Engineering “off-the-shelf” TCR-T cells: A transient mRNA platform for balanced alloreactivity and functionality Authors: Shan He, Yirong Dai, Qi Chen, Adeline Chia, Joey Lim, Shou Kit Hang, Joey Aw, Andrea Pavesi, Rajneesh Kumar, Thinesh Lee Krishnamoorthy, Antonio Bertoletti, Anthony Tanoto Tan Published: 2026 ABSTRACT AbstractAdoptive T cell therapy using T cell receptor (TCR)-engineered T (TCR-T) cells shows promise for the treatment of cancer. Our lab employs mRNA electroporation to redirect T cell specificity toward HBV antigens expressed on HBV-related HCC, while exploiting mRNA’s transient nature to limit toxicity. Here, we first demonstrated that peripheral blood mononuclear cells (PBMCs) from solid cancer patients are inferior manufacturing sources compared with healthy donors, showing reduced manufacturability and lower naive T cell frequencies. Since these limitations necessitate shifting to allogeneic “off-the-shelf” TCR-T products from healthy donors, we developed a non-gene-editing platform to produce such TCR-T cells. We addressed host-versus-graft rejection by combining transient tacrolimus-induced immunosuppression with immunosuppressive drug-resistant armored (IDRA) TCR-T cells. To reduce the risk of graft-versus-host disease, we optimized manufacturing using an IL-2/4/7 cytokine cocktail. The resulting TCR-T cells exhibited reduced alloreactivity, a less NK-like phenotype, and elevated IL-4/IL-10 secretion, while maintaining robust antigen-specific cytotoxicity and migratory capacity. This study provides an in vitro proof-of-concept for a feasible, multi-layered strategy to produce allogeneic TCR-T cell therapies, warranting further preclinical validation.Keywords: allogeneic TCR-T cells, off-the-shelf cell therapy, mRNA electroporation, host-versus-graft rejection, graft-versus-host disease, tacrolimus resistance, immunosuppressive drug-resistant armored T cells, IDRA, IL-2/4/7 cytokine cocktail, hepatitis B virus-related hepatocellular carcinoma Introduction Adoptive T cell therapy using T cell receptor (TCR)-engineered T (TCR-T) cells holds significant promise for treating both infectious diseases and cancer—yet its clinical translation remains severely constrained by the prohibitive cost and complexity of autologous cell manufacturing.1,2,3 Each patient requires a bespoke production process: T cells must be harvested, expanded, and engineered individually, driving up costs and extending diagnosis-to-treatment times to a degree that is clinically untenable for many patients. Compounding this, T cells sourced from late-stage, heavily pre-treated cancer patients are frequently of poor quality, and this compromised T cell fitness can result in manufacturing failure. Our lab has nonetheless advanced this approach for patients with hepatitis B virus-related hepatocellular carcinoma (HBV-HCC), engineering expanded autologous T cells via mRNA electroporation to redirect their specificity toward HBV-associated tumor antigens.4,5 This strategy exploits the labile nature of exogenously introduced mRNA to temporally limit TCR-T cell function—a built-in safety feature—while multi-dose infusion regimens allow precise dosage control. Despite demonstrating meaningful anti-tumor efficacy in some patients,3,6,7,8,9 the autologous paradigm faces fundamental logistical and biological barriers that have severely reduced the number of HBV-HCC patients that we have so far treated. To overcome these obstacles, the field is moving toward allogeneic, “off-the-shelf” T cell therapy products derived from healthy donors—enabling scalable, cost-effective manufacturing and timely patient access.10 However, allogeneic cell sources introduce two critical hurdles: host-versus-graft (HvG) rejection, and graft-versus-host (GvH) disease, where infused T cells attack the recipient’s healthy tissue. Genetic modification strategies to reduce alloreactivity,10 while widely explored, add manufacturing complexity, raise safety concerns,11,12 and are incompatible with our transient mRNA workflow. A more feasible approach combines temporary recipient immunosuppression with immunosuppressant-resistant T cells to control HvG reaction.13,14 We adopted a strategy of conferring tacrolimus resistance by overexpressing a mutant CnB in the presence of intact endogenous wild-type CnB established by Brewin et al.15,16 Applying this strategy, we engineered an immunosuppressive drug-resistant armored (IDRA) TCR-T cell product transiently resistant to tacrolimus—a standard-of-care immunosuppressant in liver transplant patients—through co-electroporation of T cells with mRNAs encoding both the TCR and a mutated form of calcineurin B (CnB), tacrolimus’s target protein.17 This creates a protected window during which tacrolimus-induced recipient immunosuppression suppresses HvG rejection while sparing the engineered T cells. Here, we developed a non-gene-editing platform to produce “off-the-shelf” TCR-T cells. We first demonstrated the necessity of such allogeneic therapies by showing the inferiority of patient-derived peripheral blood mononuclear cells (PBMCs) for autologous therapy production. We then show that combining tacrolimus-induced immunosuppression with tacrolimus-resistant IDRA TCR-T cells overcomes HvG rejection in vitro. To address the parallel threat of GvH disease, we optimized our cell manufacturing process using an IL-2/4/7 cytokine cocktail. We demonstrate that expanding these IDRA TCR-T cells in this specific cytokine environment reduces their alloreactivity in vitro without compromising their robust anti-tumor cytotoxicity, ultimately yielding a balanced allogeneic product. Our in vitro findings establish a proof-of-concept for an allogeneic TCR-T cell therapy approach that warrants further preclinical validation. Results PBMCs from solid cancer patients exhibit impaired manufacturability of TCR-T cell production We compared the ability of PBMCs from healthy donors and cancer patients to generate TCR-T cells for clinical use. PBMCs of 21 solid cancer patients (Table 1), and 