Enhancing the anti-leukemia immunity of acute lymphocytic leukemia-derived exosome-based vaccine by downregulation of PD-L1 expression

Abstract

Cell-released nanovesicles can induce anti-leukemia immunity. Leukemia cell-derived exosomes (LEXs) are promising anti-tumor vaccine components for cancer immunotherapy. Nonetheless, LEX-based vaccines show modest potency in vivo, likely due to the presence of immunosuppressive PD-L1 proteins in the exosomes. We hypothesized that targeting exosomal PD-L1 could optimize LEX-based vaccines. To test this hypothesis, we compared the capacity of exosomes derived from PD-L1-silenced acute lymphocytic leukemia-derived leukemia cells (LEX_PD-L1si) and non-modified exosomes to induce anti-leukemia immunity. Lentivirus-mediated PD-L1 shRNA was used to downregulate PD-L1 expression in parental leukemia cells and LEXs. LEX_PD-L1si were characterized by electron microscopy, Western blotting, nanoparticle tracking analysis and flow cytometry, and their anti-leukemia immune effects were tested on immune cells and in animal models. In the present study, lentivirus-mediated PD-L1 shRNA successfully downregulated PD-L1 expression in parental leukemia cells and in LEXs. LEX_PD-L1si induced better DC maturation and subsequently enhanced T cell activation, as compared with non-modified LEXs. Consistently, immunization with LEX_PD-L1si induced greater T cell proliferation and Th1 cytokine release. LEX_PD-L1si was a more potent inducer of antigen-specific cytotoxic lymphocyte (CTL) response. Finally, we vaccinated DBA/2 mice with exosome formulations to test their ability to induce both protective and therapeutic anti-tumor CTL responses in vivo. Vaccination with LEX_PD-L1si strongly inhibited tumor growth and prolonged survival of immunized mice. Downregulation of exosomal PD-L1 expression in LEXs effectively induces more potent anti-leukemia immunity. Therefore, our strategy for optimizing LEX-based vaccine has a potential application in leukemia immunotherapy.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00262-021-03138-5.

Introduction

Acute lymphoblastic leukemia (ALL) is a hematologic malignancy which historically had an inferior prognosis with cure rate less than 40% [1, 2]. Despite recent advances in conventional chemotherapy, numerous patients still experience severe toxicity, including infections, gastrointestinal reactions and cerebral hemorrhage [3–5]. Moreover, up to half of the patients may be insensitive to or unfit for chemotherapy [6]. Thus, the 5-year overall survival rate of patients with acute lymphoblastic leukemia (ALL) is approximately 40% [7, 8]. Elderly patients with acute leukemia generally have worsened prognosis and show a 5-year overall survival rate of less than 20% [9]. These findings highlight the need for alternative treatment strategies in patients ineligible for intensive chemotherapy and those with recurrent or refractory diseases [10].

Recent studies examining the interactions between acute lymphocytic leukemia(ALL) cells and the immune system have yielded immunotherapeutic approaches that can be used to improve prognosis and survival in patients with ALL [11]. A well-known immunotherapeutic approach that is entering clinical application is CD19-chimeric antigen receptor (CAR) T cell therapy, targeted to kill CD19-positive lymphoblastic leukemia cells via gene modification. Although CD19-CAR T cell therapy is promising, it still shows off-target effects and has a high proportion of side effects. Besides, many responses are not durable. Relapses occur in approximately 50% ALL patients undergoing CAR-T treatment, including CD19-negative relapses. [12, 13]. Therefore, using CAR-T for the treatment of patients with ALL does not meet expectation [13]. Another immunotherapy strategy in ALL, antibody–drug conjugate targeting CD22, improves the complete remission rate and MRD negativity rate of patients with R/R ALL to a certain extent. However, it provides limited survival benefits and is associated with high incidence of hepatotoxicity [14]. Alternatively, the direct and sustained activation of tumor-specific T cells in vivo via local inoculation of anti-leukemia vaccines carrying leukemia-associated antigens shows a superior therapeutic spectrum and safety profile [15].

Exosomes, which are bioactive vesicles released by eukaryotic cells and have a diameter of 30–130 nm, carry various information components derived from their parental cells [16]. Tumor cell-derived exosomes carry tumor-associated antigens, which can be used in novel tumor vaccines to stimulate the priming of T cells in immunotherapy [17]. Our previous studies have shown that similar to other tumor cells, ALL cells can release significant quantities of exosomes that harbor native tumor-associated antigens derived from their parental cells. These findings indicate that leukemia cell-derived exosomes (LEX) can be utilized as an anti-leukemia vaccine for targeted elimination of leukemia cells [18–21]. Additionally, LEX-based vaccines are relatively more stable than cell-based vaccines and can cross the blood–brain barrier non-invasively [22]. These characteristics of LEX vaccines are useful in clinical applications. Conversely, unmodified tumor exosomes (TEX) show poor immunogenicity and can promote immune tolerance, thereby substantially compromising their therapeutic performance [23]. However, the immunogenicity of LEX-based leukemia vaccines requires optimization. The unsatisfactory immune response induced by LEX can be attributed to two factors. First, deficiencies in immune-stimulating factors, such as adhesion and co-stimulatory molecules, compromise the immunogenicity of exosome-derived vaccines [18]. Second, high levels of immunosuppressive factors in TEXs mediate immune tolerance [24, 25]. Among these immunosuppressive factors, programmed cell death protein–ligand 1 (PD-L1), one of immune checkpoint molecules that interacts with programmed cell death protein-1 (PD-1), is expressed on the surface of the tumor cells and on TEXs [24, 26]. Exosomal PD-L1, which is resistant to anti-PD-L1 therapy, can transmit immunosuppressive signals to T cells and anti-apoptotic signals to tumor cells, inducing local and systemic immunosuppression and promotion of tumor growth [24] Our previous study showed that PD-L1 is highly enriched in LEX from ALL cells, thus creating a barrier for therapeutic vaccination against ALL.

