been used to demonstrate efficient serial delivery of mRNA ex vivo. Emerging approaches to engineering a robust T cell product employ more than a single genetic modification. Some modifications can notably increase the durability of patient response, including the addition of exogenous cytokines,39 recombinant transcription factors,40,41,42,43 transient mRNA reprogramming as seen in stem cells,44 and genome editing.45,46 Genome editing, transcription factor, and cytokine payloads range in size from guide RNAs (∼0.1 kb) to CRISPR editor mRNA (∼4 kb).47 Achieving precise dose and temporal control of these payloads is essential for programming a desired cell phenotype.48,49,50 Sequential delivery of mRNA may offer dynamic control of expression levels over time, minimizing toxicity and allowing flexible modulation. For instance, administration of Trop2 CAR mRNA via LNPs four times every four days has demonstrated safety in murine models without eliciting systemic toxicity.51 Recombinant transcription factors delivered as mRNA or DNA can be electroporated into cells for transient expression, as well as delivered through nanoparticles.41 This approach has been applied successfully with transcription factors like TERT, FOXO1, and FOXP3 to induce varied T cell phenotypes without permanent genomic alteration.52,53,54 In cases requiring a series of transcription factors reach terminal differentiation, sequential mRNA delivery may be necessary.55 Here, we describe an LNP formulation using 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE) and 1,2-dioleoyl-3-trimethylammonium-propane (DOTAP), two lipids commonly used in LNP formulations, to enable serial delivery of mRNA while maintaining high cell yields. We compared our strategy in primary human T cells with an established commercial LNP formulation containing the ionizable lipid SM102, termed SM102-DSPC (1,2-distearoyl-sn-glycero-3-phosphocholine) LNP. Both DOTAP- and DOPE-containing LNPs significantly improved mRNA delivery compared to the SM102-DSPC formulation. We also identified media conditions for robust T cell transfection using the DOTAP/DOPE LNPs. Lastly, to mimic multi-step, non-viral T cell manufacturing where sequential transfection may take place, we evaluated the use of LNPs to serially deliver both reporter mRNA, and an mRNA encoding a urokinase plasminogen receptor (uPAR)-targeting CAR, a therapeutically relevant payload whose target is expressed on senescent cells within diseased and aged tissues.37,56,57,58 The LNP platform produced an equivalent number of transfected, viable, and functional T cells without the toxicity associated with electroporation. Results Titrating DOTAP in an LNP formulation improved mRNA delivery to primary human T cells ex vivo Adding synthetic cationic lipid DOTAP in an LNP formulation significantly enhanced mRNA expression in T cells while maintaining acceptable viability at lower cationic lipid ratios. DOTAP provides a permanent positive charge that significantly improves transfection in hard-to-transfect cell types.59,60,61 This improvement is due in part to the ability of DOTAP to complex with negatively charged nucleic acids and facilitate effective endosomal escape.62,63 As a first-generation synthetic transfection agent, DOTAP has a quaternary ammonium head group that interacts with both nucleic acids and cellular membranes, facilitating intracellular delivery.64,65 Building on the cationic nature of DOTAP, an SM102-DSPC formulation used in coronavirus disease 2019 (COVID) vaccines66 was modified by replacing DSPC with varying amounts of DOTAP from 5% to 40% by molar ratio, while proportionally decreasing cholesterol content (Figures 1A and S1A).63 Dynamic light scattering (DLS) measurements confirmed that all DOTAP LNPs had a hydrodynamic radius under 300 nm and a polydispersity index (PDI) below 0.3 (Figure 1B). Encapsulation efficiency was 90% or higher and was not affected by increasing the DOTAP percentage in the LNP (Figure 1B). Titrating DOTAP in an LNP formulation improves mRNA delivery to primary human T cells ex vivo (A) Schematic representations of lipid nanoparticle (LNP) formulations show that SM102-DSPC LNPs are composed of SM102, DSPC, cholesterol, and DMG-PEG2000, while SM102-DOTAP LNPs are composed of SM102, DOTAP, cholesterol, and DMG-PEG2000. (B) Particle sizes, polydispersity indices (PDIs), and encapsulation efficiencies of eGFP reporter mRNA for all LNPs in the screen (from 3 separately prepared samples, mean ± standard deviation). (C) Delivery of 400 ng eGFP mRNA to T cells using the SM102-DSPC formulation, various formulations containing increasing amounts of DOTAP, and MessengerMAX transfection reagent (mmax) (n = 9, 3 donors, denoted by different shapes). (D) Representative flow cytometry histograms for eGFP protein expression in T cells following treatment with SM102-DOTAP and SM102-DSPC LNPs. Statistical analysis was performed using one-way ANOVA (∗∗∗∗p ≤ 0.0001). Next, we evaluated the delivery performance of DOTAP-containing LNPs. The addition of DOTAP to LNP formulations enhanced the delivery of firefly luciferase (ffluc) and eGFP mRNA to T cells cultured in ImmunoCult-XF compared with the SM102-DSPC formulation or the commercially available transfection agent MessengerMAX (mmax). Expression increased as the percentage of DOTAP rose (Figures 1C and S1B). However, cell viability, assessed using GhostRed780 dye, decreased with increasing DOTAP ratios (Figure S1C).67 The GhostRed780 dye binds to intracellular amines in membrane-compromised cells. SM102-DOTAP5 exhibited the lowest cytotoxicity to T cells and was therefore selected as the lead formulation for subsequent experiments. Adding DOPE to DOTAP-containing LNP improved mRNA delivery to T cells Incorporating DOPE into DOTAP-containing LNPs further enhanced mRNA delivery to primary human T cells at lower doses. DOPE is a neutral helper phospholipid that promotes fusion with the endosomal membrane through its small head group and unsaturated oleoyl chains, which give the lipid a cone-shaped geometry.64 This configuration stabilizes the non-bilayer hexagonal (HII) phase seen in transitional structures during membrane fusion and bilayer disruption.64 More recently, DOPE and DOTAP have been used to deliver mRNA and protein gene-editing cargo ex vivo to neural progenitor cells, human embryonic kidney (HEK) 293 cells, and human bronchial epithelial cells as selective organ-targeting LNPs, highlighting their potential in cell therapy engineering.68,69,70 The SM102-DOTAP5 formulation (containing 5% DOTAP from the LNP in Figure 1) was modified to include 5% DOPE to evaluate the combinatorial effects of DOPE and DOTAP, replacing an equivalent amount of cholesterol to create the SM102-DOTAP5-DOPE formulation, or SM102-DD (Figures 2A and S2A). T cells were transfected with reporter mRNA encoding eGFP or ffluc using SM102-DD and SM102-DOTAP5 at varying doses, and expression levels were measured 24 h post-transfection. Cells treated with SM102-DD exhibited higher ffluc and eGFP expression at lower mRNA doses (Figures 2B, 2C, and S2B). Cell viability decreased with increasing mRNA doses (Figures S2C and S2D). DLS measurements showed that all modified LNPs had hydrodynamic radii below 300 nm. Furthermore, SM102-DD also maintained high encapsulation efficiency (Figure 2E). Adding DOPE to DOTAP-containing LNPs to make SM102-DD LNPs improve mRNA delivery to primary human T cells ex vivo (A) Schematic representations of lipid nanoparticle (LNP) formulation show that SM102-DD LNPs are composed of SM102, 5% DOPE, 5% DOTAP, cholesterol, and DMG-PEG2000, while SM102-DOTAP5 LNPs are composed of SM102, 5% DOTAP, cholesterol, and DMG-PEG2000. (B) Delivery of 400 ng eGFP mRNA to T cells using the SM102-DSPC formulation, SM102-DOTAP5 and SM102-DD (n = 9, 3 donors, denoted by different shapes). (C) Dose-dependent delivery of eGFP mRNA to T cells using SM102-DD and SM102-DOTAP5 LNPs (n = 9, 3 donors). (D) Representative flow cytometry histograms demonstrate eGFP expression in T cells following treatment with SM102-DSPC, SM102-DOTAP5 and SM102-DD LNPs. (E) Particle size and polydispersity indices (PDIs) from 2 separately prepared samples, and encapsulation efficiencies for the SM102-DD LNP formulation from 3 separately prepared samples (mean ± standard deviation). Statistical analysis was performed using one-way ANOVA (∗p ≤ 0.05; ∗∗p ≤ 0.01; ∗∗∗p ≤ 0.001; ∗∗∗∗p ≤ 0.0001). LNP-mediated