Tumor vaccine based on extracellular vesicles derived from γδ-T cells exerts dual antitumor activities

Abstract

γδ-T cells are innate-like T cells with dual antitumor activities. They can directly eradicate tumor cells and function as immunostimulatory cells to promote antitumor immunity. Previous studies have demonstrated that small extracellular vesicles (EVs) derived from γδ-T cells (γδ-T-EVs) inherited the dual antitumor activities from their parental cells. However, it remains unknown whether γδ-T-EVs can be designed as tumors vaccine to improve therapeutic efficacy. Here, we found that γδ-T-EVs had immune adjuvant effects on antigen-presenting cells, as revealed by enhanced expression of antigen-presenting and co-stimulatory molecules, secretion of pro-inflammatory cytokines and antigen-presenting ability of DCs after γδ-T-EVs treatment. The γδ-T-EVs-based vaccine was designed by loading tumor-associated antigens (TAAs) into γδ-T-EVs. Compared with γδ-T-EVs, the γδ-T-EVs-based vaccine effectively promoted more tumor-specific T-cell responses. In addition, the vaccine regimen preserved direct antitumor effects and induced tumor cell apoptosis. Interestingly, the allogeneic γδ-T-EVs-based vaccine showed comparable preventive and therapeutic antitumor effects to their autologous counterparts, indicating a better way of centralization and standardization in clinical practice. Furthermore, the allogeneic γδ-T-EVs-based vaccine displayed advantages over the DC-EVs-based vaccine through their dual antitumor activities. This study provides a proof-of-concept for using the allogeneic γδ-T-EVs-based vaccine in cancer control.

Keywords: extracellular vesicle; immunotherapy; tumor; vaccine; γδ-T cells.

FIGURE 1

γδ‐T‐EVs‐based vaccine efficiently induces tumor‐specific T‐cell responses. (a) Size distributions; (b) Expression of EVs markers CD63 and CD81, Gray histograms represent isotype control; (c) Expression of Alix, EBNA1, and LMP2a proteins in γδ‐T‐EVs and γδ‐T‐EVs (TAAs). (d–g) Equivalent amounts of PBS, TAAs, γδ‐T‐EVs, γδ‐T‐EVs + free TAAs, or γδ‐T‐EVs (TAAs) were used to treat huPBMCs. After 7 days, the cells were restimulated with peptide pools and subjected to detecting intracellular IFN‐γ in T cells. (D,E) Induction of EBNA1‐specific CD4 and CD8 T cells, (F,G) induction of LMP2a‐specific CD4 and CD8 T cells by γδ‐T‐EVs (TAAs) and corresponding controls. Representative data were shown as mean ± SEM from three independent experiments. *p < 0.05, ** p < 0.01, ***p < 0.001. EVs: γδ‐T‐EVs; HIV: huPBMCs pretreated with γδ‐T‐EVs (TAAs) were restimulated with HIV p17 peptide pool.

FIGURE 2

γδ‐T‐EVs‐based vaccine displays immune adjuvant effects on antigen‐presenting cells. (a) Completed huPBMCs or HLA‐DR⁺ cell‐depleted huPBMCs were treated with γδ‐T‐EVs (TAAs) for seven days, then LMP2a‐specific T cells were detected. (b) Purified CD3 T cells were treated with γδ‐T‐EVs or γδ‐T‐EVs (TAAs) in the absence or presence of iDCs. TAAs treatment was also used as a control (Ctr). Seven days later, LMP2a‐specific T cells were detected. (c) iDCs were treated with γδ‐T‐EVs or γδ‐T‐EVs (TAAs) for 48 h. TAAs was used as control (Ctr). The APC functional makers (MHC‐II, CD83, CD86, and CD40) were detected by flow cytometry. (d‐e) Secretion of TNF‐α and IL‐6 from iDCs after 48 h treatment with γδ‐T‐EVs or γδ‐T‐EVs (TAAs). (f) iDCs were treated with γδ‐T‐EVs, or γδ‐T‐EVs (TAAs) for 48 h, then cocultured with CFSE‐stained allogeneic CD3 T cells. Seven days later, proliferating T cells were determined by CFSE dilution using flow cytometry. (g) Secretion of IFN‐γ in the coculture system of pretreated iDCs with allogeneic CD3 T cells. (h) Expression of IFN‐γ on γδ‐T‐EVs or γδ‐T‐EVs (TAAs). (I‐J) iDCs were treated with γδ‐T‐EVs, or γδ‐T‐EVs (TAAs), in the presence of neutralizing anti‐IFN‐γ antibody or isotype control. 48 h later, the pretreated iDCs were cocultured with CFSE‐stained allogeneic CD3 T cells. After seven days, proliferating T cells were determined. Representative data were shown as mean ± SEM from three independent experiments. MFI: median fluorescence intensity. *p < 0.05, **p < 0.01, ***p < 0.001.