To overcome these limitations and improve the immunogenicity of ALL cell-derived exosomes-based vaccines, we downregulated exosomal expression of PD-L1. For this, we used a lentiviral vector containing PD-L1 small hairpin RNA (shRNA) to silence PD-L1 expression in L1210 leukemia cells. Then, we isolated and analyzed exosomes derived from these genetically engineered parental cells. Finally, we investigated the anti-leukemia efficacy of exosomes derived from PD-L1 silenced leukemia cells (LEX_PD-L1si).

Materials and Methods

Reagents

RPMI-1640 medium, fetal bovine serum (FBS) and serum-free medium AIM-V were purchased from Invitrogen (Shanghai, China). Recombinant mouse IL-2 protein was purchased from Abcam (Shanghai, China). Recombinant mouse granulocyte–macrophage colony-stimulating factor (rmGM-CSF), recombinant human interleukin (rhIL)-4 and lipopolysaccharide (LPS) were purchased from PeproTech (Shanghai, China). Rabbit anti-mouse shock protein 70(catalog numbers: 4872 T)was obtained from Abcam(Shanghai, China). TSG101(catalog numbers: ab125011) and CD9(catalog numbers: ab223052) antibodies were obtained from Cell Signaling Technology(Shanghai, China). Rabbit anti-mouse antibodies PD-L1 and calnexin were purchased from Abcam (Shanghai, China). PE-labeled anti-MHC Ia/Ib, PE-cyanine7 conjugated anti-CD80 and APC-labeled anti-CD86 were purchased from eBioscience (Shanghai, China). EasySep™ Mouse CD4⁺ and CD8⁺ T cell isolation kits were purchased from StemCell Technologies (Vancouver, Canada). Aldehyde/sulfate latex beads were purchased from Invitrogen (Shanghai, China). Enzyme-linked Immunosorbent Assay (ELISA) kits for quantitative detection of mouse IL-12p70(catalog numbers BMS6004), TNF-α(catalog numbers BMS607-3), IFN-γ(catalog numbers: 88–7314) and IL-2 (catalog numbers: BMS601) were purchased from Invitrogen( Shanghai, China). Anti-CD8 monoclonal antibody clone 2.43 (catalog numbers: BE0061), anti-CD4 monoclonal antibody clone GK1.5 (catalog numbers: BE0003-1) and a rat isotype control IgG clone LFT-2(catalog numbers: BE0090) were purchased from BioXcell( Shanghai, China).

2Cell lines and animals

The murine acute lymphocytic leukemia lines, L1210 and p388, were purchased from the Shanghai Institute for Biological Science (Shanghai, China) and cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS). DBA/2 female mice (6–8 weeks old) were purchased from the Shanghai Laboratory Animal Center and were housed in a specific pathogen-free, regularly controlled animal house at 18–22℃ in a 12-h light/dark cycle and fed standard chow and water ad libitum. All procedures involving animals were approved by the Ethics Committee of Xinhua Hospital Affiliated with the Shanghai Jiaotong University School of Medicine.

Generation of bone marrow-derived DCs

DBA/2 mice were killed, and dendritic cells (DCs) were generated from bone marrow-derived precursors as previously described and cultured in RPMI-1640 medium containing 10% fetal bovine serum (FBS), 100 u/mL penicillin, 100 mg/mL streptomycin, 10 ng/mL rmG-MSF and 10 ng/mL rmIL-4. Following cell culture for 6 days, DCs were collected and incubated with PBS, LEX, LEX_GFP, LEX_PD-L1si or LPS in complete medium containing 10% exosome-free FBS for 24 h. Then DCs and the supernatants were collected and stored for flow cytometry and ELISA analysis.

Mixed lymphocyte reaction (MLR)

DCs from DBA/2 mice were stimulated with 10 µg/ml of exosomes, 100 µg/ml PBS or 1 μg/ml LPS for 48 h, respectively. Then, the DCs were irradiated with 4000 rad and used as stimulators. Allogeneic T cells, which were isolated with the EasySepTM T cell separation kit, were used as responders (0.1 × 106cells/well). DC/T cell co-culture was maintained for 72 h at 37 °C in RPMI 1640 complete medium. Following that, [3H] thymidine (0.5 μCi/well) was added per well. After a further 24 h of co-culture, T cell proliferation was determined via [3H] thymidine incorporation.

Lentivirus vector construction and cell infection

Three pairs of self-complementary oligonucleotides carrying shRNA sequences targeting mouse PD-L1 were synthesized at the Shanghai Hanbio Co., Ltd (Shanghai, China). A scrambled shRNA sequence was used as negative control. Oligonucleotides encoding PD-L1 shRNAs and scrambled shRNA were introduced into lentiviral frame plasmids, pHBLV-U6-Scramble-Zsgreen (Shanghai HanbioCo., Ltd, Shanghai, China). The recombinant plasmid DNAs were then transfected into Escherichia coli for construction of the recombinant plasmid. After confirming successful ligation, 293 T cells were co-transfected with the recombinant lentiviral vector (10 μg), pSPAX2 vector (10 μg) and pMD2G vector (10 μg) to pack the vector. The harvested lentiviruses were titrated as previously described [27]. The targeted cell line was transduced using previously described protocols [19]. To evaluate the efficiency of interference, RT-PCR and Western blotting were used to detect the mRNA and protein expression levels of PD-L1 in the targeted cell line. Of the three lentiviral vectors containing PD-L1 shRNA, the vector containing PD-L1 shRNA3 showed the highest interference efficiency for PD-L1 mRNA and protein expression and was, therefore, selected for use in further procedures. The shRNA3 sequence targeting mouse PD-L1 was as follows: Top strand: CCGGGAAGCAAAGTGATACACATCTCAAGAGAATGTGTATCACTTTGCTTCTTTTTTTG; Bottom strand: AATTCAAAAAAGAAGCAAAGTGATACACATTCTCTTGAGATGTGTATCACTTTGCTTC. The scramble shRNA sequence used as negative control was as follows: Top strand: GATCCCCTTCTCCGAACGTGTCACGTTTCAAGAGAACGTGACACGTTCGGAGAATTTTTGGAAA-3; Bottom strand: AGCTTTTCCAAAAATTCTCCGAACGTGTCACGTTCTCTTGAAACGT GACACGTTCGGAGAAGGG.

