Carboxyfluorescein diacetate succinimidyl ester (CFDA-SE) was from Invitrogen™ Life Technologies. 4-20% MP TGX Stain-Free Gel, Trans-Blot Turbo Midi PVDF transfer packs, and ECL Substrates for High-Sensitivity Western Blot Detection were purchased from Bio-Rad. Anti-β actin (C4) (sc-47778), anti MHC-I (F3) (sc-32235) were purchased from Santa Cruz Biotechnology. Anti HLA-ABC (polyclonal) (PA5115364) was obtained from Thermo Fischer Scientific. Anti β2-microglobulin (EP2978Y) (ab75853) was from Abcam. Ovalbumin peptide (257–264) was purchased from Sigma-Aldrich, and ovalbumin was purchased from InvivoGen.

AGS human gastric adenocarcinoma cells were purchased from the European Collection of Cell Cultures (ECACC). CT26.WT undifferentiated colon carcinoma cells, 4T1 murine breast cancer cells and J774A.1 murine monocytes were purchased from American Type Culture Collection (ATCC). The T2K^b cell line is a human T2 cell line transfected with mouse H-2K^b class I genes. T2K^b cells express empty H-2 class I on their surface and can efficiently present exogenous peptides to murine cytotoxic T lymphocytes (10). In this experimental model, we used the OT-I murine CTL cell line, K^b-restricted, specific for the OVA 257-264 peptide (pOVA) of ovalbumin protein (11). The T2K^b and OT-I cell lines were kindly gifted by Prof. C.T. Baldari, University of Siena.

CT26, 4T1, and HDF were cultured in DMEM supplemented with 2 mM glutamine, penicillin (100 U/mL), streptomycin (100 μg/mL), and 10% fetal bovine serum (FBS, Euroclone). AGS cells were cultured in RPMI supplemented with 2 mM glutamine, penicillin (100 U/mL), streptomycin (100 μg/mL), and 10% fetal bovine serum (FBS, Euroclone). J774A.1 cells were cultured in DMEM supplemented with 2 mM glutamine, penicillin (100 U/mL), streptomycin (100 μg/mL), and 10% HyClone defined fetal bovine serum (HyClone defined FBS, Cytiva).

T2K^b cells and OT-I were cultured in RPMI 1640 complete medium (supplemented with 2 mM L-glutamine, penicillin 100 U/mL, streptomycin 100 μg/mL) with 7.5–10% HyClone defined fetal bovine serum (HyClone defined FBS, Cytiva). Cells were incubated at 37°C in a humidified atmosphere containing 5% CO₂. Human-derived monocytes and CD8⁺ T cells were obtained from a voluntary donor after tetanus-toxoid (TT) immunization.

The transfer of proteins from the donor to recipient cells was evaluated using CFDA-SE. Donor cells were labeled with the dye at a concentration of 10 μM in PBS buffer for 15m, then detached and plated with recipient cells both in co-culture and in Transwell^® inserts. For flow cytometry analysis, cells were detached after 24h or 40h of co-culture, fixed in 3% paraformaldehyde, and analyzed using a BDFACS Canto-II. The ability of donor cells to transfer proteins was evaluated by measuring the ratio of the fluorescence of recipient cells after co-culture with donor cells and their autofluorescence.

Cancer cells were grown in starvation medium without FBS and after 24h the medium was collected, centrifuged at 1500 ×g for 10 min to discard cell debris and filtered through 0.2 µm size pore filters.

Activation of monocytes, both lineage J774.A1 or T2K^b and healthy donor-derived monocytes, was done by treating them for 24 hours with tumor conditioned media (t.c.m.) derived from cancer cell lines (AGS or 4T1), as described above.

MVs were purified as described by Santi et al. (2015) (8). Briefly, activated and non-activated monocytes (40 × 10⁶ cells) were cultured for 24 h in starvation medium without FBS, then their supernatant was centrifuged at 1000 ×g for 5 min to discard the cells. The supernatant was then centrifuged at 1500 ×g for 10 min to remove cell debris. The supernatant was centrifuged at 10,000 ×g for 45 min to isolate the MVs fraction. The pellet was resuspended in PBS and washed via centrifugation at 10,000 ×g for 45 min and finally resuspended in culture medium.

Nanoparticle tracking analysis (NTA) measurements were performed with a NanoSight NS300.

