The human embryonic kidney cell line 293F (RRID: CVCL_6642), along with three human mesothelin-expressing cell lines A431-H9 (a transfected A431 cell line stably expressing mesothelin), KLM-1 (RRID: CVCL_5146), and T3M4 (RRID: CVCL_4056) were utilized in this study. The mesothelin-negative A431 cell line (RRID: CVCL_0037) was also included. A431 and 293F were obtained from the China Center for Type Culture Collection, while KLM-1 and T3M4 were purchased from the American Type Culture Collection. All cells were cultured in DMEM medium (Invitrogen, Carlsbad, CA) supplemented with 10% FBS (HyClone, Logan, UT), 2mM L-glutamine and 1% penicillin-streptomycin (Invitrogen, Carlsbad, CA) and incubated in 5% CO2 with a balance of air at 37°C. All the cell lines were retrovirally transduced to express GFP/luciferase fusion protein.

Mesothelin-targeted monoclonal antibody R47, developed in our laboratory from previous work (32), was utilized to construct the bispecific antibody. The bispecific antibody (MSLN/CD3 bsAb) format was R47scfv-hFc-G4S-OKT3scFv (shortened as scFv-hFc-scFv). The MSLN/CD3 bispecific antibody was expressed in HEK-293F cells in a secreted form. The bispecific antibody plasmids carrying the fusion gene were introduced into 293F cells using polyethyleneimine for transient expression. Six days post-transfection, the supernatant from the transfected cells was collected. Protein purification was performed using a Protein A column, followed by buffer exchange with a Sephadex G-25 desalting column. The purity of the preparation was assessed by SDS-PAGE.

Human PBMCs were isolated from whole blood of healthy donors (Wuhan Blood Center) by Ficoll separation (Stem Cell Technologies, Vancouver, BC, Canada) according to the manufacturer’s instruction.

To establish NKT cell lines, PBMCs were cultured in RPMI 1640 containing 10% FBS the presence of α-GalCer (100 ng/ml) and recombinant interleukin-2 (rIL-2) (100 U/mL). α-GalCer is a potent ligand for NKT cells, and rIL-2 was used to promote the growth and activation of these cells. After 10–14 days, human NKT cells were restimulated by co-culture with irradiated PBMCs (typically irradiated at 25–30 Gy) that had been pulsed with α-GalCer (100 ng/mL) for 2–4 hours prior to the culture. This restimulation step was repeated every 2 weeks for a period of at least 1 month to maintain NKT cell expansion (33).

To establish γδT cell lines, PBMC were cultured in RPMI 1640 supplemented with 10% FBS (HyClone, Logan, UT), 1% L-glutamine, and 1% penicillin-streptomycin (Invitrogen, Carlsbad, CA). The cells were plated at a density of 1 × 10⁶ cells/mL in 6-well plates. γδT cells were activated by adding Zoledronate (5 μmol/L) (Adooq Bioscience), a bisphosphonate that induces the activation and expansion of γδT cells. Zoledronate was added to the culture medium at the beginning of the culture and refreshed twice a week. Every 2 days, the culture medium was supplemented with 100 U/mL rIL-2 support the proliferation of γδT cells. The culture period for the γδT cell lines lasted 14 to 21 days. After 14–21 days, γδT cells were harvested for further analysis or experimental procedures.

This research used a FACS Calibur (BD Biosciences, Franklin Lakes, NJ) to measure the fluorescence associated with live cells. All cells were harvested by detaching with trypsin-EDTA (ThermoFisher, Waltham, MA), washed by centrifugation, and resuspended in ice-cold PBS containing 5% BSA. The phenotype of NKT cells was assessed using Vα24-Jα18 mAb (6B11) followed by staining with a secondary goat anti-mouse-IgG-Cy5(Sangon biotch, China). The phenotype of γδT cells was assessed using TCR gamma/delta antibody (5A6.E9) followed by staining with a secondary goat anti-mouse-IgG-Cy5. Flow cytometry results were analyzed to determine the frequency of NKT and γδT cells in the sample. For the identification of CD3+ cells, PBMCs, NKT, and γδT cells were stained with the MSLN/CD3 bispecific antibody for 60 minutes at 4°C, followed by secondary staining with goat anti-mouse IgG-Cy5.