Exosome isolation

L1210 cells were pre-cultured overnight in AIM-V medium without FBS to avoid contamination with serum exosomes. Cell-culture supernatant was collected and used for exosome extraction. Leukemia cell-derived exosomes were accumulated using differential centrifugation [28]. The harvested exosome pellets were washed twice with glacial phosphate-buffered saline (PBS) and re-concentrated using ultra-centrifugation (at 100,000 g for 1 h). Exosomes derived from non-modified L1210 cells were designated as LEX. Exosomes derived from L1210 cells transfected with the lentiviral vector containing the scrambled shRNA sequence and PD-L1 shRNA were designated as LEX_GFP and LEX_PD-L1si, respectively. Exosome morphology and typical exosomal proteins (HSP70, TSG101 and CD9)were identified using transmission electron microscopy(Philips CM12) and Western blotting analysis as described previously [18, 19].

Nanoparticle tracking analysis (NTA)

NTA which was applied to analyze the size distribution and average size of exosomes was performed using NanoSight LM10 (NanoSight Ltd., Amesbury, UK) [29]. Briefly, the exosomes were diluted in PBS to the concentration into the recommended range for NTA. Each exosome sample was recorded five times for 30 s. Captured nanoparticles were analyzed using NTA software (Version 3.2 Build 16).

Flow cytometry

To quantify the expression levels of PD-L1 on the exosomal surface, 30 μg of exosomes was first incubated with aldehyde/sulfate latex beads at 4 °C overnight; the reaction was then blocked by the addition of 100 mmol/L glycine. As described in our previous study [18, 19], exosome-loaded latex beads were washed twice in PBS containing 1% fetal taurine and were then stained with either a specific antibody against mouse PD-L1 or an isotype control. Fluorescence intensity of exosome-loaded latex beads was analyzed using a BD FACScan TM flow cytometry.

To analyze the effects of the exosomes on the phenotype of bone-marrow-derived DCs, DCs were incubated with PBS, LEX, LEX_GFP, LEX_PD-L1si or LPS for 24 h. Afterward, DCs were collected and stained with PE-cyanine7-labeled anti-CD80, PE-labeled anti-MHC Ia/Ib and APC-labeled anti-CD86 antibodies and then analyzed using FACScan as described previously [18, 19].

T cell proliferation assay

Splenic T cells were collected from 6 to 8-week-old female DBA/2 mice immunized seven days earlier. The isolated splenic T cells (1 × 10⁵ cells/well) were then co-cultured for 72 h with irradiated L1210 (1 × 10⁴ cells/well) or p388 cells (1 × 10⁴ cells/well; used as controls) in the presence of PHA (20 μg/ml) at 37˚C and 5% CO₂. Then, [³H] thymidine (0.5 μCi per well) was added to the cultures and allowed to incubate for an additional 16 h. Subsequently, the cells were harvested, and [³H]-thymidine uptake was detected by MicroBeta counter (Beckman Coulter, Krefeld, Germany).

Cytotoxicity assay

For the CTL assay, single-cell suspensions of splenocytes were collected from 6–8-week-old female DBA/2 mice immunized seven days earlier. The CD8⁺ T cells were then isolated from splenocytes using an EasySep™ mouseCD8⁺T cell Isolation Kit and re-stimulated with irradiated L1210 cells and mouse IL-2 ((100 μg/mL) for 7 days; the resulting effector cells were then harvested. L1210 cells (used as specific target cells) or p388 cells (serving as controls) were seeded in a 96-well plate at 1 × 10⁴cells/well. The magnitude of cytotoxic response at different effector/target (E/T) ratios was evaluated by a lactate dehydrogenase (LDH) release assay, and percentage of specific lysis was calculated as follows: (experimental LDH release − effector cells − target spontaneous LDH release)/(target maximum LDH release) × 100.

Enzyme-linked immunosorbent assay (ELISA)

Cytokine (IL-12p70, TNF-α, IFN-γ and IL-2) levels secreted by immune cells were detected using ELISA kits following the manufacturer's instructions. The concentrations of these cytokines were determined according to a standard curve.

Assessment of LEXPD_-L1si efficacy in vivo

To evaluate the protective effect of exosome vaccines, 100μL PBS (blank control), 10 μg LEX, 10 μg LEX_GFP or 10 μg LEX_PD-L1si was injected subcutaneously (s.c.) into the inner side into the inner side of the right hind limbs of DBA/2 6–8-week-old female mice on Day 0. Immunization was boosted twice on Days 7 and 14. On Day 21, the immunized mice were challenged with L1210 cells (0.5 × 10⁶ cells/mouse) injected s.c. into the lateral thigh. The survival rate of the mice was recorded every 2 days after tumor inoculation; tumor size, calculated as length × width² xπ /6, was also measured every 2 days. Percentage of increase life-span (%ILS) was calculated as follows: 100 x (the median day of death of the exosome treated tumor bearing mice—median day of death of mice in PBS group)/ median day of death of mice in PBS group. Percentage of tumor growth delay (%TGD) on Day 20 was calculated as follows: 100–100 x (the mean tumor volume for the mice treated by exosomes/the mean tumor volume for the mice in PBS group).

To evaluate the therapeutic immune effect of our exosome-based vaccine, we pre-established a tumor-bearing mouse model by s.c. inoculating L1210 cells (0.5 × 10⁶ cells/mouse) into the lateral part of the right thigh on Day 0. Then, 10 μg LEX, 10 μg LEX_GFP or 10 μg LEX_PD-L1si was injected s.c. into the inner side of the right thigh on Day 5. An identical treatment regimen was performed on Days 10 and 15, and PBS was used as blank control. Tumor-bearing mice were monitored every 2 days to evaluate the therapeutic efficacy of the vaccines. Survival rate and tumor size were recorded as described above.