MVs isolated from 40x10⁶ J774A1 and from 8x10⁶ T2Kb as described above were diluted in PBS to a final volume of 1 ml. The samples were diluted to reach the optimal concentration of particles per frame value. The instrument was set up in accordance with the manufacturer’s software manual (NTA 3.4 Build 3.4.4): camera level was increased until all particles were distinctly visible not exceeding a particle signal saturation over 20%. The ideal detection threshold was determined to include as many particles as possible with the restrictions that 10–100 red crosses were counted while only ~10% were not associated with distinct particles. Blue cross count was limited to 5. For each measurement, five 1-min videos were captured under the following conditions: cell temperature: 25°C; syringe speed: 30 µl/s; laser: green; camera: sCMOS.

For SDS-PAGE and Western blot analysis, MVs were lysed in 30 µL Laemmli electrophoresis buffer (without β-mercaptoethanol and bromophenol blue) and assayed for protein content using the Bicinchoninic Acid protein assay (BCA). Each sample (25 μg) was supplemented with β-mercaptoethanol and bromophenol blue and separated using SDS-PAGE. The gels were then electroblotted onto PVDF membranes for detection. Blots were incubated with anti-β-actin, anti HLA, anti MHC-I, and anti-β2 microglobulin. After incubation with secondary antibodies, the blots were developed using the ECL plus immunodetection system and visualized using Amersham Imager 600.

The experiment reported in Table 1 was performed by activating T2K^b in the CT26.WT cancer cell conditioned medium. Activated T2K^b cells were then preloaded with OVA peptide (pOVA) (10μg/ml) and co-cultured with CT26.WT cells for 24h. OVA-specific OT-I CD8⁺ T cells were then added to the T2K^b - CT26.WT co-culture for an additionally 18h in the indicated rows. Thereafter, T2K^b and CT26.WT cell viability was determined by flow cytometry analysis (ViobilityTM 488/520 fixable dye, FITC) of specific cell subpopulations using anti-CD7-PE and/or anti-CD3e-Vioblue. To inhibit cytotoxic activity, cells were incubated with anti-mouse H-2K^b (BioLegend, USA), 1h before exposure to pOVA. In a similar experiment ( Table 2 ), monocytes were isolated from a blood sample of a healthy human donor immunized against tetanus-toxoid (TT). First, peripheral blood mononuclear cells (PBMCs) were purified from blood by Ficoll-Hypaque density gradient centrifugation (Lymphoprep, Alere Technologies, Oslo, Norway). Then, PBMCs were washed, counted and seeded in a cell culture flask for 2 hours in order to allow monocytes adhesion on its surface and let lymphocytes culturing in suspension. After incubation the cell suspension was harvested and subjected to CD8+ T lymphocytes purification protocol; the monocyte portion was harvested by scraping, washing and collecting. T cytotoxic lymphocytes were purified from PBMCs isolated by the donor blood sample subjected to positive selection magnetic labelling with anti-CD8 microbeads (MACS, Miltenyi biotec, Bergisch Gladbach, Germany). One million of T cytotoxic lymphocytes were then cultured in each well of a 24 cell culture plate with TT (0.5 μg/ml) and 5 x 10⁵ autologous irradiated monocytes in medium RPMI 1640 (Bio Concept Ltd., Allschwil, CH) and Human Serum (HS) 5% (Sigma-Aldrich, St. Louis, MO, USA). After 5 days 30U/ml of interleukine (IL)-2 (PeproTech EC, Ltd., London, UK) was added to promote T cell proliferation of the selected TT-specific T CD8+ lymphocytes. Ten days later, the TT specificity was analyzed by carboxyfluorescein succinimidyl ester (CFSE) (CellTrace CFSE dye, Invitrogen, USA) method, following manufacturer’s instructions. After CFSE staining, 2 × 10⁵ T CD8+ lymphocytes were incubated with or without TT (0.5 μg/ml) and the proliferative response was investigated by flow cytometry on BD FACS Canto II and analyzed using the FACSDiva software (Becton Dickinson, Franklin Lakes, NJ, USA) after 3 days of cell culture. Thus, monocytes isolated from human donor were pulsed with TT (0.5 μg/ml) for 24h and cultured with AGS (human gastric adenocarcinoma) cells for an additional 24h. Cytotoxic TT-specific autologous CD8+ T lymphocytes were added and 18h later the cell viability was determined by flow cytometry analysis (Viobility 488/520 fixable dye, FITC) of specific cell subpopulations using anti-CD44 Allophycocyanin conjugated mAbs, anti-CD8 PE, and anti-CD14 PerCP conjugated mAbs. An anti HLA class I blocking antibody (AbCam, UK) was added to the control conditions 1h before exposure to TT to test the inhibition of the restricted cytotoxic activity.