mRNA delivery to T cells depended on media conditions LNP-mediated mRNA delivery to primary T cells was strongly dependent on culture conditions, with serum-free ImmunoCult-XF supporting substantially higher mRNA expression than RPMI with 10% fetal bovine serum (FBS). We delivered LNPs encapsulating eGFP or ffluc mRNA to T cells in either RPMI with 10% FBS or serum-free ImmunoCult-XF, followed by a 24-h incubation. In T cells cultured in RPMI/10% FBS, both ffluc and eGFP expression were significantly reduced compared to T cells in serum-free ImmunoCult-XF when delivered using SM102-DD (Figures 3A, 3C, S3A, and S3C). Notably, there were nearly 19-fold more live eGFP+ T cells when treated with SM102-DSPC LNPs in RPMI/10% FBS than in ImmunoCult-XF, although mean fluorescence intensity (MFI) and overall cell viability remained similar (Figures 3A, S3A, and S3B). This difference suggests that the effect of media composition on protein expression is formulation-dependent. mRNA delivery to primary T cells ex vivo via SM102-DD LNPs is influenced by media conditions (A) eGFP expression in T cells cultured in RPMI with 10% FBS or ImmunoCult following treatment with SM102-DOTAP5, SM102-DD, or SM102-DSPC LNPs (n = 9, 3 donors labeled as circles, triangles, and squares). (B) SM102-DD LNPs exhibit distinct zeta potentials depending on the culture media used, while SM102-DSPC LNPs exhibit consistent zeta potentials (from 2 separately prepared samples in squares and circles). (C) Representative flow cytometry histograms showing eGFP expression in T cells cultured in RPMI/10% FBS versus ImmunoCult. Statistical analysis was performed using one-way ANOVA (∗p ≤ 0.05; ∗∗p ≤ 0.01; ∗∗∗p ≤ 0.001; ∗∗∗∗p ≤ 0.0001). SM102-DD LNPs and SM102-DSPC LNPs were incubated with PBS, PBS with 2% bovine serum albumin (BSA), ImmunoCult-XF and RPMI/10% FBS to evaluate the influence of protein corona formation under different media conditions on LNP surface charge. Both formulations cultured in RPMI/10% FBS showed a significant difference in zeta potential compared to ImmunoCult-XF, consistent with protein corona modification (Figure 3B). Additionally, we treated Jurkat cells, an immortalized T cell line typically maintained in RPMI/10% FBS, with SM102-DSPC and all SM102-DOTAP formulations. Transfection in Jurkat cells by LNPs differed from that in primary T cells cultured in ImmunoCult-XF across all formulations (Figure S3D), indicating that reporter gene expression is significantly influenced by the cell line, the serum components in the culture medium, or a combination of both. ImmunoCult-XF supplemented with 100 IU/mL IL-2 was used in all experiments moving forward, due to the higher delivery efficiency observed under these conditions. SM102-DD LNPs enter T cells primarily via macropinocytosis The delivery of Cy5-labeled eGFP mRNA via LNPs to primary human T cells cultured in ImmunoCult-XF was evaluated by simultaneously tracking the levels of the mRNA and the translated protein production using flow cytometry (Figure 4A). This labeling strategy enabled us to characterize the relationship between mRNA uptake and translation across multiple LNP formulations.71 Representative contour plots of Cy5 (uptake) versus eGFP (translation) signals in cells treated with SM102-DSPC, SM102-DOTAP5, and SM102-DD LNPs demonstrated that SM102-DD achieved the highest amounts of both uptake and translation (Cy5+eGFP+ cells in Figures 4B and 4C). Notably, a population of Cy5+/eGFP- cells was detectable across formulations, indicating that not all delivered mRNA is translated into protein. This gap was most pronounced in the lower-performing formulation, SM102-DSPC. SM102-DD LNPs enter human T cells primarily via macropinocytosis (A) Schematic of lipid nanoparticle encapsulating Cy5-labeled eGFP mRNA. (B) Representative flow cytometry contour plots of Cy5 signal versus eGFP expression in primary human T cells following transfection with SM102-DSPC, SM102-DOTAP5, and SM102-DD LNPs encapsulating Cy5-labeled eGFP mRNA, distinguishing LNP uptake from mRNA translation (n = 6, 2 donors shown in red and blue, Cy5+eGFP+ highlighted in green). (C) Quantification of live Cy5+eGFP+ transfected cell counts across LNP formulations (n = 6, 2 donors). (D) Representative flow cytometry histograms showing eGFP expression in T cells pre-treated with each inhibitor, with corresponding MFI values indicated. Inhibitors: 5-(N-ethyl-N-isopropyl)amiloride (EIPA, macropinocytosis inhibitor), chlorpromazine (chlor., clathrin-mediated endocytosis inhibitor), methyl-β-cyclodextrin (MβCD, cholesterol-dependent pathway inhibitor), Dyngo-4a (Dyngo, dynamin inhibitor), and nystatin (caveolae-mediated endocytosis inhibitor). (E) Transfected cell yield (eGFP+/GhostRed780-) in primary human T cells pre-treated with endocytosis inhibitors prior to SM102-DD LNP delivery (n = 9, 3 donors labeled as circles, triangles, and squares). (F) Schematic of DiI lipophilic dye-labeled LNP used for lysosomal colocalization imaging. (G) Representative confocal microscopy images of primary human T cells treated with SM102-DSPC, SM102-DOTAP5, and SM102-DD LNPs at 5 h post-transfection. Nuclei are stained with Hoechst (blue), lysosomes with LysoTracker Green (green), and LNPs labeled with DiI (red). Colocalization of DiI and LysoTracker is highlighted using white arrowheads. Scale bars, 10 μm. (H) Pearson’s correlation coefficient (PCC) values quantifying colocalization of DiI-labeled SM102-DOTAP5 and SM102-DD LNPs with LysoTracker Green signal in primary human T cells (n = 4 images). Statistical analysis was performed using Welch’s t test (∗∗p ≤ 0.01, ∗∗∗∗p ≤ 0.0001) in (H), otherwise a Brown-Forsythe and Welch’s ANOVA (∗∗∗p ≤ 0.001, ∗∗p ≤ 0.01, ∗∗∗∗p ≤ 0.0001). To identify pathways responsible for LNP uptake, T cells were pre-treated with pharmacological inhibitors targeting distinct internalization pathways before transfection and then imaged shortly thereafter. Inhibition of macropinocytosis with 5-(N-ethyl-N-isopropyl)amiloride (EIPA), and of cholesterol-dependent membrane organization with methyl-β-cyclodextrin (MβCD), resulted in a significant reduction in transfected cell yield compared to the no-drug control (Figures 4D and 4E). EIPA and MβCD treatment produced the lowest eGFP expression (MFI of 1,370 and 1,710, respectively), approximately 3-fold lower than the untreated control (MFI of 4,483) (Figure 4D). Both treatments also significantly reduced the yield of transfected live cells (eGFP+/GhostRed780-) relative to the no-drug control (Figure 4E). Inhibitors targeting clathrin-mediated endocytosis (chlorpromazine), dynamin-dependent processes (Dyngo), and caveolae-mediated uptake (nystatin) did not significantly reduce transfected cell yield relative to the no-drug control (Figure 4E). These findings suggest that macropinocytosis is the dominant route of SM102-DD LNP internalization in primary human T cells, consistent with previous studies.34,72 The sensitivity to MβCD further supports this, as macropinocytosis is known to depend on cholesterol-rich membrane organization.73,74 To follow LNP subcellular trafficking following uptake, 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate (DiI), a lipophilic tracer, was incorporated into SM102-DD, SM102-DOTAP5, and SM102-DSPC LNPs (Figure 4F). The resulting DiI-labeled LNP formulations were then added to human T cells. Imaging via confocal microscopy at 5 h post-transfection was performed after adding a live LysoTracker Green dye to label lysosomes. SM102-DSPC-treated cells showed minimal intracellular DiI signal (Figure 4G), suggesting that few LNPs were internalized. In contrast, SM102-DD and SM102-DOTAP5 LNPs both showed robust intracellular DiI signal (Figure 4G; white arrowheads), indicating similar levels of internalization. The partial colocalization observed of DiI and LysoTracker signals (Pearson’s correlation coefficient; PCC < 1) indicates that a fraction of LNP-associated DiI signal resides outside lysosomes at 5 h post-transfection, which may reflect LNP material in earlier endosomal compartments or undergoing endosomal escape (Figure 4H). In SM102-DSPC-treated cells, the near-zero DiI signal inside cells precluded meaningful colocalization analysis. Altogether, these tracing studies indicate that SM102-DD LNPs are internalized primarily via macropinocytosis, involving trafficking through the endolysosomal system. Four rounds of serial LNP-transfection