In vivo depletion of T cell subsets

DBA/2 mice were challenged with L1210 cells (0.5 × 106 cells/mouse) into the lateral part of the right thigh on Day 0. Then, 10 μg LEX_PD-L1si was injected s.c. into the inner side of the right thigh on Day 5,10 and 15. For T cell selective depletion, mice were injected intraperitoneally with 500 µg of anti-CD8 monoclonal antibody clone 2.43, anti-CD4 monoclonal antibody clone GK1.5 and a rat isotype control IgG clone LFT-2 one day before vaccination with LEX_PD-L1si. Another three injections of antibodies were performed with a 2-day interval. The depletion of individual cell type was evaluated by flow cytometry and was found > 90% specific depletion in splenocytes (Supplementary Fig. 1). Tumor volume was monitored as described above.

Statistical analysis

All experiments were performed in triplicate. Data are presented as mean ± SD or SEM. The log-rank test was used to analyze survival data, and differences between the two groups were analyzed by Student’s t test. Statistical significance was determined at p < 0.05.

Results

Gene-mediated PD-L1 silencing in leukemia cells decreases PD-L1 expression in LEX

L1210 cells were transduced with a scramble or PD-L1 shRNA-incorporated lentivirus. It has been shown that the gray value ratios of PD-L1/internal control in L1210, L1210_GFP and L1210_PD-L1si were 0.9100 ± 0.1200, 0.8600 ± 0.0900 and 0.2200 ± 0.0320, respectively (p < 0.05). Besides, flow cytometry analysis also revealed that PD-L1 protein on the cell surface was markedly reduced in L1210_PD-L1si, indicating that PD-L1 protein was enriched in L1210 cell and could be significantly downregulated by stable transduction of L1210 cells with PD-L1 shRNA-modified lentiviral vector. Moreover, we further analyzed the properties of exosomes derived from PD-L1-silenced L1210 cells (LEX_PD-L1si) by electron microscopy, NTA, flow cytometry and Western blotting. As shown in Fig. 1c, electron microscopy illustrated that both LEX and LEX_PD-L1si were ranging from 30 to 120 nm in diameter and had the dimpled, cup-shaped characteristic morphology, which were consistent with previously reported exosomal morphologic characteristics30. Western blotting showed that expression of specific exosomal markers HSP70, TSG101 and CD9 was abundant in all types of exosome preparations (Fig. 1d, calnexin was used as a negative control). The gray value ratios of PD-L1/CD9 in LEX, LEX_GFP and LEX_PD-L1si were 0.8400 ± 0.1200, 0.9200 ± 0.1400 and 0.1800 ± 0.04, respectively (Fig. 1d; p < 0.05). Flow cytometry indicated that PD-L1 expression in LEX_PD-L1si was significantly downregulated compared with those of non-modified LEX and LEX obtained from GFP-transduced L1210 cells (LEX_GFP)(Fig. 1e; p < 0.05;). The NTA revealed that the size of LEX, LEX_GFP and LEX_PD-L1si was 104.5 ± 16.2 nm, 108.5 ± 22.2 nm and 102.2 ± 25.7 nm. No significant difference in size and concentration of LEX_PD-L1si was observed compared with LEX and LEX_GFP (Fig. 1f).

Fig. 1.

Characterization of LEX_PD-L1si a Comparison of PD-L1 protein expression in L1210 cells, L1210 cells transduced with control vector (L1210_GFP) and L1210 cells transduced with PD-L1 shRNA modified vector (L1210_PD-L1si), as assessed using Western blotting. b Representative images show flow cytometric analysis of PD-L1 expression on L1210, L1210_GFP and L1210_PD-L1si cells. c Exosomes derived from L1210_PD-L1si cells were visualized by electron microscopy and are visible as dimpled micro-vesicles, ranging between 40 and 100 nm. Scale bar is 100 nm. d Expression pattern of PD-L1 and typical exosome markers HSP70, TSG101 and CD9, in exosome preparations. Calnexin was used as a negative control e Membrane-bound PD-L1 protein levels in LEX, LEX_GFP and LEX_PD-L1si were measured using flow cytometry. All experiments were performed in triplicate. One representative experiment is shown. (f)The size distribution of exosomes determined by nanoparticle tracking analysis.*p < 0.05 and **p < 0.01 denote statistically significant differences. Data are representative of three independent experiments

LEX_PD-L1si efficiently promotes maturation and function of dendritic cells

DCs are indispensable for antigen presentation during T cell priming, which is critical for anti-leukemia immunity [31, 32]. Therefore, we explored the influence of LEX_PD-L1si on DC phenotype and function. The expression of CD86, CD80 and MHC-II on DCs is essential for antigen presentation and T cell activation [33]. It has been shown that immature DCs (imDCs) express a relatively low level of CD86, CD80 and MHC-II and secrete a scant amount of IL-12p70 and TNF-α(Fig. 2a, b, c). Following incubation with the three types of exosomes (10 µg/ml) for 24 h, DC surface expression of CD86, CD80 and MHC-II was markedly upregulated. Stimulation with LEX_PD-L1si exerted the most significant effects on upregulating CD86, CD80 and MHC-II expression on the DC surface(Fig. 2a, p < 0.05). DC-produced pro-inflammatory factors, IL-12p70 and TNF-α, act as essential elements to block T cells to differentiate into effectors [33]. It has been shown that DCs in the LEX_PD-L1si-stimulated group secreted significantly more IL-12p70 and TNF-α compared with secretion levels of DCs in the LEX- and LEX_GFP-stimulated groups (Fig. 2b, c, p < 0.05). Mixed lymphocyte reaction (MLR) assay showed that LEX_PD-L1si-stimulated DCs acted as more potent inducers of T cell proliferation than LEX- or LEX_GFP-treated DCs at stimulator/responder ratios of 1:5 and 1:20 (Fig. 2d, p < 0.05). These results suggest that LEX_PD-L1si promoted maturation and function of dendritic cells more efficiently than the other two exosomes.