Human PBMCs, NKT, or γδT cells were co-cultured with KLM-1 cells in a 10:1 effector-to-target cell ratio in the presence of MSLN/CD3 bispecific antibody (bsAb) for 24 hours in flat-bottom 96-well plates (BD Biosciences). To assess T cell activation, surface expression of CD69 was analyzed by flow cytometry, with CD69 serving as a marker of early T cell activation. Following the incubation, cytokines in the culture supernatants were collected and analyzed using human IL-1β and IL-6 ELISA kits (Mlbio, China), according to the manufacturer’s instructions, to quantify the secretion of pro-inflammatory cytokines.

Cell growth and viability were assessed by firefly luciferases reporter assay. All the tested cancer cells were stably transfected to constitutively express a fire fly luciferase reporter gene (ffLuc2). One hundred microliters of stably transduced cells were seeded on a 96-well plate (5 × 10³ cells per well), followed by the addition of PBMC, NKT and γδT cells and bispecific antibodies at the indicated concentrations. After being incubated at 37°C for 72 h, cells were collected by spinning down, followed by lysis via two rounds of freeze-thaw. Released luciferase activity was measured to represent the cell viability.

All animal experiments were conducted in accordance with the regulations of the Laboratory Animal Care Committee of Huazhong Agricultural University (HZAUMO-2023-0304). All the operations and experimental procedures were complied with the national standard of Laboratory Animal-Guideline for ethical review of animal welfare (GB/T 35892–2018). Female NSG mice, were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd and maintained under specific pathogen-free conditions. For in vivo therapeutic experiments, 1-2 million A431(H9) cells or five million KLM-1 were subcutaneously injected into NSG mice. Seven days after tumor engraftment, the tumor formed and reached the size of 100-200 mm³, approximately 5 × 10⁶ PBMCs, NKT or γδT and different concentrations of MSLN/CD3 bsAb were injected intravenously into tumor-bearing mice. PBMCs, NKT or γδT was administered twice a week and MSLN/CD3 bsAb was administered every three days. Tumor dimensions were measured every two days with a caliper and calculated by the following formula: Tumor volume (mm³) = (length × width × width) × 0.5.

The sections obtained from organs (including brain, heart, liver, spleen, lung and kidney) from each group were paraffin-embedded and sliced into 4-µm per section. Then, the slides were baked at 65°C for 1 h, deparaffinized in xylene, rehydrated by graded ethanol, and stained with H&E successively.

All statistical analyses were conducted using GraphPad Prism5 (GraphPad Software, Inc., La Jolla, CA). Comparison of two groups was performed using unpaired Student’s t-test (two tailed). Comparisons among three or more groups were performed using one- way analysis of variance. Two-way ANOVA analysis of variance was used for tumor growth curve. P<0.05 was considered statistically significant.

NKT cells were selectively expanded and activated from whole PBMCs using α-GalCer. On day 0, PBMCs were activated with α-GalCer, resulting in NKT cell-specific proliferation. By day 14, the purity of NKT cells exceeded 90% ( Figure 1A ). The focus of this study is the circulating γδT cells expressing Vγ9/Vδ2 TCR heterodimers, as they exhibit potent anti-tumor functions (34). Vγ9/Vδ2 T cells were selectively expanded and activated from whole PBMCs using Zoledronate. On day 0, PBMCs were also activated with Zoledronate, inducing γδT cell-specific proliferation. By day 14, the purity of γδT cells exceeded 90% ( Figure 1B ). In addition, after being stimulated by MSLN-positive KLM-1 cells and MSLN/CD3 bsAb, NKT cells and γδT cells showed higher percentage of CD69 expression than PBMC cells ( Figure 1C ). And NKT cells had significantly higher expression levels of PD-1, TIM3 and LAG3 than γδT cells and PBMC cells. γδT cells had significantly lower expression levels of TIM3 and LAG3 than NKT cells and PBMC cells ( Figure 1D ).