yielded more viable T cells than serial electroporation SM102-DD supported efficient and sustained serial transfection of T cells with multiple mRNA payloads. When reporter mRNAs encoding eGFP or mCherry were delivered using SM102-DD, robust transgene expression was maintained across four sequential doses of 0.5 μg per 500,000 cells, or approximately 60 ng per 60,000 cells administered every 24 h (Figures 5A and 5B). In a direct comparison, SM102-DSPC-treated cells exhibited consistently lower cumulative expression under identical dosing conditions, indicating that SM102-DD more effectively facilitates repeated cytosolic delivery and translation of synthetic mRNA (Figures 5B and S4). Identical mRNA doses were delivered to parallel cultures using electroporation with pulse codes EO115 and CM138 on the Lonza 4D Nucleofector to benchmark performance against standard non-viral methods. These are pulse codes commonly employed in T cell mRNA transfection and engineering (Figure S4).24,75,76 Serial mRNA transfection with SM102-DD LNPs generates fewer dead transfected T cells than serial electroporation (A) Timeline of serial transfections: 500,000 T cells were serially electroporated or treated with LNPs delivering either 0.5 μg eGFP mRNA or 0.5 μg mCherry mRNA every 24 h. (B) Representative contour plots of flow cytometric analysis from each time point during the serial transfection experiments, with bolded blue gates indicating eGFP expression and bolded red gates indicating mCherry expression. (C) Representative contour plots of flow cytometric analysis of eGFP+ cells treated with SM102-DSPC or SM102-DD, or electroporated with EO115 or CM138 pulse codes at 120 h. (D) Number of dead eGFP+ cells (within bolded black gates) at 120 h (n = 12, 3 donors, labeled as circles, triangles, and squares). (E) Representative contour plots of flow cytometric analysis of live mCherry+ cells treated with SM102-DSPC or SM102-DD, or with EO115 or CM138 pulse codes at 120 h. (F) Number of dead mCherry+ cells (within bolded black gates in E) at 120 h (n = 12, 3 donors). Statistical analysis was performed using Brown-Forsythe and Welch’s ANOVA (∗p ≤ 0.05; ∗∗p ≤ 0.01). While electroporation and SM102-DD LNPs could both deliver mRNA in a serial fashion, transfected cells died at higher rates in the electroporated samples. Electroporation produced a higher absolute number of live transfected cells at early time points compared to SM102-DD LNPs (Figures S5E and S5F), consistent with its higher per-round transfection efficiency. However, this early advantage was offset by pronounced loss of total viable cells over the serial transfection course (Figure S5B). At the 120 h time point, transfected cells in the electroporated group underwent further cell death (Figures 5C and 5E), while a greater proportion of transfected cells in the LNP-treated group persisted (Figures 5D and 5F). Electroporation resulted in a lower or equivalent absolute cell count over time compared to the SM102-DD treated group, a trend observed for both total cell number (Figures 6A and S5A) and viable cell number determined by GhostRed780 staining (Figures 6B and S5B). Serial mRNA transfection with SM102-DD LNPs produces high cell yields for both transfected and untransfected cells compared to electroporation (A) Number of total cells at each time point during serial transfection (timeline shown in Figure 5A). (B) Number of live cells at each time point. (C) Viability at each time point. (D) Number of live single-positive eGFP+, single-positive mCherry+ and double-positive eGFP+mCherry+ cells at 96 h. (E) Number of live eGFP˗, mCherry˗, and eGFP ˗ mCherry-cells across all time points. n = 12, 3 donors, labeled as circles, triangles, and squares. Statistical analysis was performed using Brown-Forsythe and Welch’s ANOVA (∗p ≤ 0.05; ∗∗p ≤ 0.01; ∗∗∗p ≤ 0.001; ∗∗∗∗p ≤ 0.0001). Overall cell viability percentages over four transfections were statistically similar between LNP and electroporation-treated groups (Figures 6C and S5C). At the 96 h time point, the number of eGFP+ and mCherry+ cells was similar between electroporation and SM102-DD conditions (Figures 6D and S5E). At 120 h, there was no statistically significant difference in cell number between CM138- and EO115-electroporated cells (Figure S5D). Therefore, we moved forward with EO115 as the electroporation comparator for subsequent experiments as it is widely employed in primary T cell mRNA delivery and gene editing workflows, providing the most clinically relevant benchmark for comparison.77,78,79 SM102-DD LNP-treated groups produced broadly similar cumulative yields of transfected cells when compared to electroporation. Although some expected donor-to-donor variation was observed in the absolute yields, the enhanced viability and consistent transfection were observed across all evaluated replicates. In addition, the SM102-DD LNP-treated groups showed a significantly higher overall viable cell number (Figures 6B and S5B) compared to electroporation. SM102-DD LNPs enabled serial transgene delivery to T cells while maintaining high cell viability across all tested conditions. Across all experimental conditions, electroporated cells contained fewer live, untransfected cells than their LNP-treated counterparts, highlighting substantial bystander cell loss attributable to the electroporation process itself (Figures 6E and S5G). In contrast, the LNP method, while slightly less efficient per transfection event, exhibited significantly lower cytotoxicity and preserved a larger population of viable, untransfected cells over multiple transfection rounds, supporting the potential for successive transfections (Figures 6E and S5G). Serial delivery of uPAR CAR mRNA via LNPs produces functional CAR T cells To evaluate whether the cell yield advantage of SM102-DD LNPs extended to a therapeutically relevant payload, we encapsulated mRNA encoding an uPAR-targeting CAR (uPAR CAR). This second-generation CAR comprised an uPAR single chain variable fragment (scFv), a FLAG tag, a CD28 co-stimulatory domain, and a CD3ζ signaling domain, along with 5′ and 3' untranslated regions (Figure 7A). SM102-DD LNPs encapsulating this mRNA maintained physicochemical properties consistent with reporter-loaded formulations, with a hydrodynamic radius below 300 nm and PDI below 0.3 (Figure S6A). Delivery of uPAR CAR mRNA to primary human T cells at doses ranging from 250 to 1,000 ng per 60,000 cells resulted in dose-dependent increases in CAR surface expression as measured by BV421-anti-FLAG staining, with MFI increasing from 722 at the lowest dose to 3,535 at the highest dose, while the number of CAR+ cells remained similar across doses (Figures S6B and S6C). Serial delivery of chimeric antigen receptor mRNA using SM102-DD LNPs produces high cell yields and exhibits functional mRNA-dependent activity (A) Schematic of the uPAR CAR mRNA construct comprising a uPAR single chain variable fragment (scFv), FLAG tag, CD28 hinge and co-stimulatory domain, and CD3ζ chain, as well as 5′ and 3′ untranslated regions (UTRs) delivered to primary human T cells. (B) Timeline of serial uPAR CAR mRNA deliveries at T0, T1 (24 h), T2 (48 h), T3 (72 h), and readouts at T4 (96 h) and 120 h. (C) Representative flow cytometry contour plots of CAR via FLAG-BV421 and viability (GhostRed780+ stain) at each time point for SM102-DD (pink) and SM102-DD Mock (black) conditions. The top left quadrant highlighted in a black box denotes the live CAR+ cells (FLAG-BV421+/GhostRed780˗). 