12 healthy donors were collected and different parameters across multiple stages of our manufacturing process were measured (Figure 1A). Although this patient cohort encompassed highly diverse clinical backgrounds and treatment regimens, the observed T cell manufacturing defects did not correlate with any specific clinical history or prior therapy. Solid cancer patients yielded significantly fewer extractable PBMCs per milliliter of blood (Ficoll recovery rate, Figure 1B) and exhibited lower tolerance toward cryopreservation (Thaw recovery rate, Figure 1B), showing reduced cell viability after one round of freeze-and-thaw cycle. The rate of T cell expansion following anti-CD3 stimulation follows a similar pattern (expansion fold change, Figure 1B). With few exceptions, patient-derived T cells frequently failed to expand to numbers sufficient for TCR-T cell engineering. This reduced fitness extended to the engineering phase (EP survival rate, electroporation of mRNA encoding a TCR specific for HBV-S183/HLA-A∗02, Figure 1B). Patient-derived activated T cells displayed significantly lower survival than healthy donor-derived T cells when quantified 18 h post-electroporation and yielded a final product with reduced frequencies of CD8+ cytotoxic T cells (%CD8+/CD3+, Figure 1B). Despite the evidence of lower T cells general fitness, TCR expression at single T cell level was identical between the TCR-T cells derived from the two groups (%TCR+/CD8+, Figure 1B). The final HBV-specific TCR-T products derived from cancer patients and healthy donors also exhibited comparable IFN-γ+ and TNF+ frequencies and direct cytotoxicity against HepG2/2.2.15 targets (Figures 1C and 1D). Table 1Clinical history of solid cancer patientsPatient IDAgeGenderEthnicityPrimary cancerNumbers of cancer treatments receivedaSurgerybAblationcEmbolizationdRadiotherapyeChemotherapyfTargeted therapygImmunotherapyTC0170MChineseHCC04150010TC0264FChineseStomach0000003TC0369MChineseLiver00100028TC0467MChineseLiver0400000TC0568MChineseCutaneous SCC100351601TC0678FEurasianBreast0000020TC0771FChineseLiver1031015TC0870FNAEndometrial10001211TC0958FChineseBreast0000205TC1065FChineseBreast0000201TC1169MChineseLiver1000000TC1263FChineseLiver0000000TC1362FChineseLiver0000000TC1462FChineseLiver0010000TC1571MChineseLiver01000018TC1674MChineseLiver00000115TC1774FChineseLiver0220000TC1880MChinesePancreas0000000TC1972FChineseLiver1130000TC2075FChinesePancreas0000000TC2173FChineseColorectal2200000Each treatment category includes the following treatments.aLap resection, liver resection, parotidectomy with parapharyngeal space dissection and R selective neck dissection, TAHBSO PLND PAND (total abdominal hysterectomy with bilateral salpingo-oophorectomy, pelvic lymph node dissection, para-aortic lymph node dissection).bMicrowave ablation (MWA), radiofrequency ablation (RFA).cTransarterial chemoembolizaiton (TACE), Yttrium-90 Radioembolization (Y90).dConcurrent chemoradiotherapy (CCRT), stereotactic body radiation therapy (SBRT).eCarboplatin, Cisplatin, Docetaxel, Paclitaxel, Paclitaxel (Abraxane).fLenvatinib, Letrozole, Megace, Palbociclib.gAtezolizumab, Atezolizumab/Bevacizumab (Combination), Bevacizumab, Enhertu (Trastuzumab deruxtecan), Nivolumab, Pembrolizumab, Pertuzumab, PHESGO (Fixed-dose combination of Pertuzumab, Trastuzumab, and Hyaluronidase), Transtuzumab, TCR-T cell therapy. Clinical history of solid cancer patients Each treatment category includes the following treatments. Lap resection, liver resection, parotidectomy with parapharyngeal space dissection and R selective neck dissection, TAHBSO PLND PAND (total abdominal hysterectomy with bilateral salpingo-oophorectomy, pelvic lymph node dissection, para-aortic lymph node dissection). Microwave ablation (MWA), radiofrequency ablation (RFA). Transarterial chemoembolizaiton (TACE), Yttrium-90 Radioembolization (Y90). Concurrent chemoradiotherapy (CCRT), stereotactic body radiation therapy (SBRT). Carboplatin, Cisplatin, Docetaxel, Paclitaxel, Paclitaxel (Abraxane). Lenvatinib, Letrozole, Megace, Palbociclib. Atezolizumab, Atezolizumab/Bevacizumab (Combination), Bevacizumab, Enhertu (Trastuzumab deruxtecan), Nivolumab, Pembrolizumab, Pertuzumab, PHESGO (Fixed-dose combination of Pertuzumab, Trastuzumab, and Hyaluronidase), Transtuzumab, TCR-T cell therapy. PBMCs from solid tumor patients are inferior sources for TCR-T cell manufacturing, exhibiting reduced naive CD8+ cell and elevated terminally differentiated effector memory CD8+ cell frequencies compared with those from healthy individuals (A) Schematics of the TCR-T cell manufacturing process, with stages compared between PBMCs from healthy donors and solid cancer patients highlighted in red. The experiments were always performed with patient and healthy donor-derived samples concurrently to avoid batch effects. (B) Bar chart showing manufacturing parameters compared between healthy donors (HC, n = 12) and solid tumor patients (CP, n = 21). Each dot represents individual values measured from each CP donor normalized against the average of HC donors in each parameter group. The bars show respectively the number of cells extracted per mL of blood using Ficoll-Paque, percentage of viable cells after one freeze-thaw cycle, fold change in viable cell number after eight days of expansion (n = 20 for CP, PBMCs from one patient did not survive cryopreservation), percentage of viable cells 24 h after HBV-S183 TCR mRNA electroporation relative to baseline (n = 20 for CP), percentage of CD8+ cells under the CD3+ gate 18 h after electroporation (n = 12 for HC, n = 18 for CP). The sample size for CP reduced to 18 since expanded T cells from two patients did not survive electroporation, and percentage of S183 TCR+ T cells under the CD8+ gate. S183 TCR positivity was determined by Dextramer staining (n = 12 for HC, 18 for CP). Significance was calculated using Mann-Whitney U test. Two outliers in the expansion fold change bar were labeled in red and excluded in the significance calculation. ∗∗p < 0.01, ∗∗∗p < 