Mouse melanoma cell line B16F10, mouse colorectal carcinoma cell lines MC38, and HEK293T cells were cultured in DMEM medium supplemented with 10% FBS (Gibco), 100 U/ml penicillin and 100 μg/ml streptomycin). To generate B16F10 and MC38 cell line stably expressing human EGFRvIII, B16F10 and MC38 cells were transduced with a lentiviral vector encoding human EGFRvIII, firefly luciferase and puromycin N-acetyltransferase (pLV-EGFRvIII-P2A-FLuc-T2A-puro) and selected on 7.5 μg/ml puromycin (Beyotime). To generate B16F10 and MC38 cell line stably expressing human CD19, the cells were transduced with a lentiviral vector encoding truncated human CD19, luciferase and puromycin N-acetyltransferase (pLV-hCD19t-P2A-FLuc-T2A-puro) and selected on 7.5 μg/ml puromycin (Beyotime). The expression of human EGFRvIII and CD19 was evaluated by flow cytometry (CytoFLEX S, Beckman Coulter) using an anti-EGFRvIII antibody (abcam, EPR28380-83) and APC anti-human CD19 antibody (Biolegend, clone: HIB19).

Six-to eight-week-old female C57BL/6J mice were used for all studies and maintained in individually ventilated cages under 12 h light/dark cycles in a specific pathogen-free animal facility. All animal studies were performed in accordance with the National Institutes of Health guidelines for the use and care of live animals and approved by the Institutional Animal Care and Use Committee of the Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences (SIAT-IACUC-241226-YYS-WXC-A2900).

To generate murine CAR-T cells, a previously described CAR sequence targeting the human EGFRvIII and CD19 was used, which contains single-chain fragment (scFv) domain targeting EGFRvIII and CD19 (derived from antibody clone 139 and FMC63) (14, 15), a CD28 transmembrane domain, a CD28 costimulatory signal and a CD3ζ domain. based on the CAR sequence, a flag-tag was inserted between the signal peptide and the scFv domain. The plasmid MIGR1 was used as a retroviral vector (RV) backbone to construct pRV-EGFRvIII CAR and pRV-hCD19 CAR plasmids. The sequences were inserted in the MIGR1 plasmid between the EcoR I and Sal I restriction enzyme sites. All sequences were synthesized and subcloned by GENEWIZ.

Retroviral supernatants used for the transduction of murine T cells were prepared as previously described (16). In brief, retrovirus was produced by seeding 6-6.5 million HEK293T cells into 10 cm tissue culture dishes and transfecting the next day with 7.5 μg retroviral vector and 5 μg pCL-Eco plasmid that encodes gag/pol/env using polyethylenimine (Yeasen) at a 3:1 ratio. Supernatant containing the retrovirus was collected 48 h after transfection, and filtered with 0.45 μm filters. Retroviral supernatants were stored at 4°C or frozen at −80°C.

Spleen T cells were isolated from 6-8-week-old C57BL/6J mice and mechanically disrupted. T cells were isolated using MojoSort™ Mouse CD3 T Cell Isolation Kit (Biolegend). The T cells were then activated for 2 days using plates coated with anti-mouse CD3 (5 μg/ml, Biolegend, clone: 145-2C11), soluble anti-mouse CD28 (2 μg/ml, Biolegend, clone: 37.51) monoclonal antibodies and 50 U/ml murine IL-2(10 ng/ml, Peprotech) in complete medium (RPMI-1640, Gibco) supplemented with 10% FBS (Gibco), 2 mM GlutaMAX, 50 μM β-mercaptoethanol, 100 U/ml penicillin and 100 μg/ml streptomycin). T cells were transduced with retroviral supernatants using retronectin-coated plates (Takara). 12 to 24 h after virus transduction, T cells were replaced with a new complete medium with 50 U/ml murine IL-2. On day 5, transduction efficiency was evaluated by surface staining of an APC anti-DYKDDDDK tag antibody (Biolegend, Clone: L5). T cells were then harvested and used for further assays.

The cytotoxicity assay was performed using luciferase-expressing target cells. Murine T cells were co-cultured with firefly luciferase-tagged target cells EGFRvIII⁺MC38 at indicated ratios for 24 hours, in a total volume of 200 μl in a 96-well plate. Luciferase expression and activity were used as indicators of target cell viability. The luminescent signal was detected by the POLARstar Omega multifunctional microplate reader (BMG LABTECH) after removal of cell culture supernatant and addition of D-Luciferin potassium salt substrate (Beyotime).