700,000 T cells were serially electroporated or treated with LNPs delivering 2 μg uPAR CAR mRNA every 24 h. (D) Representative flow cytometry contour plots of FLAG-BV421 versus GhostRed780 staining at the experimental endpoint for EO115 (purple) and SM102-DD (pink) conditions, and number of dead CAR+ cells (FLAG-BV421+GhostRed780+; black box) tracked through 120 h. (E) Number of dead CAR+ cells at all time points for EO115 (purple) and SM102-DD (pink) conditions. (F) Percentage of dead CAR+ cells at all time points for EO115 (purple) and SM102-DD (pink). (G) Total number of cells at each time point during serial transfection (n = 9, 3 donors). (H) Number of live CAR+ cells at each time point across SM102-DD-CAR and EO115-CAR conditions (n = 9, 3 donors). (I) Number of live uPAR+ cells at each time point across SM102-DD-CAR and EO115-CAR conditions (n = 9, 3 donors). (J) Number of live uPAR+ cells at T1 (24 h) across SM102-DD CAR, EO115-CAR, SM102-DD Mock and EO115-Mock conditions (n = 9, 3 donors). (K) Percentage of dead cells within the uPAR+ population (% dead uPAR+ cells) at each time point for EO115 (purple) and SM102-DD (pink) conditions (n = 9, 3 donors). (L) Number of live CAR- cells at each time point across SM102-DD-CAR and EO115-CAR conditions (n = 9, 3 donors). Statistical analysis was performed using Brown-Forsythe and Welch’s ANOVA (∗p ≤ 0.05; ∗∗p ≤ 0.01; ∗∗∗p ≤ 0.001; ∗∗∗∗p ≤ 0.0001). Primary human T cells were then serially transfected with uPAR CAR mRNA to evaluate CAR T cell production. The dose level was fixed at 171 ng per 60,000 cells and transfection occurred every 24 h to mimic prior manufacturing strategies.80,81,82,83,84 mRNA was delivered using either SM102-DD LNPs, or an electroporation control with pulse code EO115 (Figure 7B). CAR was expressed and trafficked to the cell surface after each dose using the SM102-DD LNPs (Figure 7C). The number of dead CAR+ cells in electroporated cultures remained significantly higher than that of LNP-treated cultures at all time points (Figures 7D, 7E, and S6E). This is consistent with the compounding cytotoxicity observed with serial reporter mRNA delivery in Figures 5C–5F. The percentage of CAR+ dead cells was comparable or lower in SM102-DD-treated cultures relative to electroporated cultures up to 96 h, with electroporated cultures showing significantly greater relative CAR+ cell death at 24 h and 72 h (Figure 7F). Additionally, SM102-DD-treated cells maintained significantly higher or comparable cell counts and specifically live cell counts compared to electroporated cells at all timepoints, (Figures 7G and S6D), consistent with our findings using reporter mRNA (Figures 6A and 6B). The number of live CAR+ cells was similar to that of electroporated cultures at the 24 h and 72 h time points (Figures 7H and S6F). At the 48 h and 96 h time points post serial transfection, the number of CAR+ cells in the electroporated cultures was slightly higher than in LNP treated cells (Figures 7H and S6F). As uPAR is endogenously expressed on a subset of T cells within ex vivo culture, we hypothesized that successful delivery and translation of the CAR would enable self-killing in a subset of cells, or directed fratricide,57 whereby CAR-expressing T cells kill uPAR+ cells within the same culture. Live uPAR+ cell numbers declined in both CAR conditions beginning as early as 24 h after the first transfection and continuing through 96 h, at which point depletion was equivalent between LNP- and electroporation-treated cultures (Figure 7I). This fratricide-driven elimination was further confirmed at the T1 time point, where both CAR conditions showed markedly reduced live uPAR+ populations within 24 h of initial mRNA delivery compared to mock controls in which no payload (only water) was delivered by either modality (Figure 7J). The percentage of dead cells within the uPAR+ population (Figure 7K) differed significantly at 24 h, with electroporated cultures showing greater relative target cell killing than SM102-DD-treated cultures, consistent with more rapid onset of CAR expression following electroporation.24 This difference resolved between 48 and 96 h, during which both conditions achieved comparable, near-complete elimination of the uPAR+ population (Figure 7K). The cell death tracked in the cultures in Figure 7 is consistent with CAR-dependent cytotoxicity, but these assays did not fully distinguish the effects of electroporation-mediated cell death from the CAR-dependent cytotoxicity. Finally, the number of live CAR- cells was also higher in SM102-DD LNP-treated populations across all time points compared to electroporated cells, indicating that untransfected cells are similarly protected from the bystander toxicity associated with electroporation (Figure 7L), consistent with experiments involving fluorescent mRNAs (Figure 6E). Together, these analyses indicate that SM102-DD LNP delivery achieves target cell killing comparable to electroporation. To benchmark the functional activity of CAR T cells generated by serial transfection, a target cell line, uPAR293, was engineered to express a defined level of uPAR, as described previously.57 uPAR surface expression on uPAR293 target cells was confirmed by (PE)-anti-uPAR staining before co-culture (Figure S6G). The target uPAR293 cells were co-cultured with CAR T cells at a 1:1 effector-to-target ratio and monitored by live imaging while stained with Incucyte Cytotox Green Dye (Figure 8A). One group received four rounds of uPAR CAR mRNA delivery via SM102-DD LNPs, and the other via EO115 electroporation (Figure 8B). At the beginning of co-culture at 0 h, both CAR T groups displayed low baseline fluorescence. By the 8 h time point, dense, bright, localized fluorescent cell clusters appeared, arising from robust target-cell killing. The uPAR293 negative control group maintained minimal background fluorescence throughout the entire duration, confirming high baseline cell viability in the absence of treatment. Meanwhile, the uPAR293 DMSO positive control group exhibited a rapid shift to intense green fluorescence across the entire cell monolayer, capturing non-specific chemical-induced cell death. Across three different manufacturing runs from different donors, the total green integrated intensity of SM102-DD-CAR-treated co-cultures was significantly higher than that of uPAR293 target cells alone, indicating robust CAR-dependent cytotoxic activity at the 8 h time point (Figure 8C; Videos S1, S2, S3, and S4). SM102-DD-CAR-treated cells also exhibited a significantly higher proportion of activated CD69+CD25+ cells compared to untransfected controls, while SM102-DD-eGFP and SM102-DD-Mock conditions showed intermediate activation levels (Figure 8D). Additionally, IFN-γ secretion in the culture supernatant was measured by ELISA following co-culture with uPAR293 target cells. SM102-DD CAR T cells secreted significantly higher levels of IFN-γ following co-culture with uPAR293 target cells compared to mock SM102-DD controls (Figure 8E), consistent with the higher activation status and cytotoxicity results. Electroporated uPAR CAR cells were also activated and secreted IFN-γ following co-culture with uPAR293 target cells. Overall, the functionality of a CAR mRNA payload within a single T cell is comparable when delivered through either electroporation or SM102-DD LNPs. CAR T cells generated by SM102-DD LNP delivery are cytotoxic and exhibit antigen-specific activation with cytokine secretion (A) Schematic of the co-culture cytotoxicity assay, in which uPAR CAR T cells are plated with uPAR-HEK293 (uPAR293) target cells. (B) Representative Incucyte phase and green fluorescence overlay images at 0, 8, 24, and 48 h after the start of the co-culture for SM102-DD-CAR T cells with uPAR293 and EO115-CAR T cells with uPAR293 at a 1:1 effector to target ratio. Controls include uPAR293 alone, and uPAR293 treated with DMSO as a positive control for target cell death. Green cells indicate dead uPAR293 cells as detected by Incucyte Cytotox Green Dye via live imaging. Scale bars, 100 μm. (C) Cytotoxicity, total green integrated intensity change, in the assays in (B) (n = 9, 3 donors labeled as circles, triangles, and squares). Fold change is the integrated intensity at the endpoint relative to green intensity at t = 0. (D) Percentage of CD69+CD25+ cells across SM102-DD-CAR, SM102-DD-eGFP, SM102-DD Mock, EO115-CAR, EO115-eGFP, EO115-Mock, and untransfected conditions following co-culture (n = 9, 3 donors). (E) IFN-γ concentration measured by ELISA in supernatants collected after co-culture of uPAR293 target cells with SM102-DD CAR, EO115-CAR, SM102-DD-Mock, or EO115-Mock T cells (n = 9, 3 donors). Statistical analysis was performed using Brown-Forsythe and Welch’s ANOVA (∗p ≤ 0.05; ∗∗p ≤ 0.01; ∗∗∗p ≤ 0.001; ∗∗∗∗p ≤ 0.0001; ns, not significant). Video S1. uPAR293 treated with DMSO Time-lapse fluorescence imaging using IncuCyte of uPAR293 target cells treated with DMSO and green IncuCyte Cytotox Dye. Video S2. uPAR293 Time-lapse fluorescence imaging using IncuCyte of uPAR293 target cells treated with Green IncuCyte Cytotox Dye. Video S3. uPAR293 co-cultured with EO115-CAR T cells Time-lapse fluorescence imaging using IncuCyte of uPAR293 target cells co-cultured with electroporated T cells expressing uPAR CAR mRNA at a 1:1 effector-to-target ratio, and treated with green IncuCyte