0.001. (C) Bar charts showing the percentage of IFN-γ ICS-positive and TNF ICS-positive cells under the CD8+ gate. The HBV-S183 TCR-T cells were cocultured overnight with HepG2/2.2.15 target cells at a 1:1 effector-to-target ratio for stimulation before ICS was performed (n = 12 for HC, 18 for CP). Significance was calculated using Mann-Whitney U test. (D) Normalized Cell Index curves generated by xCELLigence RTCA showing cell indices of 1 × 105 HepG2/2.2.15 cells in each RTCA E-Plate after 24 h of resting, with 1 × 105 CD8+TCR+ TCR-T cells added at the 24-h time point. All indices were normalized to 1 at the 24.5-h time point. The gray curve represents wells where C18 TCR-T cells, which does not recognize HepG2/2.2.15 cells, were added. The blue curve represents wells where TCR-T cells manufactured from healthy donors (n = 4) were added, while the maroon curve represents wells were TCR-T cells manufactured from solid cancer patients (n = 6) were added. (E and F) Bar charts showing percentages of naive and terminal effector memory CD8+ T cells defined by expression of CCR7 and CD45RA within the PBMC compartment of healthy donors (n = 13) and solid cancer patients (n = 9). Significance was calculated using Mann-Whitney U test. Phenotypic analysis revealed that patient CD8+ T cells have significantly fewer naive T (Tn) cells (CCR7+CD45RA+) and more terminally differentiated effector memory (TemRA) cells than healthy donors (Figures 1E and 1F). Given that Tn cells possess superior proliferative capacity and resist exhaustion,18,19,20 we hypothesized that their scarcity impaired manufacturability. Indeed, sorting and reconstituting PBMC fractions demonstrated that CD8+ Tn cells are the primary drivers of α-CD3-mediated in vitro expansion, yielding functional TCR-T cells comparable with unsorted controls (Figure S1). Taken together, PBMCs of cancer patients are a poor starting material for TCR-T cell manufacturing due to deficits in abundance, expandability, and survival after TCR electroporation in comparison with PBMCs of healthy donors, while the overall function of the final product does not show differences between products derived from cancer patients and healthy donors. Tacrolimus suppresses alloreactivity in vitro with minimal impact on HBV-specific IDRA TCR-T cell function We have already demonstrated the successful production of immunosuppressive drug-resistant armored (IDRA) TCR-T cells that maintain their function in the presence of tacrolimus.17 To mimic the adoptive transfer of IDRA TCR-T cells into an allogeneic recipient, carboxyfluorescein succinimidyl ester (CFSE)-labeled HBV S183 IDRA TCR-T were mixed with allogeneic recipient PBMCs (80%), unpulsed and pulsed with HBV-S183 peptide (20%, non-pulsed as control; Figure 2A). Experiments were performed in the presence or absence of clinically relevant amounts of tacrolimus (10 ng/mL). Activation of IDRA TCR-T cells, which was directly analyzed with intracellular cytokine staining (ICS), was shown to be mostly preserved in the presence of tacrolimus and was identical when mixed with autologous or allogeneic PBMCs (Figure 2B). Combining tacrolimus-mediated immunosuppression with tacrolimus-resistant IDRA TCR-T cells provides a feasible strategy to mitigate graft rejection (A) A schematic representation of the MLR performed to test the function of IDRA TCR-T cells in an allogeneic environment in the presence of tacrolimus. (B) Bar charts showing the percentage of CD8+ TCR-T cells that are IFN-γ ICS-positive under different MLR experimental conditions (n = 2) and representative flow cytometry plots. TCR-T cells were produced from each of the two donors and were mixed with either their own PBMCs (autologous) or the other donor PBMCs (allogeneic). IDRA TCR-T cells were mixed with PBMCs at a ratio of 1:6. (C) A schematic representation of the MLR performed to measure alloreactivity generated by TCR-T cell infusion. (D) Bar charts showing concentrations of different cytokines measured from the MLR are shown in (C). Each dot represents one donor/recipient combination. T cells were expanded from the same donors’ PBMCs from which whole blood was taken. The expanded T cells and whole blood were mixed in a matrix fashion, generating three autologous combinations (n = 3) and six allogeneic combinations (n = 6). Significance was calculated using t test. (E) Bar charts showing concentrations of different cytokines generated from MLR with IDRA TCR-T cells done in the same way as in (D) (n = 3 for autologous and n = 6 for allogeneic). Each dot represents one donor/recipient combination. Significance was calculated using t test. (F) Bar charts displaying the relative levels of IFN-γ and IL-2 in allogeneic co-cultures (n = 6). Data are presented as the percentage of cytokine concentration measured in tacrolimus-treated conditions relative to their respective untreated controls. To assess tacrolimus-mediated suppression of alloreactivity, we added non-electroporated donor T cells into allogeneic whole blood. After overnight incubation in the presence or absence of tacrolimus, the supernatant was collected for cytokine quantification using Ella (Figure 2C). This approach was used because the protein transport inhibitor (Brefeldin A) used in ICS would inhibit cytokine secretion and interfere with the assessment of alloreactive responses (Figure 2B). We used plasma concentrations of multiple cytokines (IFN-γ, TNF, IL-2, etc.) as a surrogate measurement for alloreactivity induced by T cell infusion. We demonstrated that alloreactivity triggered by T cell infusion can be detected with the Ella assay in the form of cytokine release as was indicated by the elevated concentrations of key cytokines when allogeneic T cells were added into recipient blood compared with when autologous T cells were added (Figure 2D). More importantly, we were able to show that such alloreactivity can be significantly suppressed with tacrolimus indicated by the significant drop in the concentrations of various cytokines when the whole blood is treated