Tumor lysates and supernatants from co-cultures of T cells and target cells were used for cytokine assays. Mouse IFN-γ and IL-2 ELISA kits (4A Biotech) were used to test the secretion of IFN-γ and IL-2 according to the manufacturer’s instructions. Duplicate wells were provided for each sample. The concentration of protein was calculated by measuring the optical density (OD) value of the absorbance at 450 nm.

The B16F10 and MC38 cells were seeded in 96-well plates and treated with varying concentrations of LTN for different time periods, and then incubated with CCK-8 reagent (NCMBio) for 4 hours. Absorbance was measured at a wavelength of 450nm using a microplate reader (Thermo).

Total RNA of cells or tissues was isolated using the EasyPure RNA Kit (TransGen). cDNA was then synthesized with Hifair AdvanceFast 1st Strand cDNA Synthesis Kit (Yeasen) according to the manufacturer’s instructions. qRT-PCR was performed with TB Green Fast qPCR Mix (Takara) on a Bio-Rad CFX96 system (Bio-Rad), with reactions run in triplicate. Relative mRNA levels were normalized to the endogenous control gene Gapdh and determined using the comparative threshold cycle (Ct) method. Primer sequences are listed in Supplementary Table 1 .

Female C57BL/6J mice (6–8 weeks old) were inoculated s.c. with 1×10⁶ EGFRvIII⁺MC38, EGFRvIII⁺B16F10, hCD19⁺MC38 or hCD19⁺B16F10 cells in 100 μl DPBS into the right flank. Mice were randomly assigned 7 days later, when tumors had grown to approximately 100 mm³. Then, the mice were then treated with 1×10⁷ T cells injected into the lateral tail vein.

In addition, mice in the CAR-T + LTN group were intraperitoneally injected with 1.25 mg/kg LTN (Solarbio) daily for 7 days. Tumor size and body weight were measured every two days. Mice were euthanized when the tumor volume reached 1500 mm³. On day 10 after infusion, some mice were euthanized and tumor tissues were collected and stored at -80°C for further analysis.

Tumors were collected and enzymatically digested by Collagenase I (Gibco) into single-cell suspensions. The single-cell suspensions were washed with PBS containing 1% FBS. Mouse TruStain FcX™ PLUS (BioLegend) was used for reducing non-specific immunofluorescent staining. For analysis of memory-type phenotypes and degranulation, the cells were stained with PE anti-mouse CD45 antibody (BioLegend, Clone:30-F11), FITC anti-mouse CD3 antibody (BioLegend, Clone:17A2), Brilliant Violet 510™ anti-mouse CD4 antibody (BioLegend, Clone: RM4-5), Alexa Fluor^® 700 anti-mouse CD8a antibody (BioLegend, Clone: 53-6.7), APC/Fire™ 750 anti-mouse CD62L antibody (BioLegend, Clone: MEL-14), APC anti-mouse CD44 antibody(BioLegend, Clone: IM7) and Brilliant Violet 421™ anti-mouse CD107a (LAMP-1) Antibody (BioLegend, Clone:1D4B). For analysis of the exhausted phenotype, cells were stained with PE anti-mouse CD45 antibody (BioLegend, Clone:30-F11), FITC anti-mouse CD3 antibody (BioLegend, Clone:17A2), Brilliant Violet 510™ anti-mouse CD4 antibody (BioLegend, Clone: RM4-5), Alexa Fluor^® 700 anti-mouse CD8a antibody (BioLegend, Clone: 53-6.7), APC/Fire™ 750 anti-mouse CD279 (PD-1) antibody (BioLegend, Clone: 29F.1A12), APC anti-mouse CD223 (LAG-3) antibody (BioLegend, Clone: C9B7W), Brilliant Violet 421™ anti-mouse CD366 (Tim-3) antibody (BioLegend, Clone: RMT3-23), FITC anti-mouse CD317 (BST2, PDCA-1) antibody (BioLegend, Clone: 927), APC/Fire™ 750 anti-human/mouse Granzyme B recombinant antibody (BioLegend, Clone: QA16A02), TCF1/TCF7 (C63D9) Rabbit mAb (Alexa Fluor^® 488 Conjugate, CST), Brilliant Violet 421™ anti-mouse CD127 (IL-7Rα) antibody (BioLegend, Clone: A7R34) and PE anti-mouse KLRG1 (MAFA) recombinant antibody (BioLegend, Clone: 2F1). After that, 7-AAD Viability Staining Solution (BioLegend) was added to stain the dead cells, and then the samples were detected and analyzed.