Fig. 2.

Co-incubation with LEX_PD-L1si effectively promotes phenotypic and functional maturation of DCs. Bone marrow-derived DCs(BMDCs) were co-incubated with 30 µg LEX, LEX_GFP or LEX_PD-L1si for 24 h. BMDCs stimulated with PBS were used as negative controls, while BMDCs stimulated with LPS (1 ug/ml) were used as positive controls. a Expression levels of CD86, CD80 and MHC-II b IL-12p70 c and TNF-α secretion level of the supernatant of each group of exosomes co-incubated with DCs, as measured by ELISA. d The effect of each type of exosome on the capacity of DCs from 8-week-old DBA/2 female mice to stimulate proliferation of allogeneic T-lymphocytes. *p < 0.05 and **p < 0.01 denote statistically significant differences. Data are representative of three independent experiments and are expressed as mean ± SEM

Immunization with LEX_PD-L1si promotes T cell activation and antigen-specific CTL response

It has been proved that exosomal PD-L1 suppresses T cell activation and function26. In our study, we examined whether LEX_PD-L1si could reverse immunosuppression and restore anti-tumor immunity. After pre-labeling our exosomes with CFSE, we co-cultured them with splenic T cells for 2–12 h and then analyzed exosome internalization efficiency by T cells. After 8 h of co-incubation, CFSE-positive T cells were observed using confocal fluorescence microscopy (Fig. 3a). Furthermore, exosomal uptake efficiency was detected by flow cytometry at different time points. As shown in Fig. 3b, CFSE-positive T cells (8.1 ± 2.7%) were confirmed as early as 2 h after incubation, and the percentage of CFSE-positive T cells reached a plateau at 12 h after incubation. Therefore, these results implies that besides influencing T cell activation through DC maturation, exosomes could directly act on T cells.

Fig. 3.

LEX_PD-L1si effectively promotes T cell activation and antigen-specific CTL response. a First, 10⁵/ml splenic T cells were incubated with 20 μg CFSE-labeled LEX_PD-L1si for 8 h. CFSE-positive T cells were then detected using confocal fluorescence microscopy. Splenic T cells incubated with 20 μg CFSE-unstained LEX_PD-L1si was used as a negative control. b Splenic T cells were co-incubated with CFSE-labeled LEX_PD-L1si for 1–12 h. Time-dependent curve of percentage of CFSE-positive T cells was constructed based on flow cytometry data. c DBA/2 mice were immunized subcutaneously with 100 μL PBS or 10 μg each exosome type three times at 1-week intervals. At Day 7 after the last immunization, splenic T cells obtained from immunized mice were co-incubated with irradiated L1210 cells or p388 cells for 72 h. d T cell proliferation was evaluated using ³H thymidine incorporation. d and e IFN-γ and IL-2 secretion levels in splenic CD4⁺ T cells isolated from immunized mice in each group were detected by ELISA. f Splenic CD8⁺T cells obtained from immunized mice were re-stimulated with irradiated L1210 (4000 rad) cells in the presence of mIL-2 in vitro for 7 days. The separated viable CD8⁺ T cells served as effector cells. L1210 or p388 cells served as target cells and were mixed with effector cells at different ratios. Data were obtained using three independent experiments and are expressed as mean ± SD. *p < 0.05 and **p < 0.01 denote statistically significant differences

After revealing the favorable exosomal uptake efficiency by T cells, we then focus on the effects of LEX_PD-L1si on T cell activation and function by analyzing splenic T cells obtained from mice immunized with LEX, LEX_GFP or LEX_PD-L1si. As shown in Fig. 3c, 100μL PBS (blank control), 10 μg LEX, 10 μg LEX_GFP or 10 μg LEX_PD-L1si was injected subcutaneously (s.c.) into the inner side into the inner side of the right hind limbs of DBA/2 6–8-week-old female mice on Day 0. Immunization was boosted twice on Days 7 and 14. On Day 21, splenic T cells from the immunized mice were isolated for the detection of T cell proliferative response, Th1 response and CTL activity. Immunization with LEX, LEX_GFP or LEX_PD-L1si promoted T cell expansion in response to challenge, with L1210 cells used as specific targets. Immunization with LEX_PD-L1si exerted the strongest effects on boosting T cell proliferation (Fig. 3d, p < 0.05). We also measured cytokine production in splenic CD4⁺ T cells, which is indicative of CD4⁺ T cell activation. The results of our ELISA assay showed that splenic CD4⁺ T cells obtained from mice immunized with LEX_PD-L1si secreted the highest levels of IFN-γ and IL-2 compared with those secreted by CD4⁺ T cells obtained from mouse immunized with LEX and LEX_GFP groups (Fig. 3e and f). To further assess the CTL activity elicited by the optimized vaccine, we analyzed splenic CD8⁺ T cells obtained from mice immunized with LEX, LEX_GFP or LEX_PD-L1si. Tumor-specific CTL response in CD8⁺ T cells was evaluated by a LDH release assay. Notably, LEX_PD-Lsi immunized CD8⁺ T cells exhibited the highest lysis rate in response to L1210 cells (p < 0.05). However, the CTL response induced by LEX_PD-L1si did not show significant eradicating activities against p388 cells, indicating that superior CTL activity induced by LEX_PD-L1si is likely tumor-specific.