FIGURE 3

γδ‐T‐EVs‐based vaccine preserves direct antitumor activities both in vitro and in vivo. (a) Apoptosis of EBV‐LCL and (b) SNU‐719 cells were determined after cultured with γδ‐T‐EVs or γδ‐T‐EVs (TAAs) for 18–24 h. An equivalent amount of PBS or TAAs was used as a control. (c) EBV‐LCL cells were injected s.c. in Rag2^−/−γc^−/− mice. After 14 days, mice with subcutaneous tumors were treated with γδ‐T‐EVs, or γδ‐T‐EVs (TAAs) at indicated (n = 10). TAAs was used as control (Ctr). The tumor volume (d) and mice survival (e) were determined at the indicated time. (f) Histochemical analysis of human Ki‐67 in tumor tissues at the endpoints, scale bar = 25 μm. (g) Mice with subcutaneous SNU‐719 tumors were treated with γδ‐T‐EVs, or γδ‐T‐EVs (TAAs) as indicated (n = 12). TAAs was used as control (Ctr). The tumor volume (h) and mice survival (i) were determined at the indicated time. (j) Histochemical analysis of human Ki‐67 in tumor tissues at the endpoints, scale bar = 25 μm. For the bar graphs, representative data were shown as mean ± SEM from three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001. ns: not significant.

FIGURE 4

γδ‐T‐EVs‐based vaccine inhibits tumor progression in humanized mice. (a) Tumor models were established by injection s.c. of autologous EBV‐LCL cells in humanizedmice. Then, γδ‐T‐EVs or γδ‐T‐EVs (TAAs) were injected into humanized mice i.p. as indicated (n = 12). TAAs was used as control (Ctr). The tumor volume (b) and mice survival (c) were determined at the indicated time. (d) Representative histochemical analysis of human Ki‐67 in tumor tissues at the endpoints, scale bar = 25 μm. (e‐f) TAAs, γδ‐T‐EVs or γδ‐T‐EVs (TAAs) were injected into EBV‐induced B cell lymphoma‐bearing humanized mice biweekly for two doses (n = 4). Seven days post the booster injection, the percentage of LMP2a‐ and EBNA1‐specific T cells in peripheral blood nucleated cells were detected. (g) Death of EBV‐LCL cells after 6 h cocultured with autologous T cells induced by γδ‐T‐EVs or γδ‐T‐EVs (TAAs). EBV‐LCL cells alone were used as control (Ctr). (h) Death of EBV‐LCL cells with preloading of EBV peptides after 6 h cocultured with autologous T cells induced by γδ‐T‐EVs (TAAs). EBV‐LCL cells preloaded with HIV p17 peptides were used as control (Ctr). For the bar graphs, representative data were shown as mean ± SEM from two to three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001. ns: not significant.

FIGURE 5

Allogeneic γδ‐T‐EVs‐based vaccine promotes antigen‐presenting cells to induce tumor‐specific T‐cell responses. Expression of MHC‐II (a) and CD83 (b) on iDCs after treated with autologous, allogeneic γδ‐T‐EVs or γδ‐T‐EVs (TAAs) for 48 h. TAAs was used as control (Ctr). (c,d) Secretion of IL‐6 and TNF‐α from iDCs after treatment. (e) iDCs were pretreated with autologous, allogeneic γδ‐T‐EVs or γδ‐T‐EVs (TAAs) for 48 h, then cocultured with CFSE‐stained allogeneic CD3 T cells. Seven days later, proliferating T cells were determined by CFSE dilution using flow cytometry. (F,G) Autologous, allogeneic γδ‐T‐EVs or γδ‐T‐EVs (TAAs) were used to treat huPBMCs. After 7 days, the cells were restimulated with EBNA1 (f) or LMP2a (g) peptide pools and subjected to detecting intracellular IFN‐γ in T cells. (h) Completed huPBMCs or HLA‐DR⁺ cell‐depleted huPBMCs were treated with the autologous or allogeneic γδ‐T‐EVs (TAAs) for seven days, then the LMP2a‐specific T cells were detected. For the bar graphs, representative data were shown as mean ± SEM from three independent experiments. MFI: median fluorescence intensity. *p < 0.05, **p < 0.01, ***p < 0.001. ns: not significant.