Cytotox dye, showing progressive accumulation of green fluorescent signal indicative of CAR-mediated cytotoxicity over time. Video S4. uPAR293 co-cultured with SM102-DD-CAR T cells Time-lapse fluorescence imaging using IncuCyte of uPAR293 target cells co-cultured with SM102-DD-transfected T cells expressing uPAR CAR mRNA at a 1:1 effector-to-target ratio, and treated with green IncuCyte Cytotox dye, showing progressive accumulation of green fluorescent signal indicative of CAR-mediated cytotoxicity over time. T cell subset composition and exhaustion markers were also assessed following serial transfection with the CAR mRNA. In LNP-delivered (SM102-DD) groups, PD-1+TIM-3+ co-expression did not differ significantly between CAR, eGFP, and mock conditions, whereas in electroporated (EO115) groups, CAR-transfected cells showed significantly higher PD-1+TIM-3+ co-expression than both eGFP- and mock-transfected controls (Figure S6H). Co-expression of the exhaustion markers PD-1, TIM-3, and LAG-3 remained below 1.5% across all conditions, indicating that serial transfection did not induce substantial T cell exhaustion overall. CD4 and CD8 subset composition was comparable across SM102-DD and EO115-transfected conditions (Figure S6I). Altogether, the SM102-DD platform produced CAR T cell products with similar composition and function to those generated by electroporation strategies while preserving a larger population of viable, untransfected cells over multiple transfection rounds. Discussion Our work demonstrates the feasibility of using LNPs to deliver mRNA to T cells while maintaining viability and yield, aligning with findings from other studies in the field.24,32,36,85 Additionally, we show that serially transfecting T cells up to four times over 96 h when using SM102-DD LNPs maintained higher T cell yields compared to electroporation using the Lonza 4D Nucleofector. The transfection efficiency with LNPs remained lower than with electroporation during the first two transfections. However, the lower toxicity of LNP transfection offered significantly increased cell yield and could be advantageous during complex manufacturing workflows requiring multiple genetic modifications.24,32 Even optimized approaches consolidate multiple editing components into a single electroporation round to minimize genotoxicity and preserve cell yield. This strategy reduces handling steps but does not eliminate the underlying trade-off between delivery efficiency and cellular toxicity, reinforcing the need for alternative strategies that can sustain genomic integrity and viable cell numbers across multi-step workflows.79 In complex workflows, LNP-mediated mRNA delivery may offer advantages over electroporation for CAR T cell engineering, and our study extends this to a serial multi-round delivery context where the cell yield advantage compounds across transfection rounds.86 For some applications, a higher single-round efficiency may remain preferable where high bursts of mRNA activity are required, such as biallelic knockout strategies. Notably, transfection efficiency with SM102-DD LNPs increased in a dose-dependent manner (Figure S6C), suggesting that per-round efficiency could be further improved through dose optimization or formulation adjustments without necessarily sacrificing the demonstrated yield advantage. Adjusting seeding density or culture vessel geometry may further improve delivery efficiency, as has been shown to influence nanoparticle-cell interactions in other contexts.87 Additionally, the transfection efficiency required is highly dependent on the application. At the transfection levels achieved in this study, serially engineered uPAR CAR T cells retained robust CAR-mediated cytotoxicity against target cells (Figures 7H–7J and 8C), indicating that lower per-round efficiency does not necessarily compromise downstream functional potency for applications where a threshold level of expression is sufficient. Serial electroporation at 24-h intervals was selected as a comparator to benchmark the cumulative damage of repeated membrane disruption via electroporation against repeated LNP exposure under equivalent dosing schedules. Short-interval sequential electroporation has been previously reported, with studies demonstrating repeated nucleofection within 16- to 24-h windows for multiplex genome editing and CAR T manufacturing workflows, including sequential delivery of Dicer-substrate small interfering RNA (siRNA) and TCR mRNA to primary T cells,80 and two-step electroporation of transcription activator-like effector nucleases (TALENs) and double stranded DNA template for non-viral gene insertion.83 Future studies comparing our delivery schedule to other optimized multi-step protocols would further clarify the translational advantages of the SM102-DD LNP platform. LNP-mediated delivery preserved not only the population of modified cells, but also sustained a significantly larger cohort of viable, untransfected cells throughout serial transfections (Figures 6E and S5G). Donor-to-donor variability was observed across 16 donors for absolute yields of transfected cells (Figure 7H), but LNPs outperformed electroporation for each donor with respect to total live cells at the end of serial delivery (Figure 7G). Larger and demographically diverse cohorts will be needed to identify any donor characteristics that predict transfection outcomes. In contrast, electroporation led to pronounced depletion of untransfected cells, likely due to bystander cytotoxicity, which restricted the pool of cells available for subsequent rounds of gene modification. This cell preservation with LNP-based delivery extends the platform’s utility by allowing additional rounds of cell engineering and ensuring robust starting material for processes that require iterative dosing, pooled modifications, or cell expansion. The cell yield advantage of SM102-DD LNPs extended to serial delivery of a therapeutically relevant uPAR CAR construct, with serially engineered CAR T cells demonstrating functional cytotoxic activity comparable to electroporated cells. The higher proportion of CD69+CD25+ cells observed in SM102-DD-treated CAR conditions compared to electroporation suggests that LNP-mediated delivery may influence T cell activation state. This finding may have implications for delivery to resting T cell populations, as activated T cells have increased macropinocytosis,72 which is the dominant uptake route identified for SM102-DD LNPs in this study. Across multiple functional and phenotypic readouts, CAR-dependent effects were more robustly resolved in the electroporation condition than in the LNP condition. Within the electroporation arm, CAR-expressing cells showed significantly higher activation (Figure 8D, EO115-CAR vs. EO115-eGFP, p < 0.01), IFN-γ secretion (Figure 8E, p < 0.001), and PD-1+TIM-3+ frequency (Figure S6H, p < 0.0001) compared to non-CAR controls. Within the SM102-DD arm, this CAR-dependent effect was either not statistically significant (activation, Figure 8D; PD-1+TIM-3+; Figure S6H) or only weakly significant (IFN-γ secretion, Figure 8E, p < 0.05). Notably, all SM102-DD conditions showed elevated PD-1+TIM-3+ frequency relative to their EO115 counterparts regardless of payload, suggesting that LNP treatment by itself may raise this marker independent of CAR expression and obscure a CAR-specific signal against an elevated baseline. Clarifying how LNP treatment without payloads activates and potentially exhausts T cells will be an important area for future formulation optimization. Broader applications of this platform may include transcription factor delivery for cellular reprogramming, or allogeneic CAR T cell manufacturing, where previous studies have explored delivering gene-editing cargo such as CRISPR-Cas9, base editors, TALENs, or mRNA multiple times to knockout genes, along with viral transduction to knockin CAR.33,52,53,88,89,90 Together, these data suggest that the lower bystander toxicity and sustained viable cell population provided by LNPs can directly benefit scalable cell therapy manufacturing and the development of advanced genetic engineering protocols. Additionally, our results highlight that the maintenance of absolute T cell numbers and viability across multiple engineering rounds is not always predicted by percentage metrics alone. This distinction is particularly critical in clinical and manufacturing settings where maximum output from