LEX_PD-L1si exerts immuno-protective effects against tumor challenge in vivo

Next, we evaluated the immuno-protective effects of LEX_PD-L1si in our mouse model. DBA/2 mice were vaccinated thrice with either PBS, LEX, LEX_GFP or LEX_PD-L1si at 7-day intervals(each group contained 10 mice). The mice were subcutaneously challenged with L1210 cells 7 days after the last vaccination (Fig. 4a). Tumor growth and survival rate were monitored daily during the 24-day observation period. As shown in Fig. 4b, vaccination with LEX and LEX_GFP showed moderate inhibitory effects on tumor growth compared with those in the PBS-treated group, whereas immunization with LEX_PD-L1si performed substantially better than that with LEX or LEX_GFP in delaying tumor growth. As depicted in Table 1, LEX_PD-L1si immunization promoted in vivo tumor growth delay (%TGD 49.8 ± 3.3%) more efficiently than LEX (%TGD 28.1 ± 6.2%, p < 0.01) or LEX_GFP (%TGD 32.9 ± 4.7%, p < 0.01). Accordingly, the survival of tumor-challenged mice was prolonged most significantly by LEX_PD-L1si vaccination. As shown in Fig. 4c, mice in the PBS-treated group died in 24 days after the tumor challenge. The median survival time (MST) of mice in the PBS-treated group was 20 days. Immunized with LEX or LEX_GFP could prolong MST to 26–28 days (% ILS:30–40%). However, MST was prolonged to 32 days post tumor challenge by vaccination with LEX_PD-L1si(% ILS:60%). These results indicate that LEX_PD-L1si induced a stronger protective immune response against leukemia cells than that induced by LEX or LEX_GFP.

Fig. 4.

LEX_PD-L1si immunization induces potent anti-leukemia preventive immunity. a Female DBA/2 mice, 6–8-week-old, were immunized with 10 μg of each exosome type or injected with 100 μl PBS, on day 0 (prime), 7 (booster I) and 14 (booster II). On day 21, mice were challenged with 5 × 10⁵ L1210 cells subcutaneously. Each group contained 10 mice. b Tumor volume was measured using calipers after the tumor challenge every 2 days. c Survival rate of immunized mice for up to 50 days after tumor challenge. All experiments were performed in triplicate. *p < 0.05 and **p < 0.01 denote statistically significant differences

Table 1.

The protective effect of LEX_PD-L1si vaccines

parameters	PBS group	LEX group	LEXGFP group	LEXPD-L1 si group
MST(days)	20.0	26.0	28	32.0
%TGD	–	28.1 ± 4.2	32.9 ± 4.7	49.8 ± 3.3a*
% ILS	–	30	40	60

MST: median survival time; %TGD: percentage of tumor growth delay on Day 20; % ILS: percentage of increased life span.

^a: LEX_PD-L1si group vs LEX_GFP group or LEX group. Data are representative of three independent experiments. %TGD was expressed as mean values ± SEM.

*p < 0.05 and **p < 0.01 denote statistically significant differences

LEX_PD-L1si induces a robust therapeutic effect against leukemia cells in vivo

Next, we examined whether LEX_PD-L1si could induce therapeutic anti-tumor effects against established tumors in vivo. For this, 5 × 10⁵ L1210 cells were pre-inoculated subcutaneously into each mouse on Day 0. Then, tumor-bearing mice were injected with different formulations on Days 5, 10 and 15 (Fig. 5a, each group contained 10 mice). Tumor mass was measured for 22 days after tumor inoculation. Our results indicate that all the exosome formulations examined in our present study significantly inhibited the growth of pre-established tumors. LEX_PD-L1si inhibited tumor growth more effectively than LEX or LEX_GFP (Fig. 5b; Table 2, %TGD 50.1 ± 4.1, p < 0.05). Moreover, vaccination with LEX_PD-L1si improved the survival rate and prolonged the MST of tumor-bearing mice significantly compared to vaccination with either LEX or LEX_GFP (Fig. 5c, Table 2). 20% of the mice in the LEX_PD-L1si-treated group were alive at 6 weeks after tumor inoculation, whereas mice in the PBS group died in 24 days, and mice in the LEX or LEX_GFP groups all died within 32 days (Fig. 5c). These results suggest that LEX_PD-L1si prolonged the survival of tumor-bearing mice more efficiently than LEX or LEX_GFP.

Fig. 5.

LEX_PD-L1si immunization induces robust anti-leukemia therapeutic immunity. a Female DBA/2 mice, 6–8-week-old, were subcutaneously inoculated with 5 × 10⁵ L1210 cells on Day 0 and were then vaccinated with 10 μg each exosome type or injected with 100 μl PBS, on Day 5 (prime), 10 (booster I) and 15 (booster II). Each group contained 10 mice. b Tumor volume was measured using calipers from Day 6 to Day 22. c Survival rate of tumor-bearing mice was recorded from Day 6 to Day 50. *p < 0.05 and **p < 0.01 denote statistically significant differences.(d) Assessing the role of CD4⁺ and CD8⁺ T cells in the LEX_PD-L1si induced therapeutic immunity against L1210 cells. DBA/2 mice were challenged with L1210 cells (0.5 × 106 cells/mouse) into the lateral part of the right thigh on Day 0. Then LEX_PD-L1si (10 μg/mouse) was injected s.c. into the inner side of the right thigh on Days 5, 10 and 15. For selective depletion, mice were injected intraperitoneally with 500 µg of GK1.5 (anti-CD4), 2.43 (anti-CD8) or normal rat IgG one day before vaccination. Another three injections of antibodies were performed with a 2-day interval. Normal rat IgG was used as control antibody and the tumor volume was measured using calipers. All experiments were performed in triplicate. *p < 0.05 and **p < 0.01 denote statistically significant differences

Table 2.

The therapeutic effect of LEX_PD-L1si vaccines

parameters	PBS group	LEX group	LEXGFP group	LEXPD-L1 si group
MST(days)	20.0	28.0	26.0	31.0
%TGD	–	23.1 ± 4.7	31.7 ± 5.2	50.1 ± 4.1*
% ILS	–	40	30	55

MST: median survival time; %TGD: percentage of tumor growth delay on Day 20; %ILS: percentage of increased life span.

^a: LEX_PD-L1si group vs LEX_GFP group or LEX group. Data are representative of three independent experiments. %TGD was expressed as mean values ± SEM.