To investigate the effect of LTN on CAR-T cell function, we first constructed a CAR targeting the mutant EGFRvIII ( Figure 1A ). Flow cytometry confirmed efficient CAR surface expression on mouse T cells ( Figure 1B ). We then co-cultured these CAR-T cells with EGFRvIII-overexpressing MC38 colon cancer cells and concurrently treated them with varying LTN concentrations. While literature reports direct cytotoxic effects of LTN on tumor cells (17, 18), no obvious toxicity was observed at any dose tested here ( Supplementary Figure S1A ). However, LTN at a concentration of 10 ng/ml significantly improve the tumor-killing activity of CAR-T cells ( Figure 1C ). While the antitumor effect increased with higher LTN concentrations, the response was not strictly concentration-dependent ( Figure 1C ). Specially, at 10 ng/ml LTN, the lytic ability of CAR-T cells against tumor cells gradually increased with escalating effector-to-target (E/T) ratios ( Figure 1D ). Moreover, LTN treatment resulted in an increased percentage of CD107a⁺ CAR-T cells ( Figure 1E ), indicating enhanced degranulation and cytotoxic potential. LTN treatment also upregulated the secretion of key cytokines, including IL-2 and IFN-γ, further demonstrating its role in boosting CAR-T cell-mediated anti-tumor responses ( Figure 1F ). These findings indicate that LTN can significantly improve the antitumor activity of CAR-T cells at relatively low concentrations in vitro.

Central memory T cells are characterized by self-renewal capabilities and long-term survival, and their reactivation by tumor antigens leads to robust tumor cell killing (19). Given the importance of memory T cells for sustained anti-tumor immunity, we next assessed the effect of LTN on the differentiation of CAR-T cells into a central memory (Tcm, CD44⁺CD62L⁺) phenotype. After 48 hours of co-culture with EGFRvIII⁺MC38 target cells, we analyzed the phenotypes of CAR⁺CD8⁺T cells by flow cytometry. Our results showed that LTN treatment significantly increased the proportion of central memory CAR-T cells, with a marked effect observed at 10 ng/mL LTN ( Figure 2A ; Supplementary Figure S1B ). Notably, this increase was not concentration-dependent, as the proportion of Tcm cells plateaued within the concentration range tested. This suggests that LTN’s effect on CAR-T cell differentiation reaches an optimal threshold at a low dose. We further analyzed the impact of LTN on CAR-T cell subsets, including naïve T, effector memory T (Tem), effector T cells (Teff) and stem cell-like memory T cells (Tscm). LTN had no significant effect on the differentiation of naïve T, Tem and Teff cells ( Figures 2B, C ). Similarly, it did not alter CAR-T cell proliferation (Ki67⁺) or granzyme B expression levels ( Figures 2D, E ), potentially due to universally high baseline granzyme B expression in these CAR-T cells. However, LTN significantly promoted differentiation toward a Tscm phenotype ( Figure 2F ) and enhanced stemness of CAR-T cells ( Figure 2G ). To explore how LTN promotes memory differentiation in CAR-T cells, we assessed transcription of T cell memory-related genes. LTN increased mRNA levels of the memory reprograming regulators Tcf7 (encoding TCF1) and Foxo1, without altering effector-associated genes (Prdm1 and Tbx21) ( Figure 2H ). This selective upregulation implicates the TCF1/FOXO1 pathway in LTN-mediated induction of a memory phenotype.

CAR-T cell exhaustion is a major challenge, especially in the treatment of solid tumors, where CAR-T cells often exhibit a functional decline over time (20). To explore whether LTN could mitigate CAR-T cell exhaustion, we analyzed the expression of exhaustion markers such as CTLA-4, LAG-3, PD-1, and TIM-3. While LTN treatment did not change the expression of CTLA-4, LAG-3 and PD-1 ( Figures 3A-C ) nor the co-expression patterns of exhaustion markers (TIM3⁺LAG3⁺, TIM3⁺PD1⁺ and TIM3⁺CTLA4⁺) ( Supplementary Figure S2A ), it significantly reduced the expression of TIM-3 ( Figure 3D ) and, more notably, CD317 ( Figure 3E ). CD317 is a newly identified exhausted T cell marker in HCC and had been shown to exert immunosuppressive function in NK cells (21, 22). This reduction in TIM-3 and CD317 expression suggests that LTN may alleviate CAR-T cell exhaustion, further promoting their anti-tumor activity. Taken together, our findings provide compelling evidence that LTN significantly enhances the anti-tumor activity of CAR-T cells in vitro by promoting the differentiation of CAR-T cells into central memory phenotypes and by modulating key markers of exhaustion.