with tacrolimus compared with the untreated group (Figure 2D). As expected, when tacrolimus-resistant IDRA TCR-T cells were used in the mixed lymphocyte reaction (MLR), tacrolimus failed to suppress IFN-γ and IL-2 secretion as effectively as it did with non-electroporated T cells (Figures 2E and 2F). The ability of IDRA TCR-T cells to evade tacrolimus-mediated immunosuppression and potentially cause GvH complications highlighted the need for a strategy to mitigate the inherent GvH potential of the allogeneic product. Manufacturing IDRA-TCR-T cells with lower GvH potential To reduce GvH potential, healthy donor-derived IDRA TCR-T cells were expanded using conventional high-dose IL-2 (600 IU/mL), IL-4/7,21 or IL-4/7 supplemented with low-dose IL-2 (20 IU/mL) (Figure 3A). Our choice of an IL-4/7-based cytokine cocktail was informed by a randomized phase 1/2 trial in which allogeneic SARS-CoV-2-specific T cells expanded with IL-4/7 were infused into 60 patients with severe COVID-19 without any signs of GvHD.19 We adopted the same IL-4/7 concentrations and assessed the manufacturability of mRNA-electroporated TCR-T cells generated with this protocol, using our conventional high-dose IL-2 protocol as a reference. Expanding TCR-T cells with IL-2/4/7 results in a product that inherently possesses less GvHD potential with comparable manufacturing parameters compared with IL-2 product (A) A schematic representation of the MLR performed to measure the GvH potential of TCR-T cell products expanded with different cytokine cocktails (a. 600 IU/mL IL-2, b. 800 IU/mL IL-4 and 20 ng/mL IL-7, c. 20 IU/mL IL-2, 800 IU/mL IL-4, and 20 ng/mL IL-7). (B) Key TCR-T cell production parameters measured for TCR-T cells expanded with the three different cytokine cocktails described in (A). n = 10 for Expansion rate, and n = 8 for the other parameters. Significance was calculated using t test. (C) Bar charts showing concentrations of key cytokines measured with Ella from the MLR are shown in (A). Each dot represents one donor/recipient combination. IDRA TCR-T cells were expanded from the same donors’ PBMCs from which whole blood was taken (n = 3). The expanded T cells and whole blood were mixed in a matrix fashion, generating three autologous combinations (n = 3) and six allogeneic combinations (n = 6). Significance was calculated using t test. The number of T cells expanded from PBMCs with IL-4/7 was lower than with IL-2 and IL-2/4/7, the latter two of which showed comparable rates (Figure 3B). A similar trend was observed in electroporation survival rates, defined as the percentage of viable T cells (trypan blue-negative) 24 h post-electroporation relative to the total number of expanded T cells used. No significant difference in TCR expression efficiency was observed, as the S183 TCR was expressed at nearly 100% across all three T cell types. Because T cells expanded with IL-4/7 alone showed lower expansion and electroporation survival than those expanded with high-dose IL-2 (Figure 3B), we supplemented the IL-4/7 cocktail with low-dose IL-2 (IL-2/4/7), which restored these manufacturing parameters to levels comparable with the IL-2 condition while preserving the reduced inflammatory cytokine secretion characteristic of IL-4/7 expansion. Notably, T cells expanded with IL-4/7 or the IL-2/4/7 cytokine cocktail exhibited more balanced CD8+/CD4+ ratio compared with those expanded with high-dose IL-2 (Figure 3B). Then, to mimic the adoptive transfer of allogeneic IDRA TCR-T cells in patients treated with tacrolimus, IDRA TCR-T cells expanded with different cytokine cocktails were mixed with allogeneic recipient blood in the presence of tacrolimus. As control, IDRA TCR-T cells were also mixed with autologous blood. Note that in the presence of tacrolimus, the cytokine levels measured from the plasma after overnight incubation should have solely come from the IDRA TCR-T cells and thus can be indicative of their GvH potential. Figure 3C shows that when mixed in allogeneic blood, IL-4/7 and IL-2/4/7 derived IDRA TCR-T cells produced significantly less IFN-γ, IL-2, and TNF than their IL-2 expanded counterparts. Meanwhile, the IL-4/7 and IL-2/4/7-derived IDRA TCR-T cells also produced more IL-10 and exhibited a constitutive expression of Th2 cytokine IL-4. IL-2/4/7 TCR-T cells are characterized by reduced susceptibility to bystander activation and elevated IL-4 and IL-10 secretion We further characterized the functionality of IL-2/4/7 TCR-T cells and compared them with IL-2 TCR-T cells by conducting NanoString nCounter with the Immunology Panel and Olink Target 48 Cytokine analyses on both cell types. The TCR-T cells were either left unstimulated or stimulated overnight with anti-CD3/28 Dynabeads. Following this, the cell culture supernatant was collected for Olink analysis, and mRNA was extracted for NanoString analysis (Figure 4A). Transcriptomic and secretomic analyses reveal that IL-2/4/7 cells are less NK-like and express distinctively more IL-4 and IL-10 upon stimulation compared with IL-2 cells (A) Schematics of the experimental flow for NanoString nCounter Immunology Panel (transcriptomic) and Olink Target 48 Cytokine (secretomic) characterization of IL-2 and IL-2/4/7 TCR-T cells. The TCR-T cells were either unstimulated or stimulated with anti-CD3/28 Dynabeads. Each experimental condition involved 1 × 106 TCR-T cells and 5 × 105 anti-CD3/28 Dynabeads (stimulated). (B) Heatmap showing the NanoString nCounter Immunology Panel results. The z-scores of the mRNA count were calculated and used to plot the heatmap. Clustering was performed with Euclidean distance as the similarity metric. Each row represents one donor’s TCR-T cells (n = 3), and each column represents one gene. (C) Volcano plot of the NanoString nCounter Immunology Panel results (n = 3) showing differential gene expression of unstimulated and stimulated IL-2 and IL-2/4/7 TCR-T cells compared separately with key DEGs labeled. The x axis represents the log2 fold change, and the y axis represents the −log10 adjusted p value. Genes