*p < 0.05 and **p < 0.01 denote statistically significant differences

As observed above, immunization with LEX_PD-L1si could activate more potent specific Th1 and CTL response. Moreover, in order to further investigate the roles of CD4⁺T cells and CD8⁺T cells in anti-leukemia immunity induced by LEX_PD-L1si immunization in therapeutic models, the DBA/2 mice were depleted of CD4⁺T cells or CD8⁺T cells one day before LEX_PD-L1si vaccination, and tumor mass was measured. It has been shown that depletion of CD8⁺T cells blocked the antitumor immunity induced by LEX_PD-L1si. Besides, depletion of CD4⁺T cells could compromise antitumor immunity tumor to a certain extent (Fig, 5d). These data indicated that CD8⁺T cells, as the main effector cells, play a requisite role in the antitumor immunity induced by LEX_PD-L1si, and CD4⁺T cells which can provide help for CD8⁺T cell induction were also involved in antitumor response induced by LEX_PD-L1si immunization.

Discussion

Exosomes derived from tumor cells, including leukemia cells, are a rich source of tumor antigens; these antigens originate from parental cells and reflect the tumor content and activities of these parental cells [34, 35]. TEX carrying tumor-associated antigens can act as potent inducers of the immune response [36]. Therefore, TEXs were expected to be a promising cell-free anti-cancer vaccine. However, studies using animal models and clinical trials have shown that treatment with non-modified TEXs does not induce an effective anti-tumor CTL response that specifically eliminates tumor cells. Therefore, improving the efficacy of TEX-based tumor vaccines remains a challenge. Emerging evidence has shown that TEXs are enriched in immunosuppressive factors that inhibit the immune response and even facilitate tumor evasion, thereby impeding the utility of TEXs in immunotherapy [37, 38]. In our previous studies, we confirmed that LEXs are enriched in immunosuppressive factors such as TGF-β1 and PD-L1 [19], which is similar to the immunosuppressive-factor content in other tumor-derived exosomes. To improve the immunogenicity of LEX-based ALL vaccines, we also modulated exosomal components using genetic modification of parental tumor cells. We found that exosomes obtained from TGF-β1-silenced leukemia cells induced a more potent anti-tumor immune response than that induced by non-modified LEXs [19]. In our current study, we show that exosomes from PD-L1-silenced leukemia cells robustly promoted DC maturation and function, induced T cell activation, and facilitated an effective and antigen-specific CTL response.

PD-L1 is a typical immune checkpoint molecule that is highly expressed in tumor cells [39]. PD-L1 inhibits T cell anti-tumor activities by binding to the PD-1 receptor on the surface of activated T cells, thereby playing a critical role in tumor immunosuppression [40]. Tumor-derived exosomes also carry PD-L1 on their surface; exosomal surface-membrane topology is the same as that of their parental cells [37, 41]. TEXs, which carry PD-L1 on their surface, are responsible for suppressing T cell function and decreasing the frequencies of TILs [37, 42]. Moreover, TEXs enriched in PD-L1 can migrate to PD-L1-negative tumor and immune cells, thereby augmenting both local and systemic immunosuppression and even promoting tumor growth by engaging with PD-1 [43]. For these reasons, blockade of exosomal PD-L1 may be a novel therapeutic strategy for improving anti-tumor immunity and inhibiting tumor evasion. However, evidences have been shown that exosomal PD-L1 resist to the already approved antibodies to block the PD-L1/PD-1 [37]. For example, Yu et al. have demonstrated that the TRAMP-C2 prostate cancer model is resistant to current anti-PD-L1/PD-1 antibody. In contrast, genetic blockade of PD-L1 had a striking effect [44]. Similarly, the MC38 murine colon carcinoma model shows only partial responsiveness to anti-PD-L1 therapy, while deletion of the PD-L1 exhibited a more potent effect t [37]. The reason of exosomal PD-L1 resistance to current anti-PD-L1/PD-1 antibody blockade was still unclear. It is possible that how PD-L1 is presented on the TEX makes it less responsive to the current antibodies. Besides, It is also possible that exosomal PD-L1 may be produced at high enough levels that it can compete with the delivered antibody. In our study, we aimed to block the immunosuppressive effects of exosomal PD-L1 by downregulating exosomal PD-L1 expression through genetic blockade of PD-L1 in parental cells. Our results indicate that LEXs derived from PD-L1-silenced leukemia cells expressed a significantly lower level of PD-L1 than non-modified LEXs, demonstrating that artificially modulating exosomal expression of PD-L1 via genetic modification is a feasible and straightforward strategy. LEX_PD-L1si also expressed the typical exosomal markers and morphologic characteristics, indicating that genetic modification of parental cells did not affect exosomal biological properties.

Tumor antigen-pulsed dendritic cells (DC), capable of triggering antigen-specific T cell activation, play an essential role in the initiation and modulation of anti-tumor immune responses [45]. The maturation status of DCs determines their immunological potency (enhancing anti-tumor immunity or promoting immunologic tolerance). Although immature dendritic cells can participate in antigen uptake and processing, they cannot provide the signals required for the initiation of T cell response. By contrast, mature DCs induced by external signals can migrate to secondary lymphoid organs, and upregulate immunogenicity to initiate T cell response [46]. Moreover, DC-based anti-tumor immune responses can be regulated by TEXs [47]. PD-L1 on TEXs mediates suppression of DC maturation and blockage of DC-regulated T cell activation, thereby promoting tumor immune escape [47]. Our results indicate that stimulation with LEX_PD-L1si partially reversed DC tolerance by promoting DC maturation and pro-inflammatory factors production, thereby enhancing the capacity of DCs to stimulate T cell activation. These results suggest that depleting PD-L1 from LEXs may be a potential strategy for enhancing the immunological potency of DCs.