To evaluate the potential of LTN in boosting the anti-tumor activity of CAR-T cells in solid tumors, we used a syngeneic EGFRvIII⁺MC38 colon cancer model in immunocompetent mice. Following tumor cell implantation, engineered murine T cells were intravenously administered to the tumor-bearing mice ( Figure 4A ). Tumor volume and body weight were monitored over time. Our findings demonstrated that LTN could significantly improve the tumor-inhibitory effects of CAR-T cells and prolong the survival of treated mice ( Figures 4B, C ). We also investigated whether LTN alone could exhibit comparable anti-tumor effects to CAR-T cells. Although LTN alone could inhibit tumor growth, CAR-T cells combined with LTN produced the most potent tumor-suppressing effects ( Supplementary Figure S3A, B ). Furthermore, in a separate experiment using the EGFRvIII⁺B16F10 melanoma model ( Figure 4D ), LTN similarly enhanced the anti-tumor effects of CAR-T cells ( Figures 4E, F ). These results indicated that LTN could effectively enhance the anti-tumor activity of CAR-T cells in vivo across different solid tumor models.

Given the observed enhancement of CAR-T cell-mediated tumor suppression by LTN, we next explored whether LTN could affect the activation and phenotype of tumor-infiltrating T cells. Flow cytometric analysis showed that LTN, when combined with CAR-T cells, significantly increased the proportion of CD8⁺CD107a⁺T cells within tumors ( Figure 5A ), as well as the proportion of central memory cells ( Figure 5B ). Critically, while CTLA-4 expression did not significantly differ between groups ( Figure 5C ), CAR-T+LTN-treated mice exhibited markedly reduced expression of the inhibitory receptors PD-1, TIM-3, and LAG-3 in CAR-T cells ( Figures 5D–F ), along with significantly diminished co-expression frequencies of TIM-3⁺PD-1⁺, TIM-3⁺LAG-3⁺, and TIM-3⁺CTLA-4⁺ subsets ( Supplementary Figures S2B ). Similarly, increased levels of IL-2 and IFN-γ were observed in the tumors of CAR-T+LTN-treated mice ( Figure 5G ). We also assessed the durability of tumor-infiltrating CAR-T cells, finding a significant proportion of central memory-like T cells persisted within tumors at day 9 post-treatment ( Figure 5H ). These data indicate that LTN improves CAR-T cell anti-tumor efficacy by encouraging a favorable phenotype, durability, and activity. Given prior reports that LTN activates tumor-associated macrophages (TAMs) and induces inflammatory cytokine secretion in murine models (23), we also investigated its impact on TAM polarization. qPCR analysis revealed LTN increased transcription of M1-associated genes (Nos2, Cd86) and Il1b within tumors, while leaving M2-associated gene expression unaltered ( Figure 5I ). This shift toward a pro-inflammatory TAM phenotype suggests LTN also indirectly enhances CAR-T cell function by modulating the tumor microenvironment. Collectively, our findings demonstrate that LTN not only directly potentiates CAR-T cell anti-tumor activity but also reprograms TAMs toward an immunostimulatory state, thereby augmenting overall anti-tumor immunity.

To determine whether LTN could enhance CAR-T cell activity against other tumor targets, we generated human-derived CD19-expressing MC38 and B16F10 cells (hCD19⁺MC38 and hCD19⁺B16F10) as alternative tumor targets of CAR-T therapy. We then established a syngeneic hCD19⁺MC38 colon cancer model in immunocompetent mice ( Figure 6A ). As expected, LTN treatment significantly improved the tumor-suppressive activity of CAR-T cells and prolonged the survival of treated mice ( Figures 6B, C ). Similar results were observed in the hCD19⁺B16F10 melanoma model ( Figures 6D–F ). These results indicate that the enhancement of CAR-T cell function by LTN is not restricted to a certain tumor target or cancer type. Instead, LTN has broad applicability in potentiating the efficacy of CAR-T therapy.