upregulated in IL-2 cells are labeled blue, and the ones upregulated in IL-2/4/7 cells are labeled red. The cut-off for fold change is 1-fold and that for p value is 0.05. (D) Bar chart and representative ELISpot images showing IFN-γ secretion following cytokine-induced activation of IL-2 and IL-2/4/7 TCR-T cells. Cells were stimulated with IL-12 (10 ng/mL), IL-15 (100 ng/mL), and IL-18 (50 ng/mL). Significance was calculated using t test (n = 8). (E) Heatmap showing the Olink Target 48 Cytokine results plotted with log2 transformed absolute concentration (pg/mL). Cytokines with concentrations exceeding the upper limit of quantification (ULOQ) were assigned with their respective ULOQ values for visualization (n = 3). (F) Bar charts showing key cytokine concentrations selected from (E). Significance was calculated using t test. We showed through the NanoString analysis that, across all three donors, IL-2 and IL-2/4/7 TCR-T cells exhibited highly comparable transcriptomic profiles, with no distinct differences observed in either the unstimulated or stimulated groups (Figure 4B). Despite this global similarity, we identified a discrete set of differentially expressed genes, most notably the NK cell markers KLRC1, KLRC2, and KLRD1 (Figure 4C). KLRC1 and KLRC2 were consistently upregulated in IL-2 TCR-T cells compared with IL-2/4/7 TCR-T cells, irrespective of stimulation status, whereas KLRD1 upregulation in IL-2 TCR-T cells was observed only upon stimulation. This elevated expression of NK-cell markers may be associated with NK-like features.22 Indeed, IL-2 TCR-T cells were significantly more responsive to NK-activating cytokines (IL-12, IL-15, and IL-18) via IFN-γ ELISpot (Figure 4D). Olink analysis also revealed largely similar cytokine secretion profiles between IL-2 and IL-2/4/7 TCR-T cells with a few key differences (Figures 4E and 4F). While IL-2 and IL-2/4/7 TCR-T cells secreted comparable levels of most measured cytokines including IFN-γ upon stimulation (Figure 4E), IL-2/4/7 TCR-T cells produced markedly higher levels of IL-4 and IL-10, several-fold greater than IL-2 TCR-T cells (Figure 4F). IL-2/4/7 TCR-T cells were also found to produce more LTA, marginally more IL-6, and to constitutively produce IL-7. This cytokine secretion profile indicates that, upon TCR-mediated activation, IL-2/4/7 TCR-T cells secrete significantly higher levels of anti-inflammatory cytokines, while maintaining pro-inflammatory cytokine production comparable with IL-2 cells. The elevated production of IL-4 and IL-10 from IL-2/4/7 TCR-T cells was also consistent with the results of the MLR measuring GvH potential of different types of TCR-T cells (Figure 3C). Functional analysis of IL-2/4/7-derived HBV-specific TCR-T cells Having generated IL-2/4/7 TCR-T cells that are intrinsically less prone to activation by allogeneic cells, we next sought to comprehensively evaluate functional characteristics and compare it against the IL-2 TCR-T cells. To this end, we first conducted a 2D cytotoxicity assay with the xCELLigence real-time cell analysis (RTCA) technology with HepG2/2.2.15 cell as targets. The results showed that TCR-T cells expanded with the IL-2/4/7 cocktail exhibited robust cytotoxicity against HepG2/2.2.15 target cells, comparable with the standard IL-2 condition (Figures 5A and 5B). In contrast, cells expanded with IL-4/7 alone displayed significantly slower cytotoxicity. These data indicate that while IL-2/4/7 TCR -T cells maintained their cytotoxic capacity despite reduced GvH potential. TCR-T cells expanded with IL-2 and IL-2/4/7 cells showed similar antigen-specific cytotoxicity against tumor cell lines in both 2D and 3D cytotoxicity assays (A) Normalized cell index curves generated by xCELLigence RTCA showing cell indices of 1 × 105 HepG2/2.2.15 cells in each RTCA E-Plate after 24 h of resting, with 1 × 105 CD8+TCR+ TCR-T cells added at the 24-h time point. All indices were normalized to 1 at the 24-h time point. The gray curve represents wells where C18 TCR-T cells, which does not recognize HepG2/2.2.15 cells, were added. The three colored curves represent wells where S183 TCR-T cells manufactured with different cytokine cocktails were added. The experiment was done with TCR-T cells manufactured from four healthy donors’ PBMCs (n = 4). Each condition was done in duplicates on the RTCA E-Plate. (B) Bar chart showing the area under the curve (AUC) of the xCELLigence RTCA normalized cell index curves (n = 4). Significance was calculated using t test. (C) Representative confocal microscopy images of 3D cytotoxicity assays performed with the AIM Biotech idenTx devices and imaged with the Operetta CLS high-content analysis system. Different cell types were assigned with different pseudo colors with tumor target cells in green, dead cells in red, and TCR-T cells in yellow. HepG2 PreS1 GFP cells expressing the HBV-S183 epitope were used as the tumor target cells. Left: images of idenTx devices where C18 IL-2 TCR-T (does not recognize HepG2 PreS1 cells) cells were added. Middle: images of idenTx devices where S183 IL-2 TCR-T cells were added. Right: images of idenTx devices where S183 IL-2/4/7 TCR-T cells were added. Baseline images were taken immediately before TCR-T cells were added. (D) Bar charts showing a summary of the number of dead tumor target cells (GFP+DRAQ7+) in the 3D cytotoxicity assay. Left images shows data from devices where HepG2 cells (no HBV-S183 epitope expression) were used as tumor target cells and S183 TCR-T cells were added to the side channel. Middle images shows data from devices where HepG PreS1 cells were used as tumor targets and C18 TCR-T cells were added to the side channel. Right images shows devices where HepG PreS1 cells were used as tumor targets, and S183 TCR-T cells were added to the side channel. TCR-T cells manufactured from PBMCs from four healthy donors were used in this experiment (n = 4). Each condition was done in duplicates. (E) Bar chart showing numbers of killed tumor target cells calculated by subtracting the baseline