limited starting material is required. The cumulative preservation of viable, modifiable cells after repeated LNP transfection highlights the potential of this platform for scalable, flexible, and less cytotoxic T cell modification strategies. It also motivates future efforts to combine LNPs with cargo targeting or new engineering modalities to further enhance the specificity and efficiency of gene delivery in T cell therapy manufacturing. This study contributes key findings in an LNP formulation space for T cell applications and shows that substituting DSPC with 5% cationic DOTAP in the clinically approved SM102-DSPC LNP resulted in 50-fold more transfected cell yield. This improvement may result from the cationic charge of DOTAP, which may have promoted the adsorption of a favorable protein corona that facilitated T cell uptake. Such a mechanism is consistent with prior work showing that surface charge influences nanoparticle-protein interactions and subsequent cellular entry.85 Adding helper lipid DOPE augmented the improvement in transfection by 1.8-fold, boosting both protein production and the number of transfected cells, possibly due to an increase in endosomal escape mediated by the fusogenic nature of DOPE.65,70 DOPE and DOTAP are commonly used lipids, having been explored in a good manufacturing practice (GMP) context in RNA delivery using silicon nanoparticles and liposomal formulations, making them strong candidates for integration into scalable manufacturing platforms.88 The hydrodynamic radii of SM102-DD LNPs are larger than some ionizable LNP formulations reported in the literature but are consistent with DOTAP-containing formulations used for mRNA delivery to primary cells such as Selective ORgan Targeting (SORT) LNPs, where sizes in the 150–300 nm range have demonstrated efficient cellular uptake and intracellular delivery.63 As SM102-DD LNPs were synthesized using vortex mixing, more controlled microfluidic mixing methods may yield smaller and more uniform particles.91 Unlike ionizable LNP platforms requiring synthesized lipids or ApoE supplementation to achieve T cell transfection, such as those described previously,24,90 SM102-DD incorporates commercially available, well-characterized lipids without targeting ligands or additional media supplements, simplifying potential translation to scalable manufacturing workflows.24,90 The composition of T cell culture media also significantly influenced LNP-based transfection. T cells are commonly cultured in RPMI supplemented with 10% serum to support viability and proliferation in ex vivo assays.36,37,89,90 Furthermore, Jurkat cells, a frequently used T cell line, are also cultured in RPMI with 10% serum for survival and proliferation. Several previous studies have used RPMI with 10% serum for screening different LNP formulations in Jurkat cells before screening in primary T cells.90,92 In our study, the presence of FBS influenced the transfection efficiency of the SM102-DSPC, SM102-DOTAP5, and SM102-DD LNPs. Changes in zeta potential observed after incubation in different media suggest that this effect may be due to the adsorption of serum proteins onto the LNP surface, potentially altering surface charge, stability, or cellular interactions. Recent work comparing cationic DOTAP-containing LNPs to charge-neutral formulations for ex vivo mRNA delivery to primary T cells93 describes an advantage of cationic charge in enabling transfection, independent of T cell activation status and culture medium composition. These results are consistent with our findings that SM102-DD outperformed SM102-DSPC in conditions that could have low T cell activation, for some donors and in serum-free conditions. Given the established scale-up protocols for LNPs developed during COVID-19 mRNA vaccine production, their application in T cell engineering may provide an alternative to electroporation that maintains cell viability and yield.94 This approach has the potential to be integrated into automated, closed-loop manufacturing systems already in use for CAR T production, such as the Miltenyi Prodigy, Lonza Cocoon, or Cellares CellShuttle. The expression kinetics of mRNA delivered through LNPs compared to electroporation have also been shown to vary. In a study by Kitte et al., CAR T cells engineered using LNPs displayed CAR expression on their surface for three days longer than those modified by electroporation. Although initial CAR expression was higher in electroporated cells, the prolonged expression observed with LNPs led to a more potent CAR T response.24 The narrowing gap in uPAR+ cell depletion between LNP- and electroporation-treated cultures over time (Figure 7I) may reflect differences in CAR expression kinetics rather than a deficit in LNP-mediated cytotoxic potential. This interpretation is consistent with reports indicating that electroporation may produce a more rapid onset of CAR expression, whereas LNP delivery results in more gradual yet sustained expression.24 Evaluating and comparing the trade-off between expression kinetics and cell yield for mRNA delivery applications will be essential when using LNPs to engineer CAR T cells. Moreover, the vast design space of LNPs, including tunable lipid compositions, customizable cargo loading, and surface modifications, enables tailored solutions for a range of cell types and delivery goals.37,90,95,96 The modularity of LNPs allows for strategic pairing of specific lipid compositions with cargo types such as mRNA encoding transcription factors. This could enable transient modulation of T cell phenotype or function during manufacturing with drastically lower risks associated with viral integration or permanent genome editing. While not evaluated in this study, it is important to consider that LNPs may elicit inflammatory responses in T cells, as observed in some vaccine contexts,97,98,99 warranting further investigation. Studies from mRNA vaccine development have shown that certain LNP formulations can activate innate immune responses via TLRs or inflammasome pathways.97,100 Amine headgroups in ionizable lipids have been shown to trigger LNP-induced innate immune activation by engaging TLR4 and triggering NF-κB-dependent cytokine responses.101 Future work should investigate whether repeated LNP exposure alters the expression of inflammatory cytokines, such as IL-6, or affects T cell activation status, exhaustion marker expression, and differentiation, particularly in the context of manufacturing clinically relevant doses of allogeneic products. Electroporation has also been shown to alter cytokine expression23; therefore, repeated electroporation may similarly increase inflammatory cytokine expression. Additionally, our study evaluated activation and exhaustion phenotypes only at the endpoint of serial transfection. Future studies examining longer-term persistence, memory differentiation, and functional durability after extended expansion will further inform the suitability of this platform for clinical manufacturing. Finally, different mRNA doses and transcript lengths may influence cell viability and yield, and these parameters warrant further investigation. In summary, our study demonstrates that LNPs represent a promising alternative to electroporation for complex T cell engineering by improving cell viability and yield during sequential transfection. The ability to fine-tune LNP composition, including the incorporation of DOTAP and DOPE, provides a flexible platform for optimizing cellular uptake and transgene expression while remaining compatible with scalable, closed-system manufacturing workflows. Comparative studies evaluating expression kinetics, inflammatory responses, and therapeutic efficacy across electroporation- and LNP-based approaches will help define the most effective strategies for next-generation engineered T cell therapy manufacturing. Materials and methods T cell isolation and culture CD3+ T cells were harvested from Leukocyte Reduction Cones (Versiti) generated from healthy donors (16 different donors over the entire study) using the RosetteSep Human T cell Enrichment Cocktail (STEMCELL Technologies) according to manufacturer instructions and activated for 48 h using 25 μL/mL STEMCELL Human CD2/CD3/CD28 Activator. T cells were cultured at a density of 1 × 106 cells/mL in ImmunoCult-XF T cell Expansion Media (STEMCELL Technologies, Vancouver, Canada) supplemented with 100 IU/mL IL-2 (Peprotech, New