FIGURE 6

Allogeneic γδ‐T‐EVs‐based vaccine inhibits tumor progression in humanized mice. (a) Tumor models were established by injection s.c. of autologous EBV‐LCL in humanizedmice. Autologous or allogeneic γδ‐T‐EVs (TAAs) were injected into humanized mice i.p. at the indicated time. PBS, TAAs, or γδ‐T‐EVs was used as control (n = 12 or 13). Tumor volume (b) and mice survival (c) was calculated as indicated. (d) Histochemical analysis of human Ki‐67 in tumor tissues at the endpoints, scale bar = 25 μm. (e) Tumor models were established by injection of autologous EBV‐LCL in humanized mice reconstituted with complete huPBMCs or CD3 T cell‐depleted huPBMCs. TAAs or allogeneic γδ‐T‐EVs (TAAs) were injected intraperitoneally into humanized mice at the indicated time (n = 11). The tumor volume (f) and mice survival (g) were calculated as indicated. (h) Histochemical analysis of human Ki‐67 in tumor tissues at the endpoints, scale bar = 25 μm. *p < 0.05, **p < 0.01, ***p < 0.001. ns: not significant; Auto‐: autologous; Allo‐: allogeneic.

FIGURE 7

Allogeneic γδ‐T‐EVs‐based vaccine prevents tumor development in humanized mice. (a) Humanized mice were primed and boosted using the autologous or allogeneic γδ‐T‐EVs‐based vaccine as indicated, then EGFP‐expressing EBV‐LCL cells were subcutaneously injected 1 week later (n = 13). An equivalent amount of PBS, TAAs, or γδ‐T‐EVs was used as a control. (b) Tumor incidence, (c) whole‐body fluorescence images of mice 42 days after tumor cell inoculation, (d) tumor volume, and (e) mice survival were determined as indicated. ***p < 0.001. ns: not significant; Auto‐: autologous; Allo‐: allogeneic.

FIGURE 8

Allogeneic γδ‐T‐EVs‐based vaccine has advantages over DC‐EVs‐based vaccine by displaying dual antitumor activities. (a‐b) Secretion of TNF‐α and IL‐6 from iDCs after 48 h treatment with allogeneic γδ‐T‐EVs, DC‐EVs, γδ‐T‐EVs (TAAs), or DC‐EVs (TAAs). TAAs was used as control (Ctr). (c) iDCs were treated with allogeneic γδ‐T‐EVs, DC‐EVs, γδ‐T‐EVs (TAAs), or DC‐EVs (TAAs) for 48 h, and TAAs were used as control (Ctr). Then, the pretreated‐iDCs were cocultured with CFSE‐stained allogeneic CD3 T cells. Seven days later, proliferating CD3 T cells were determined by CFSE dilution using flow cytometry. (d) Secretion of IFN‐γ in the coculture system of pretreated iDCs with allogeneic CD3 T cells. (e) Allogeneic γδ‐T‐EVs‐ or DC‐EVs‐based vaccines were used to treat huPBMCs. After seven days, the cells were restimulated with LMP2a peptide pools and subjected to detecting intracellular IFN‐γ in T cells. (f) Apoptosis of EBV‐LCL cells after being cultured with allogeneic γδ‐T‐EVs‐ or DC‐EVs‐based vaccine for 18–24 h. (g) Tumor models were established by injection s.c. of EBV‐LCL cells in humanizedmice. Then, allogeneic γδ‐T‐EVs‐based or DC‐EVs‐based vaccines were injected into humanized mice i.p. at the indicated time (n = 11 or 12). TAAs was used as control (Ctr). The tumor volume (h) and mice survival (i) were calculated as indicated. (j) Histochemical analysis of human Ki‐67 in tumor tissues at the endpoints, scale bar = 25 μm. For the bar graphs, representative data were shown as mean ± SEM from three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001. ns: not significant.