number of dead tumor target cells from the number 24 h after the addition of TCR-T cells (n = 4). Significance was calculated using t test. (F) Bar chart showing the number of TCR-T cells (CMRA+) present in the gel space 24 h after the addition of them in the side channel (n = 4). Significance was calculated using t test. (G) Truncated violin plot showing the y axis position of each identifiable TCR-T cell in relation to the edge of the gel space. n = 2,533 for IL-2. n = 2,337 for IL-2/4/7. We further evaluated intrinsic killing and migration capacity using a 3D microfluidic cytotoxicity assay.23 Confocal imaging revealed that TCR-T cells expressing the C18 TCR, which does not recognize HepG2 PreS1 GFP target cells, induced minimal target cell death despite substantial T cell infiltration into the gel (Figure 5C, left). In contrast, IL-2 and IL-2/4/7 S183 TCR-T cells induced a comparable and marked increase in the number of dead HepG2 PreS1 GFP cells, with a similar level of infiltration compared with C18 TCR-T cells (Figure 5C, middle and right). These results indicate that IL-2 and IL-2/4/7-expanded cells exhibit comparable antigen-specific cytotoxicity against tumor target cells in a 3D environment that requires active migration. Quantitative data from the 3D cytotoxicity assay were also obtained to enable a more precise comparison between the cytotoxic effects of IL-2 and IL-2/4/7-expanded TCR-T cells. To precisely quantify cytotoxicity, we measured the absolute number of dead target cells (GFP+DRAQ7+) in the co-culture system (Figure 5D). Significant killing was observed only in the presence of antigen-matched HepG2-PreS1 targets and specific S183 TCR-T cells, confirming the antigen specificity of the response. Crucially, aggregated data from these assays revealed that IL-2/4/7-expanded TCR-T cells induced target cell death at levels statistically indistinguishable from the standard IL-2 group (Figure 5E). We next evaluated and compared the migratory capacity of the TCR-T cells. We quantified the number of TCR-T cells present in each gel space as well as their y axis position relative to the gel edge on the side where TCR-T cells were injected. The data show that the IL-2 and IL-2/4/7 TCR-T cells differ minimally in both aspects (Figures 5F and 5G). Taken together, these data demonstrate equivalent cytotoxicity and migratory capacity of IL-2 and IL-2/4/7 TCR-T cells. Discussion The reduced fitness of patient-derived PBMCs for TCR-T cell manufacturing reinforces the rationale for using allogeneic cell sources. Patients’ PBMCs yielded fewer cells per milliliter of blood, had lower survival rates after cryopreservation and electroporation, and showed poorer T cell expansion. Extrapolating from these yield parameters, generating a single dose of 3.5 × 108 TCR-T cells (5 million cells/kg for a 70 kg recipient) requires ∼250 mL of healthy donor blood, whereas achieving the same yield from solid tumor patients requires over 3 L. This observation aligns with increasing literature describing T cell exhaustion and dysfunction in cancer patients, often as a result of chronic antigen exposure and prior therapies,24,25,26,27 which directly impacts the feasibility of autologous cell therapy production. This necessitates the use of healthy allogeneic cell sources to manufacture TCR-T cells. However, allogeneic TCR-T cells are inherently constrained by the risks of HvG rejection and GvH disease. To overcome the HvG reaction, we showed the feasibility of using tacrolimus to temporarily suppress alloreactivity while employing transiently tacrolimus-resistant IDRA TCR-T cells to preserve their effector functions. Our in vitro models demonstrated that tacrolimus significantly suppressed alloreactivity while minimally impacting the antigen-specific activation of IDRA TCR-T cells. One important aspect of this strategy is the creation of a transient therapeutic window, a strategy designed for enhanced safety. Tacrolimus was selected not only for its efficacy but also for its relatively fast pharmacokinetics,28,29 which allows for a defined, temporary period of recipient immunosuppression. Our mRNA-based engineering of IDRA TCR-T cells can be synchronized with this approach; the expression of both the TCR and the resistance-conferring mutated CnB is transient, lasting only as long as the mRNA persists. This ensures the T cells are active and protected only during this critical window, after which they will be cleared by the recovered host immune system. Because transgene expression in this platform is limited to approximately 72 h,5,17 concerns raised for permanently engineered chimeric antigen receptor (CAR)- or TCR-T cells, such as antigen-driven upregulation of surface receptor and the consequent risk of sustained activation or enhanced alloreactivity upon repeated antigen exposure, are unlikely to apply here, as the engineered cells lose their introduced specificity once the mRNA is degraded. However, our results also showed that alloreactivity mainly came from the graft, namely, the IDRA TCR-T cells. Since these cells can resist tacrolimus-mediated immunosuppression, their potential to induce GvH reactions constitutes an unresolved risk. To address this arm of alloreactivity-GvH reaction, we built on clinical work done by Papadopoulou et al. showing that an IL-4/7 cytokine cocktail could expand virus-specific T cells that did not induce GvHD in any of the 60 treated recipients.21 We therefore developed an IL-2/4/7 cytokine cocktail that generates TCR-T cells showing markedly lower production of IFN-γ, IL-2, and TNF, as well as higher production of IL-10, upon allogeneic stimulation compared with conventional high-dose IL-2-expanded TCR-T cells, all while exhibiting comparable production parameters and tumor-specific cytotoxicity. This cytokine profile shows reduced alloreactivity and a lower propensity to induce systemic cytokine release syndrome, a known driver of acute GvHD.30,31,32 Our transcriptomic and secretomic analyses resolved more nuanced differences between