T cells play a crucial role in TEX-induced anti-tumor immunity. Previous studies have shown that directly combining T cell PD-1 receptors with their corresponding PD-L1 ligands on tumor cells or TEXs can downregulate the amplitude of T cell activation and induce T cell dysfunction, leading to tumor-cell immune escape [48]. In our present study, we show that LEX was efficiently up taken and internalized by T cells in vitro, demonstrating that the highly biologically active membrane-form of PD-L1 on LEXs can exert a direct suppressive effect on T cell activation.

Furthermore, LEX_PD-L1si outperformed non-modified LEX in inducing T cell proliferation and promoting the secretion of Th1 cytokines in an antigen-specific manner. These results suggest that downregulation of PD-L1 expression on exosomal surfaces can effectively reverse the negative immune effects of LEXs on T cells and promote T cell activation. Moreover, LEX_PD-L1si effectively induced a CTL response. Compared with non-modified LEX, LEX_PD-L1si potentiated a stronger antigen-specific cytotoxic response, which directly contributed to inhibition of leukemia cell growth.

Having shown the effectiveness of exosomes in promoting the function of DCs and T cells, we examined the anti-leukemia effects of LEX, LEX_GFP and LEX_PD-L1si in vivo. In various tumor models, TEXs has been proved to be pro-tumoral factors with the capability of suppressing the function of NK cells and T cells, promoting T regulatory cell expansion, and facilitating epithelial-mesenchymal transition (EMT) [49, 50]. However, evidence has also shown that TEXs carry a cargo of multiple stimulatory factors, such as MHC molecules, chaperones and TAA [51, 52]. The anti-cancer efficacy of TEXs has been proved in several types of tumors, including lymphoma and leukemia [34]. For example, Menayet et al. demonstrated that T cells from TEXs-immunized mice secrete interferon-γ in response to tumor stimulation [53]. Administration of the TEXs into mice induces a tumor-specific immune response [53] Consistent with the previous study [21], our study demonstrated that exosomes from L1210 cells can induce a specific anti-leukemia immunity and partly retard tumor growth in tumor-bearing mice. Bu et al. found that interstitial cell adhesion molecule (ICAM)-1 and HSP70 were enriched in L1210 derived exosomes [54]. The presence of ICAM-1 and Hsp70 on the exosomal surface has important functions. It can help antigenic peptide folding, transporting and T cell activation, thereby inducing an elevated anti-leukemia immune response [55, 56]. Therefore, TEXs is a sword with double blades in tumor immunity. It can exert either immunosuppressive effect or immunostimulatory effect. The heterogeneity of exosome phenotypic characteristics may account for the difference of immunomodulatory function of TEXs.

Besides, our results demonstrated that LEX_PD-L1si exhibited a more powerful protective effects by attenuating tumor growth and prolonged the survival time of mice in vivo, indicating that LEX_PD-L1si mediated superior anti-leukemia immunity as a preventive tumor vaccine. To evaluate therapeutic effect of exosome vaccines, tumor-bearing mice model was established in which leukemia cells injected on day 0 would constantly release LEX carrying PD-L1. Amazingly, vaccination with LEX_PD-L1si reversed the immune dysfunction state in vivo and induced robust anti-leukemia therapeutic immunity. Through in vivo T cell depletion, we confirmed that CD4⁺ T cells and CD8⁺ T cells were both involved in inducing LEX_PD-L1si therapeutic efficacy, and the CD8 + T cells were the main effector cells in anti-leukemia immunity.

As to the underlying mechanisms responsible for the superior effect of LEX_PD-L1si in therapeutic models, we proposed the following possible reasons: First, LEX_PD-L1si vaccination may facilitate the maturation and function of DCs in antigen presentation and processing stage. DCs can influence the strength or quality of CD4⁺ and CD8⁺ T cell responses, thus triggering stronger anti-tumor immunity in vivo. Second, LEX_PD-L1si themselves may contain PD-L1 siRNA and/or PD-L1 shRNA from gene-modified L1210 cells by which exosomes may exert their activities. It has been reported that exosomes contain molecular particles from their parental cells, including RNAs and proteins [57]. Meanwhile, exosomes can be an ideal vehicles for siRNA delivery [58]. Vaccination with LEX_PD-L1si which may loaded with PD-L1 siRNA and/or PD-L1 shRNA from gene-modified L1210 cells might affect the immunophenotype of tumor cells and immune cells which could uptake the LEX_PD-L1si in vivo, thus activating anti-leukemia immunity in animal models. However, our assumption needs to be further elucidated in the further studies.

Conclusions

In summary, we extensively characterized LEX_PD-L1si and showed that LEX_PD-L1si can be developed into an effective LEX-based vaccine for inducing anti-leukemia immunity via hyper-activation of DCs and T cells. This study offers a novel strategy for optimizing the immunogenicity of LEX-based tumor vaccines.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

ALL

Acute lymphoblastic leukemia

CTL

Cytotoxic T lymphocytes

DC

Dendritic cells

HSP

Heat shock protein

IL

Interleukin

IL-2

Interleukin 2

IFN-γ

Interferon-γ

LEX

Leukemia-derived exosome

MHC

Major histocompatibility complex class

PD-1

Programmed death-1

PD-L1

Programmed cell death 1 Ligand 1

TEX

Tumor-derived exosomes

TNF-α

Tumor necrosis factor-α

Authors' contributions

Fang Huang performed the experiments and prepared the manuscript. Zhichao Li analyzed the data. Wenhao Zhang and Jiaqi Li reviewed the data. Siguo Hao designed the study. All authors read and approved the manuscript.

Funding

This research was supported by the National Natural Science Foundation of China (Grant No. 81470314 and 81873435).

Declarations

Ethics approval and consent to participate

All animal experiments were conducted according to the guidelines of the Ethics Committee of Xinhua Hospital Affiliated to the Shanghai Jiao Tong University School of Medicine, Shanghai, China.

Data availability

The datasets used during the current study can be obtained from the corresponding author upon reasonable request.

Competing interests

The authors declare that they have no competing interests.