Jersey). Every 2 days, T cells were passaged with new media and fresh IL-2 at the same concentration. Donors were not age- or sex-matched across experiments, and the number of donors was selected based on conventions in the primary T cell delivery literature rather than a formal power calculation.24,86,90,93,102 LNP formulation and characterization To synthesize LNPs, an aqueous phase containing mRNA and an ethanol phase containing lipid and cholesterol components were mixed using the vortexing method at a 1.8:1 aqueous to ethanol ratio as described previously.63 Briefly, the aqueous phase was prepared using 10 mM citrate buffer and eGFP (998 nt) or ffluc mRNA (1,922 nt) with M6 cap and N1-Methyl-PseudoU substitution (Trilink Biotechnologies, SD, USA). The SM102-DSPC LNP was made using the ionizable lipid SM102 (MedKoo, Morrisville, NC, USA), phospholipid DSPC (Avanti Polar Lipids, Birmingham, AL, USA), DMG-PEG2000 (Avanti Polar Lipids, Birmingham, AL, USA) and cholesterol (Sigma-Aldrich, Burlington, MA, USA), combined in a molar ratio of 50%, 10%, 1.5%, and 38.5%, respectively into 100% ethanol.94 Similarly, the SM102-DOTAP5 LNP was made by replacing 5% DSPC with 5% cationic lipid DOTAP (Avanti Polar Lipids, Birmingham, AL, USA) and adding 5% to the molar ratio of cholesterol. The SM102-DD LNP was made using 5% phospholipid DOPE in addition to 5% DOTAP, along with SM102, DMG-PEG2000, and cholesterol. The N/P ratio of SM102-DD LNPs was 6, calculated based on the molar ratio of ionizable lipid nitrogen groups to mRNA phosphate groups. The resulting lipid concentration was held constant at 10 mM. The resulting LNP mixture was dialyzed against 1× PBS for 2 h at 4 °C using Slide-A-Lyzer MINI Dialysis Units (Thermo Fisher Scientific 10K MWCO). DLS through a Zetasizer Nano (Malvern Panalytical, Malvern, UK) was used to measure LNP size and PDI. To assess encapsulation efficiency, a Ribogreen Assay was performed using the Quant-iT RiboGreen kit (Thermo Fisher Scientific, Waltham, MA, USA). As described previously, Triton X-100 was used to lyse the LNPs for 10 min before measuring the concentration of RNA in lysed and unlysed samples using the standard curve generated.34 The encapsulation efficiency was calculated using the formula34 (mRNA concentration in lysed sample ˗ mRNA concentration in non-lysed sample)/mRNA concentration in lysed sample × 100. To measure zeta potential and the effects of different media compositions on zeta potential, LNPs were incubated with media at a 1:10 ratio by volume for 1 h while gently shaking at 37 °C before being diluted 1 to 100 in nuclease-free water to measure zeta potential using the Zetasizer Nano and DTS1070 cuvettes.103 LNP delivery to T cells or Jurkat cells in vitro After dialysis, LNPs were diluted using 1× PBS. Each dose was delivered in 60 μL PBS to 60,000 T cells in 60 μL of media or 500,000 T cells in 190 μL of media, or 60,000 Jurkat cells in 60 μL RPMI/10% FBS media and mixed using a multichannel pipette.90 After 24 h incubation, the cells were washed in 1× PBS, incubated in CellTiter Blue solution (Promega, Madison, WI, USA) for 2 h, and absorbance was measured according to the manufacturer’s instructions. For luciferase measurements, cells were lysed with 100 μL Cell Lysis Buffer (Promega, Madison, WI, USA). 20 μL of cell lysate was treated with 100 μL luciferase assay buffer made using 50 mM Tris, 10 mM MgSO4, 6 mM cysteine, 0.02 mM sodium pyrophosphate, 1 mM EDTA, and 2 g/L BSA as described previously.104 The working buffer was supplemented with 0.15 mg/mL D-Luciferin and 1 mM ATP. Luminescence was measured using a plate reader at 590 nm with a gain of 200. For eGFP expression measurement, cells were washed using 1× PBS and treated with GhostRed780 Viability Dye at a dilution of 1:200 for 15 min at room temperature. After incubation, cells were washed with 1% BSA solution in 1× PBS, and run on the Attune NxT Flow Cytometer. FCS files were analyzed using FlowJo v.10.10.0. In vitro transcription and purification of uPAR CAR mRNA A DNA template encoding the FLAG-tagged uPAR CAR construct was generated by PCR amplification with primers shown in Table S1 from a plasmid with the mRNA sequence on a pcDNA3.1(+) plasmid backbone (GenScript). mRNA was synthesized using the HiScribe T7 mRNA Kit with CleanCap AG (New England Biolabs) according to the manufacturer’s instructions. Following synthesis, dsRNA contaminants were removed by cellulose purification as described previously.105 Briefly, mRNA was incubated with cellulose (C6288, Sigma-Aldrich) in chromatography buffer containing 10 mM HEPES (pH 7.2), 0.1 mM EDTA, 125 mM NaCl, and 16% (v/v) ethanol. dsRNA-depleted mRNA was eluted and recovered using microcentrifuge spin columns (NucleoSpin Filters, Macherey-Nagel, Düren, Germany). mRNA integrity and purity were confirmed by denaturing gel electrophoresis prior to use in downstream experiments. Electroporation and LNP-mediated delivery of reporter mRNA 0.5 μg of eGFP mRNA or 0.5 μg of mCherry mRNA was delivered to 500,000 T cells in 250 μL ImmunoCult-XF supplemented with 100 IU/mL IL-2 using SM102-DD LNPs, or nucleofected with the Lonza 4D Nucleofector. Two pulse codes, CM138 and EO115, were used. For the serial transfections with reporter mRNA payloads, this was performed starting with the first transfection (T0) at 0 h, delivering 0.5 μg of eGFP mRNA to 500,000 primary human T cells. Subsequent transfections occurred every 24 h, using ½ the volume of the previous culture. This included the second transfection (T1) at 24 h using 0.5 μg mCherry mRNA on T1 cells, third transfection (T2) at 48 h using 0.5 μg eGFP mRNA on the T2 cells, and fourth transfection (T3) at 72 h using 0.5 μg mCherry mRNA on the T3 cells. A fourth of the culture was kept in culture to analyze the duration of protein expression over time. Finally, flow cytometry was used to analyze the cell populations by sampling ¼ of the volume of the culture every 24 h, specifically at 24 h (T0 cells), 48 h (T0 and T1 cells), 72 h (T0, T1, and T2 cells), and 96–120 h (T0, T1, T2, and T3 cells). To extrapolate these data, the total cell count for the main culture on any given day was calculated by correcting for the daily ¼ sample and the progressive halving of the culture, using the following formula: where n is the time point number, starting at 1 for the first day of sampling (24 h), 2 for the second day of sampling (48 h), and so on. The total cell count for each plate kept in culture for protein expression analysis over time was calculated by correcting for the daily ¼ sample and the cumulative dilution of cell density caused by the repeated sampling. The formula is: where k is the number of times that specific culture has been sampled (k starts at 1 on the first day of analysis for that culture). Serial delivery of uPAR CAR mRNA Primary human T cells were isolated on day 0 and activated for 48 h. On day 3 post-isolation, serial transfections were performed starting with the first transfection (T0) at 0 h, delivering 2 μg of FLAG-uPAR CAR mRNA, eGFP mRNA (transfected control), or water (mock control) to 700,000 T cells in 260 μL ImmunoCult-XF supplemented with 100 IU/mL IL-2 using SM102-DD LNPs or electroporation with pulse code EO115 on the Lonza 4D Nucleofector. Subsequent transfections occurred every 24 h using the same dose. This includes the second transfection (T1) at 24 h, third transfection (T2) at 48 h, and fourth transfection (T3) at 72 h, each delivering 2 μg of uPAR CAR mRNA, eGFP mRNA, or water to the previous culture. A quarter of the culture was kept at each time point to analyze cell populations over time. Flow cytometry was used to analyze cell populations every 24 h by sampling ¼ of the volume of the culture at 24 h (T0 cells), 48 h (T1 cells), 72 h (T2 cells), 96 h (T3 cells, also called the T4 time point) and 120 h. Flow cytometry was used to assess CAR surface expression (BV421-anti-FLAG, clone L5, BioLegend), uPAR expression (PE-anti-uPAR, clone VIM5, BioLegend), and viability (GhostRed780, Cytek) at each time point, and T cell subset composition (CD4, CD8, CD69, CD25, PD-1, TIM-3, and LAG-3) at the 120 h time point. Cells were incubated with GhostRed780 in 1× PBS for 15 min in the dark at 4 °C, before washing with 1% BSA in 1× PBS. Next, cells were incubated in antibodies for 1 h at 4 °C, washed with 1% BSA in 1× PBS, and analyzed on the Attune NxT Flow Cytometer. To extrapolate total cell counts, the full