The bispecific antibodies (MSLN/CD3 bsAb) were expressed in 293F cells and purified via protein A column ( Figure 2A ). The cell binding of the MSLN/CD3 bsAb was tested on MSLN-negative cell line A431, MSLN-overexpressing A431 cells (named H9), and MSLN-positive pancreatic cancer cell lines KLM-1 and T3M4, as well as different subsets of T cells (PBMC、NKT、γδT). As shown in Figure 2B , the MSLN/CD3 bsAb had specific binding to the MSLN-positive pancreatic cancer cell lines KLM-1, T3M4 and the MSLN-highly expressing H9 cells, but not to the MSLN-negative A431 cells ( Figure 2B ). The binding of MSLN/CD3 bsAb to T cells was confirmed on PBMC, NKT and γδT cells ( Figure 2C ). We evaluated the tumor killing and cytokine release mediated by the MSLN/CD3 bsAb in a co-culture assay using four MSLN+ tumor cell lines (KLM-1, T3M4 and H9) incubated with different subsets of T cells in the presence of the MSLN/CD3 bsAb. As shown in Figures 2D–G , MSLN/CD3 bsAb efficiently eradicated MSLN-positive KLM-1, T3M4, and H9 cells in a dose-dependent manner when co-cultured with PBMCs, NKT, and γδT cells in vitro, while exerting minimal impact on A431 cells. Additionally, both NKT and γδT cells demonstrated comparable anti-tumor activity, significantly superior to that of PBMCs ( Figures 2D–G ). PBMC in the presence of the MSLN/CD3 bsAb secreted high levels of IL-1β and IL-6, while NKT and γδT cells secreted low levels of IL-1β and IL-6 ( Figures 2H, I ). This also indicates that NKT and γδT cells have potential safety advantages in avoiding cytokine storm syndrome in cancer patients.

To evaluate the in vivo tumor suppression activity of the MSLN/CD3 bsAb in the presence of PBMC, NKT and γδT, a xenograft mouse model was established by subcutaneous transplantation of H9 cells in NSG mice. When the tumor reached the size of 100-200 mm³, treatment was initiated by tail vein injection of 5 million PBMC, NKT or γδT once a week and tail vein injection of 0.5, 1 or 2 mg/kg of the MSLN/CD3 bsAb every two days. As shown in Figures 3A, B , the MSLN/CD3 bsAb significantly inhibited the tumor growth of H9 cells in the presence of PBMC, and this effect was dose-dependent ( Figure 3A ). The body weight of the treated mice also significantly decreased, and this was correlated with the dose of the bispecific antibody ( Figure 3B ). As shown in Figures 3C, D , the MSLN/CD3 bsAb significantly inhibited the tumor growth of H9 cells in the presence of NKT, and this effect was independent of the dose ( Figure 3C ). The body weight of the treated mice in this study also significantly decreased, independent of the dose of the bispecific antibody ( Figure 3D ). As shown in Figures 3E, F , the MSLN/CD3 bsAb significantly inhibited the tumor growth of H9 cells in the presence of γδT, and this effect was dose-dependent ( Figure 3E ). The body weight of the treated mice did not change ( Figure 3F ).