In vivo experiments were performed in accordance with national guidelines approved by the Italian ethical committee of the Animal Welfare Office of the Italian Work Ministry (aut. No. 652/2020-PR) and conformed to the legal mandates and Italian guidelines for the care and maintenance of laboratory animals.

The animals were randomized before cancer cell injection. Twelve six- to eight-week old male severe combined immunodeficient (SCID)-bg/bg mice (Charles River Laboratories International) were subcutaneously injected with 1×10⁶ PC3 cells. The mice were monitored daily until measurable tumors were formed. Six mice were intratumorally injected with 1×10⁶ T2K^b cells (previously activated with PC3 cell-conditioned media for 16h) preloaded with 10 μg/ml pOVA for an additional 4h. Six control mice were injected with vehicle (PBS). The injection volume was 100 μl. After 4h from the injection, 5x10⁶ OT-I lymphocytes were injected into the tumor mass of all mice. The treatment was repeated every four days for a total of four times. Tumors were measured with calipers and the volumes were determined using the following formula: V = (Length × Width2)/2.

Eighteen four-week old male BALB/c mice (Charles River Laboratories International) were immunized by subcutaneous injection of 10 μg of OVA. After three weeks, immunization was repeated. Seven days later, all mice were subcutaneously injected with 1x10⁵ syngeneic 4T1 tumor cells. When the tumors reached the minimum measurable size, mice were randomized. Six mice were intratumorally injected with 3x10⁶ syngeneic J774A.1 monocyte previously activated with 4T1 cell-derived conditioned medium for 16h and preloaded with 10 μg/ml of OVA for 4h, six mice were intratumorally injected with only 10 μg/ml OVA and six mice with PBS (control). The injection volume was 100 μl. The treatment was repeated every three days for a total of three times. Tumors were measured with calipers and the volumes were determined using the following formula: V=(Length × Width²)/2.

BALB/c mice fresh spleens were placed in PBS, washed twice and disrupted using gentleMACS^® Octo Dissociator (Miltenyi Biotec, Bergisch Gladbach, Germany). Then the samples were filtered through a 70µm cell strainer and erythrocytes were removed using a Red blood cell lysis solution (Miltenyi Biotec, Bergisch Gladbach, Germany). The spleen mononuclear cells were tested for their responsiveness to OVA vaccine by measuring [³H] thymidine (Perkin Elmer, Waltham, MA, US) uptake after 120 h of stimulation. In particular, 2.2x10⁵ cell/well were seeded in triplicate with medium (RPMI 1640 complete, 10% FBS) alone, or with pOVA (2 μg/ml) or OVA (2 μg/ml). A mitogenic index (MI) > 5 was considered positive.

Longitudinal mouse data on tumor volumes and weight loss were elaborated and prepared for further analyses by calculating the means of repeated technical measurements and standard deviations in Microsoft Excel. Experiments were modeled using a linear regression model robust against clustered data in R (version 4.0.3) using the lm cluster function of the miceadds package (version 3.16-18), considering the formula Y~Time²+Treatment : Time² to model the nonlinear growth of tumor size and using individual mice as clustering factors. In all experiments, statistical significance was set at p<0.001. Bar plots were prepared using the GraphPad Prism software. Box plots and spaghetti plots were constructed with R.

All data were generated by the authors and available on request.

To evaluate whether immune cells are able to transfer proteins to cancer cells, we used human T2 cells engineered to express the mouse H-2K^b MHC-I molecule (T2K^b cells) (12) or J774A.1 murine macrophages as donor cells.

Firstly, the intracellular proteins of T2K^b or J774A.1 cells were labeled with CFDA-SE, a fluorescent probe that binds to the amino groups of proteins (8). Labeled T2K^b or J774A.1 cells were then co-cultured with AGS human gastric adenocarcinoma cells or 4T1 murine mammary carcinoma cells. Flow cytometry analysis showed that T2K^b or J774A.1 cells transfer CFDA-SE-labeled proteins to AGS and 4T1 cancer cells after 6h of co-culture ( Figure 1 , columns A and C). On the contrary, CFDA-SE labeled AGS or 4T1 cells did not transfer appreciable fluorescent proteins to either T2K^b or J774A.1 cells ( Figure 1 , columns B and D). These results indicate that immune cells are able to transfer proteins, unidirectionally, to cancer cells similarly to CAFs (8, 13). Indeed, it is well established that cells of the immune system exchange plasma membrane proteins, including MHC molecules, and that EVs can be involved in this process (14, 15). At this point we were interested in determining whether MVs produced by monocytic cell lines contained the MHC-I complex, and whether MHC-I could be transferred in this way to cells other than immune cells, such as cancer cells. Firstly we have isolated MVs from the culture media of J774A.1 or T2Kb, either treated or not with t.c.m. as described in methods. Then, we have analyzed the purified fractions by nanoparticle tracking analysis (see methods): EVs from J774A.1 or T2Kb cells indicated a similar size distribution, ranging from 100 to 800 nm, with a predominance of 140 nm vesicles ( Figure 2 ). Notably, samples derived by t.c.m activated J774A.1 or T2Kb cells contain an higher amount of MVs ( Figure 2 ). Western blot analysis of MVs secreted by T2K^b or J774A.1 revealed that they contain both HLA-I and beta-2-microglobulin (B2M), which are components of MHC-I complex, and that the amount of these proteins in the MVs were increased when immune cells were activated with t.c.m. ( Figures 3A, B ).