IL-2 and IL-2/4/7 TCR-T cells, which might provide a potential mechanistic explanation for this reduced alloreactivity. IL-2/4/7 TCR-T cells displayed a less NK-like phenotype marked by reduced expression of NK cell markers such as KLRC1 and KLRC2, which respectively encode NKG2A and NKG2C. These two molecules have been indicated to be involved in allorecognition by NK cells and have opposite functions—NKG2A being inhibitory and NKG2C being activating.22,33 The upregulation of both inhibitory and activating NK cell markers cannot directly account for the reduced alloreactivity of IL-2/4/7 TCR-T cells, hence it prompted us to examine their functional characteristics, which revealed that IL-2/4/7 TCR-T cells were markedly less responsive to NK-activating cytokines than IL-2 TCR-T cells. Furthermore, while maintaining comparable pro-inflammatory capacity, IL-2/4/7 cells produced markedly higher levels of the anti-inflammatory cytokines IL-4 and IL-10 upon TCR-mediated activation. Given their unresponsiveness to cytokine activation, these cells seem more specific to TCR-mediated activation than to bystander activation, a desirable feature for a cell therapy product. We therefore propose that this reduced propensity for cytokine-driven bystander activation may contribute to a lower GvH potential—a hypothesis that will require further mechanistic validation. While IL-4 and IL-10 have long been believed to be immunosuppressive and might affect T cell therapy efficacy,34,35,36,37 recent work suggests a more positive role. Studies have shown that CAR-T cells engineered to express IL-10 preserve anti-tumor function in the hostile tumor microenvironment (TME),38 and that the secreted IL-10 not only promotes proliferation and effector function of CAR-T cells, but also metabolically rejuvenates terminally exhausted CD8+ T (CD8+ TTE) cells.38,39 Other studies have shown that IL-4 also revitalizes and enriches functional CD8+ TTE cells in tumors by boosting their survival and anti-tumor efficacy through a metabolic reprogramming that enhances glycolysis and cellular NAD levels.40 Furthermore, single-cell atlas studies of long-term survivors have revealed that a type 2 functional state in CAR-T cells is strongly associated with 8-year remission and sustained persistence.41 These findings suggest that our IL-2/4/7 TCR-T cells are not disadvantaged by their IL-4 and IL-10 production; rather, they represent a more balanced and robust product. This unique cytokine profile may equip the cells to better resist the suppressive TME and sustain their anti-tumor activity, leading to a more durable clinical response. One limitation of this study is that the solid cancer patients and healthy donors recruited for the manufacturing comparison were not age-matched; the patient cohort was generally older. This age disparity is a significant confounding factor, as aging is a primary driver of immunosenescence—a process characterized by thymic involution, a shrinking pool of naive T cells, and the accumulation of terminally differentiated subsets.42 Consequently, the reduced manufacturing capacity observed in patient samples may partially reflect the natural decline in T cell fitness and plasticity associated with advanced age, in addition to tumor-induced dysfunction. A further limitation is that our platform was evaluated exclusively in vitro; the in vivo persistence, expansion, trafficking, and clearance kinetics of IDRA TCR-T cells under tacrolimus treatment—and the balance between transient immunosuppression, transient transgene expression, and host immune recovery—remain to be determined and will require dedicated preclinical studies. Taken together, we have developed and validated in vitro a strategy to manufacture an effective TCR-T cell therapy product for HBV-HCC patients that is marked by a unique multi-layered transient design. The temporary recipient immunosuppression, the transient expression from mRNA, and the clearance of the allogeneic T cells by the recovered host immune system constitute a self-regulating “hit-and-run” approach. Furthermore, by employing an IL-2/4/7 cytokine cocktail during manufacturing, we generated a cell product with intrinsically less GvH potential characterized by reduced bystander activation and elevated IL-4/IL-10 secretion without compromising its antigen-specific cytotoxic capacity. This work establishes a feasible, non-gene-editing platform that addresses the principal safety challenges of allogeneic T cell therapy, advancing the potential for a truly “off-the-shelf” treatment for patients with HBV-HCC. Materials and methods Patients Peripheral blood from solid cancer patients and healthy donors was acquired following written informed consent, under approval by the SingHealth Centralized Institutional Review Board (ECOS Ref. 2024-4056) and in accordance with Helsinki Declaration principles. Cells PBMCs were isolated via Ficoll-Paque (GE Healthcare) density gradient centrifugation and cryopreserved. HepG2/2.2.15 cells were cultured in DMEM (Thermo Fisher) with 10% FBS, non-essential amino acids (NEAA), 1 mM sodium pyruvate, penicillin/streptomycin (P/S), and 200 μg/mL G418. HepG2 and HepG2-PreS1-GFP cells were cultured in RPMI 1640 with 10% FBS, 0.5 mM sodium pyruvate, 20 mM HEPES, P/S, NEAA, L-glutamine, 5 μg/mL Plasmocin, and 5 μg/mL puromycin. Cells were maintained at 37°C/5% CO2. T cell expansion and mRNA electroporation PBMCs were cultured in AIM V (Gibco), 2% human AB serum, and 50 ng/mL anti-CD3 (OKT-3; eBioscience) for 7 days, supplemented with either: 600 IU/mL IL-2 (Miltenyi); 800 IU/mL IL-4 and 20 ng/mL IL-7 (IL-4/7); or the IL-4/7 cocktail plus 20 IU/mL IL-2 (IL-2/4/7). For IL-2 cultures, concentration was increased to 1,000 IU/mL on day 7. On day 8, T cells were electroporated with indicated mRNAs (S183, C18, mutant CnB) via the 4D-Nucleofector System (Lonza) and maintained overnight in AIM V with 10% AB serum and 100 IU/mL IL-2 following a previously described protocol.5,17 The α and β chains of each TCR (S183 or C18) were encoded