To evaluate the in vivo tumor suppression activity of PBMC, NKT and γδT, a xenograft mouse model was established by subcutaneous transplantation of H9 or KLM-1 cells in NSG mice ( Figure 4A ). In the H9 tumor model, PBMC, NKT and γδT significantly inhibited the tumor growth in the presence of the MSLN/CD3 bsAb ( Figure 4B ), However, the body weight of the NKT-treated mice decreased the most, followed by the PBMC-treated mice, while the γδT-treated mice maintained stable body weight ( Figure 4C ). In the KLM-1 tumor model, NKT and γδT significantly inhibited the tumor growth in the presence of the MSLN/CD3 bsAb better than PBMC, with γδT showing the best anti-tumor effect ( Figure 4D ). The body weight of the treated mice was similar to that in the H9 tumor model ( Figure 4E ). The disparity in activity observed between the H9 and KLM-1 models could be attributed to the origin of the cells; H9 cells are derived from A431 cells overexpressing MSLN and thus exhibit higher resistance, whereas KLM-1 cells are more susceptible to immune responses. Furthermore, the findings also suggest that γδT cells possess stronger and safer anti-tumor activity.

To better describe the control of different subsets of T cells on the pancreatic cancer model, we performed a more in-depth analysis of the tumors during the treatment. We collected the early and regressed specimens of KLM-1 pancreatic tumors treated with different subsets of T cells at 7 and 21 days after treatment, and measured the infiltration of CD3+ T cells by immunohistochemistry. At 7 days after treatment, T cells were enriched in the tumors treated with different subsets of T cells. At 21 days after treatment, the tumors treated with PBMC cells showed a typical immune rejection phenotype, with very limited T cell infiltration in the tumor core and most T cells accumulated at the tumor periphery. In contrast, the tumors treated with NKT and γδT cells had much more T cell infiltration throughout the tumor core ( Figure 5A ). To investigate the homing ability of PBMC, NKT and γδT cells to the tumor under the effect of MSLN/CD3 bsAb, PBMC, NKT and γδT cells labeled with Dil dye were transferred into mice bearing established KLM-1 tumors (about 1000 mm³), and 2 mg/kg MSLN/CD3 bsAb was injected at the same time. After 24 hours, the harvested organs confirmed that PBMC, NKT, and γδT cells mainly homed to the tumor, although a small fraction of these cells was also found in the liver of some mice ( Figure 5B ). Additionally, this study investigated whether PBMCs, NKT cells, and γδT cells would cause toxicity to healthy tissues in mice. Various organs such as the brain, heart, liver, spleen, lungs, and kidneys were extracted from treated mice and subjected to HE staining. Histological analysis revealed that mice treated with PBMCs and NKT cells exhibited lung injury, as evidenced by H&E staining showing severe damage to alveolar capillary structures and inflammatory cell infiltration. Furthermore, mice treated with NKT cells showed liver and kidney damage, while mice treated with PBMCs exhibited kidney damage. However, mice treated with γδT cells did not exhibit any significant damage ( Figure 5C ). These results indicate that γδT cells have stronger anti-tumor effects than PBMCs, and can effectively accumulate and survive in the tumor without causing harm to vital organs.

This study used intraperitoneal injection of α-Galcer or Zoledronate to activate or expand NKT or γδT cells in vivo, to further verify the anti-tumor effects of NKT and γδT cells. In brief, KLM-1 or T3M4 cells were subcutaneously injected into NSG mice to form palpable solid tumors. Treatment was initiated by intravenous injection of human PBMCs, followed by intravenous injection of α-Galcer or Zoledronate every other day. MSLN/CD3 bsAb was intravenously injected every 3 days ( Figure 6A ). As shown in Figures 6B, C , compared with the control and PBMC alone, tumor regression was observed in mice after injection of α-Galcer or Zoledronate. In contrast to PBMC+ MSLN/CD3 bsAb, injection of α-Galcer or Zoledronate combined with MSLN/CD3 bsAb treatment showed significant and lasting tumor growth inhibition, and the anti-tumor effects were similar. In addition, this study also measured the body weight of the mice during all treatments. Except for the PBMC+MSLN/CD3 bsAb treatment group, which showed obvious fluctuations in body weight during treatment, there was no significant difference in body weight between the other groups ( Figures 6D, E ), indicating that α-Galcer or Zoledronate induced NKT or γδT cells did not cause systemic toxicity to the mice.