To verify whether MVs mediated the transfer of MHC-I molecules from immune cells to cancer cells, we treated for 1h or 24 h the AGS or 4T1 cancer cells with MVs purified from activated T2K^b or J774A.1 cells. We found that the 1h treatment of cancer cells with purified MVs led to an increase in HLA-I protein levels in both types of recipient cells, while after 24 h, the HLA-I protein levels were equal to the control. In addition, this increase was not observed when cancer cells were treated for 1 h or 24 h with monocyte-derived conditioned medium depleted of MVs ( Figures 3C, D ).

These data support the evidence that MVs are involved in the transfer of MHC-I molecules from monocytes to cancer cells, whereas other components of the monocyte-derived secretome are not.

MHC-I molecules are important mediators of immune response. Specifically, many kind of immune cells (including dendritic cells, macrophages, neutrophils, etc) process exogenous antigens that are then loaded onto MHC-I molecules. The MHC-I/antigen complexes (pMHC-I) then translocate to the cell surface of these antigen-presenting cells (APCs), where they are exposed for recognition by CD8⁺ T cells (16). Activated cytotoxic CD8⁺ T cells proliferate and differentiate into effector cells, thereby executing immune function. Therefore, we evaluated whether the transfer of MHC-I, bound to a custom antigen, from immune cells to cancer cells can help CD8⁺ T cells to direct their cytotoxic activity towards cancer cells. To this purpose, we used as APCs the T2K^b cells, which were activated by t.c.m. and preloaded or not with an ovalbumin-derived peptide (pOVA), and we used as effector T cells the CD8⁺ T cells derived from C57BL/6 OT-I transgenic mice that are engineered to express a T cell receptor that recognizes the ovalbumin peptide 257-264 (SIINFEKL) in the context of H2K^b MHC-I molecule. These CD8⁺ T cells are known as OT-I cells (17). In vitro cytotoxicity tests were performed under various co-culture conditions, and for each of them, we determined T2K^b and CT26.WT cell viability by flow cytometry (see Methods) ( Table 1 ).

As hypothesized, OT-I cells were able to kill a higher number of T2Kb cells when pre-incubated with pOVA ( Table 1 ; row 4) compared to untreated cells ( Table 1 ; row 3). In addition, the cytotoxic effect of OT-I cells was reduced in the presence of anti-mouse H-2Kb blocking antibody ( Table 1 ; row 5). These results confirm the specificity and effectiveness of our in vitro model. When CT26.WT cells were co-cultured with pOVA preloaded T2Kb cells, we observed that OT-I cells exerted their cytotoxic effect not only on T2Kb cells, as described before ( Table 1 ; row 4), but also on CT26.WT cells ( Table 1 ; row 9). In contrast, the cytotoxic effect of OT-I cells on CT26.WT cells was very low when T2Kb cells and/or pOVA were not present ( Table 1 ; rows 1, 2, and 8), and the cancer cell death rate was similar to that observed when CT26.WT and T2Kb cells were co-cultured without OT-I cells ( Table 1 ; rows 6 and 7). This is not surprising, as loss or downregulation of MHC-I occurs commonly in many cancer types (18, 19). To test the specificity of the cytotoxic activity, co-cultured CT26.WT and T2Kb cells were incubated with the anti-mouse H-2Kb blocking antibody, which reduced the OT-I cell-mediated cytotoxic effect on both T2Kb and CT26.WT cells ( Table 1 ; row 10).