within a single mRNA transcript separated by a P2A self-cleaving peptide sequence, while the mutant CnB was encoded on a separate mRNA transcript; the two mRNAs were co-electroporated simultaneously, as previously described for this IDRA platform.17 Transient expression of mutant CnB confers functional tacrolimus resistance for approximately 3 days, after which edited T cells revert to their native, drug-sensitive phenotype.17 T cell subpopulation sorting and functional assessment CD8+ T cells were sorted into naive (Tn), effector memory (Tem), and terminally differentiated effector memory (TemRA) subsets, then reconstituted with autologous PBMC fractions at original CD8+ percentages. Reconstituted cultures were expanded, electroporated with TCR mRNA, co-cultured with HepG2.2.15 cells, and assessed for IFN-γ and TNF-α expression via ICS. MLR and alloreactivity assays To evaluate alloreactivity and GvH potential, expanded donor T cells or IDRA TCR-T cells were co-cultured overnight with recipient whole blood to generate both autologous and allogeneic combinations. These co-cultures were performed in the presence or absence of 10 ng/mL tacrolimus (Sigma-Aldrich). Following incubation, plasma was collected, and secreted cytokines (including IFN-γ, TNF, and IL-2) were directly quantified using the Ella automated immunoassay system, serving as a surrogate measurement for alloreactivity. Real-time cytotoxicity assay (xCELLigence) Cytotoxicity was measured using the xCELLigence system (Agilent). HepG2/2.2.15 targets (1 × 105) were seeded in E-plates overnight. After 25–30 h, supernatant was replaced with AIM V (2% AB serum) containing TCR-T cells at a 1:1 E:T ratio. Killing was monitored for 45 h; cell indices were normalized to values measured immediately prior to T cell addition. 3D microfluidic assay and live imaging 3D assays utilized idenTx 40 microfluidic devices.23 HepG2-PreS1-GFP targets in collagen gel were polymerized in the central channel. Media was replaced with AIM V (2% AB serum, 100 IU/mlL IL-2) containing 3 μM of the cell-impermeable nuclear dye DRAQ7 (BioLegend). CellTracker Violet-labeled TCR-T cells (3 × 106 cells/mL) were introduced into side channels. Migration and cytotoxicity were monitored via confocal microscopy (Operetta CLS, Revvity). Target death was quantified by the change in DRAQ7 fluorescence intensity over 15 h. Flow cytometry and phenotypic analysis Cells were stained with a LIVE/DEAD Fixable Near-IR Dead Cell Stain Kit (Invitrogen), followed by surface staining with anti-human CD3, CD4, CD8, CCR7, and CD45RA antibodies (BioLegend), as well as a TCR Dextramer (Immudex) specific to the S183 TCR. For ICS, cells were fixed and permeabilized using Cytofix/Cytoperm buffer following the manufacturer’s protocols, then stained with anti-human IFN-γ and TNF-α antibodies. Data were acquired on a CytoFLEX S flow cytometer (Beckman Coulter) and analyzed using FlowJo software. Statistical analysis Statistical analysis was performed using GraphPad Prism. Differences between groups were evaluated using two-tailed t tests or Mann-Whitney U test. Data and code availability For original data, please contact anthony.tan@duke-nus.edu.sg. Acknowledgments This study was supported by the Singapore Translational Research (STaR) Investigator Award from the National Medical Research Council (NMRC), Singapore (grant no. MOH-001633/MOH-STaR24jan-0001). This study was approved by the SingHealth Centralized Institutional Review Board (CIRB Ref. 2024–4056) and was conducted in accordance with the principles of the Declaration of Helsinki; all participants provided written informed consent. Author contributions S.H. conceptualized the study, designed and performed the majority of the experiments, analyzed the data, and wrote the original draft of the manuscript. Y.D. processed the patient’s blood samples and assisted with various additional experiments throughout the study. Q.C. performed the initial functional assays evaluating IDRA cells in immunosuppressive and allogeneic environments. A.C. and J.L. conducted the early experiments demonstrating tacrolimus-mediated suppression of alloreactivity. S.K.H. assisted with the execution of the Ella assays. J.A. and A.P. provided technical support, supervision, and resources for the 3D microfluidic cytotoxicity assays. R.K. and T.L.K. recruited the solid cancer patients and provided the clinical samples. A.B. and A.T.T. conceptualized the study, supervised the overall research, and critically revised the manuscript. Declaration of interests The authors declare the following financial interests/personal relationships, which may be considered as potential competing interests: Andrea Pavesi reports a relationship with AIM Biotech Pte., Ltd., that includes board membership and equity or stocks. A.P. has a patent licensed to AIM Biotech PTE. LTD. A.B. is the Scientific Founder of Lion TCR Pte. Ltd., a biotech company developing T cell receptors for treatment of virus-related diseases and cancers. A.B. and A.T.T. are the shareholders of T cell Diagnostics Pte., Ltd., a biotech company providing contract services for the analysis of virus-specific T cell responses. Other authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Declaration of generative AI and AI-assisted technologies in the writing process During the preparation of this work, the author(s) used Claude to grammatically modify the manuscript. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the published article. Footnotes Supplemental information can be found online at https://doi.org/10.1016/j.omton.2026.201335. Contributor Information Antonio Bertoletti, Email: antonio@duke-nus.edu.sg. Anthony Tanoto Tan, Email: anthony.tan@duke-nus.edu.sg. Supplemental information Document S1. Figure S1 Document S2. Article plus supplemental information References Associated Data Supplementary Materials Document S1. Figure S1 Document S2. Article plus supplemental information Data Availability Statement For original data, please contact anthony.tan@duke-nus.edu.sg.