population at each time point was calculated by correcting for the daily ¼ sample and the cumulative dilution of cell density caused by repeated sampling: where k is the number of times that specific culture has been sampled (k starts at 1 on the first day of analysis). At the 120 h time point, an additional correction factor was applied to account for the half-well volume sampled: For all calculations, the volume resuspended before flow cytometry was 260 μL, the volume run by the cytometer was 190 μL, and the fraction sampled each day was 0.25. Phenotyping Following serial transfection, T cells were stained for surface markers CD4 and CD8 without fixation. Cells were incubated with anti-CD4 (clone SK3, BioLegend) and anti-CD8 (clone RPA-T8, BioLegend) for 1 h at 4 °C, washed with 1% BSA in 1× PBS, and analyzed on the Attune NxT Flow Cytometer. FCS files were analyzed using FlowJo software v.10.10.0, using the gating strategy described in Figure S7. For activation and exhaustion marker staining, T cells were fixed in 4% paraformaldehyde for 30 min at room temperature and stored at 4 °C for up to one week prior to staining. Fixed cells were then stained with anti-CD69 (clone FN50, BioLegend), anti-CD25 (clone BC96, BioLegend), anti-PD-1 (clone EH12.2H7, BioLegend), anti-TIM-3 (F38-2E2, BioLegend), and anti-LAG-3 (C9B7W, BioLegend) for 1 h at 4 °C, washed with 1% BSA in 1× PBS, and analyzed on the Attune NxT Flow Cytometer. FCS files were analyzed using FlowJo software v.10.10.0, using the gating strategy described in Figure S7. Cytotoxicity assay uPAR-HEK293 target cells were plated at 10,000 cells per well in a 96-well plate 24 h prior to co-culture. CAR T cells generated by serial transfection were added at a 1:1 effector to target ratio (10,000 CAR+ effector cells per well) in ImmunoCult-XF supplemented with 100 IU/mL IL-2. Incucyte Cytotox Green Dye (Sartorius) was added at a final concentration of 250 nM at the start of co-culture. Co-cultures were monitored for 66 h using the Incucyte live imaging system. Green total integrated intensity was used as a measure of target cell viability and cytotoxic activity. DMSO-treated target cells served as a positive control for maximal target cell death. IFN-γ ELISA Culture supernatants were collected at 66 h at the end of the cytotoxicity co-culture and stored at −80 °C until analysis. IFN-γ concentrations were measured using the human IFN-γ ELISA MAX Deluxe kit (BioLegend) according to the manufacturer’s instructions. Inhibitor assays Primary human T cells were pre-treated with pharmacological inhibitors targeting distinct internalization routes for 30 min prior to LNP addition to characterize the endocytic pathway responsible for SM102-DD LNP uptake.34 Inhibitors used included EIPA (macropinocytosis inhibitor, 50 μM, MedChemExpress, Catalog Number: HY-101840), chlorpromazine (clathrin-mediated endocytosis inhibitor, 20 μM, Fisher Scientific, Catalog Number: AAJ6365909), MβCD (cholesterol-dependent pathway inhibitor, 5 mM, Sigma-Aldrich, Catalog Number: C4555), Dyngo-4a (dynamin inhibitor, 30 μM, ApexBio, Catalog Number: B5997), and nystatin (caveolae-mediated endocytosis inhibitor, 50 μM, Sigma-Aldrich, N4014). Following pre-treatment, inhibitor-containing media was replaced with fresh media containing SM102-DD LNPs encapsulating eGFP mRNA at 400 ng per 60,000 cells. Cells were incubated for 24 h and then analyzed by flow cytometry for eGFP expression and viability. Cy5-labeled mRNA uptake assay EZ-Cap Cy5-eGFP mRNA (ApexBio) was encapsulated in SM102-DD, SM102-DOTAP5, and SM102-DSPC LNPs and delivered to primary human T cells at 400 ng per 60,000 cells as described previously.71 Cells were analyzed by flow cytometry 24 h post-transfection to measure Cy5 signal (uptake) and eGFP expression (translation) simultaneously using the Attune NxT Flow Cytometer. Flow data were analyzed using FlowJo v.10.10.0. Confocal imaging Primary human T cells were treated with SM102-DD, SM102-DOTAP5, and SM102-DSPC LNPs at 400 ng per 60,000 cells to assess endolysosomal trafficking. LNPs were fluorescently labeled by incorporating 1% molar ratio DiI (Thermo Fisher Scientific) into the lipid phase prior to formulation with the aqueous phase. At 4 h post-transfection, cells were incubated with LysoTracker Green DND-26 (Thermo Fisher Scientific) at 100 nM and Hoechst 33,342 at 1 μg/mL for 30 min to label lysosomes and nuclei, respectively, as described by Ma et al.71 Cells were then imaged live at 5 h post-transfection using a Zeiss LSM 800 Inverted Microscope at 20× magnification. Colocalization of DiI-labeled LNPs with LysoTracker signal was assessed to evaluate endosomal trafficking across formulations. Images were analyzed using ImageJ software with the Just Another Co-localization Plugin (JACoP) tool to calculate PCC values for each image as a measure of colocalization between DiI-labeled LNPs and LysoTracker signal.71 Data and code availability Data are available upon request. Acknowledgments The authors would like to thank the members of the Saha and Murphy labs for helpful discussion and comments on the manuscript. The authors would also like to thank the University of Wisconsin Carbone Cancer Center Flow Cytometry Laboratory, supported by P30 CA014520, for use of its facilities and services. The authors gratefully acknowledge the use of facilities and instrumentation in the UW-Madison Wisconsin Center for Nanoscale Technology. The center (wcnt.wisc.edu) is partially supported by the Wisconsin Materials Research Science and Engineering Center (NSF DMR-2309000) and the University of Wisconsin-Madison. Illustrations were created in part using BioRender. The authors also acknowledge funding from NIH R01 CA278051 and R35 GM119644, the NSF Engineering Research Center (ERC) for cell manufacturing technologies (CMaT, EEC 1648035), and the UW-Madison Office of the Vice Chancellor for research and graduate education with funding from the Wisconsin Alumni Research Foundation. All human studies were conducted using a University of Wisconsin-Madison IRB-approved protocol (2018-0103, approved 08/31/23). Graphical abstract and cartoons in this paper were created using BioRender. Author contributions A.R. performed and analyzed all electroporation and LNP studies; A.R. and G.S. isolated T cells from LRS cones and performed the IVTs for mRNA synthesis; A.R., W.L.M., and K.S. wrote the manuscript with input from all authors; W.L.M. and K.S. supervised the research. All authors read and approved the final manuscript. Declaration of interests K.S. receives honoraria for advisory board membership for Andson Biotech and Bharat Biotech Therapeutics. W.L.M. is a co-founder and equity holder in Stem Pharm, Inc. and Dianomi Holdings, Inc. A.R., K.S., and W.L.M. have disclosed portions of this work to the Wisconsin Alumni Research Foundation for patent filing. Footnotes Supplemental information can be found online at https://doi.org/10.1016/j.omtn.2026.103081. Contributor Information William L. Murphy, Email: wlmurphy@ortho.wisc.edu. Krishanu Saha, Email: ksaha@wisc.edu. Supplemental information Document S1. Figures S1–S7 and Table S1 Document S2. Article plus supplemental information References Associated Data Supplementary Materials Video S1. uPAR293 treated with DMSO Time-lapse fluorescence imaging using IncuCyte of uPAR293 target cells treated with DMSO and green IncuCyte Cytotox Dye. Video S2. uPAR293 Time-lapse fluorescence imaging using IncuCyte of uPAR293 target cells treated with Green IncuCyte Cytotox Dye. Video S3. uPAR293 co-cultured with EO115-CAR T cells Time-lapse fluorescence imaging using IncuCyte of uPAR293 target cells co-cultured with electroporated T cells expressing uPAR CAR mRNA at a 1:1 effector-to-target ratio, and treated with green IncuCyte Cytotox dye, showing progressive accumulation of green fluorescent signal indicative of CAR-mediated cytotoxicity over time. Video S4. uPAR293 co-cultured with SM102-DD-CAR T cells Time-lapse fluorescence imaging using IncuCyte of uPAR293 target cells co-cultured with SM102-DD-transfected T cells expressing uPAR CAR mRNA at a 1:1 effector-to-target ratio, and treated with green IncuCyte Cytotox dye, showing progressive accumulation of green fluorescent signal indicative of CAR-mediated cytotoxicity over time. Document S1. Figures S1–S7 and Table S1 Document S2. Article plus supplemental information Data Availability Statement Data are available upon request.