To further generalize and validate our approach, we performed a similar experiment using a human model. Monocytes derived from a human donor who was previously immunized against Tetanus Toxoid (TT) were incubated for 24h with or without TT and then co-cultured with AGS cells. Subsequently, human donor-derived TT-specific CD8⁺ T cells were added to the co-culture and, after 18h, monocyte and AGS cell viability was measured by flow cytometry. The results confirmed the specificity of this in vitro model by observing that donor-derived T cells induced a higher monocyte death rate when the cells were pre-incubated with TT ( Table 2 ; row 4) compared to untreated cells ( Table 2 ; row 3), and that the cytotoxic effect was reversed after incubation with an anti HLA class I blocking antibody ( Table 2 ; row 5). In accordance with previous results, T cells exerted their cytotoxic effect on AGS cells when they were co-cultured with TT pulsed monocytes ( Table 2 ; row 9). Instead, the cytotoxic effect of T cells on AGS cells was very low when monocytes and/or TT were not present ( Table 2 ; rows 1, 2, and 8), and the cancer cell death rate measured under these conditions was similar to that observed when AGS cells and monocytes were co-cultured without adding T cells ( Table 2 ; rows 6 and 7). In addition, the T cell-mediated cytotoxic effect was reduced when AGS cells and monocytes were co-cultured in the presence of an anti HLA class I blocking antibody ( Table 2 ; row 10).

These results demonstrate that APCs (such as T2Kb cells and monocytes) are necessary to allow CD8⁺-lymphocyte cytotoxic activity on cancer cells, and that this effect is mediated by their MHC-I complex.

Finally, to demonstrate the role of MVs in the transfer of pMHC-I from activated immune cells to cancer cells and in activating CD8⁺ lymphocytes, we treated CT26.WT cells with different amounts of MVs secreted from activated T2K^b cells pre-incubated with or without pOVA. After 1h, OT-I cells were added, and after an additional 16h, CT26.WT cells were detached and subjected to cytotoxicity tests ( Table 3 ). We found that OT-I cells were able to kill a higher number of cancer cells when they were treated with pMHC-I-bearing MVs ( Table 3 , rows 4 and 6) than when cancer cells were treated with MVs containing an empty MHC-I ( Table 3 , rows 3 and 5). In fact, the treatment of cancer cells with MVs containing empty MHC-I ( Table 3 , rows 3 and 5) or with pOVA alone ( Table 3 , row 2) induced cell death to the same extent as in the untreated condition ( Table 3 , row 1).

Next, we evaluated whether the ability of immune cells to transfer pMHC-I molecules to cancer cells can be exploited as a strategy for cancer treatment. PC3 cancer cells were injected subcutaneously into the flank region of 6-8-week old male SCID-bg/bg mice. The T2K^b cells were activated with the conditioned medium of PC3 cells and preloaded with pOVA, after which they were injected intratumorally. After 4 h, OT-I cells were injected into the tumor mass ( Figures 4A, B ). Tumor growth was lower in treated mice than in control mice that were not injected with T2K^b cells or pOVA ( Figure 3B ).

To further validate that the transfer of pMHC-I molecules from immune cells to cancer cells represents a novel strategy for cancer treatment, we immunized immunocompetent BALB/c mice with a subcutaneous injection of ovalbumin protein (OVA) (scheme in Figure 5A ).

At the end of the experiment, we verified the efficacy of immunization by evaluating the response of mononuclear cells in the spleen to OVA (see Methods). Next, 4T1 cancer cells were injected subcutaneously into the flank region of BALB/c mice. When tumors had reached the minimum palpable size, mice were injected intratumorally with activated J774A.1 monocyte cells that were pre-incubated with OVA. As a control, immunized mice were injected with OVA or PBS only. Our data showed that tumor growth in immunized mice treated with monocytes and OVA was slower than that in control mice injected with OVA alone or with PBS ( Figure 5B ). In fact, while treatment with OVA alone did not show significant differences with respect to PBS (p=0.58), the exposure of OVA through J774A.1 monocytes led to a significant reduction in tumor volume with respect to PBS (p<0.001). Moreover, mice treated with activated J774A.1 monocytes plus OVA were also much less susceptible to cancer cachexia compared to those treated with OVA or PBS only. In fact, treatment with OVA alone did not significantly change the weight loss over time with respect to PBS (p=0.57), while the exposure of OVA upon J774A.1 monocytes processing led to a substantial abolishment of such a detrimental cachectic effect (p<0.001) ( Figure 5C ). Furthermore, control experiments on non-pre-immunized mice were performed to evaluate the effects of the treatments. As expected, we confirmed that in the absence of pre-immunization with OVA antigen, none of the treatments elicited a beneficial effect either in terms of tumor regression ( Figure 5D ) or cachexia reduction (data not shown).