Hsp70 preparation
Recombinant human Hsp70 was purified from E. coli (strain BL21 DE3), transformed with a pMS-Hsp70 plasmid, as previously described [20]. The Hsp70 solution was further detoxified by incubation with Detoxi-Gel Endotoxin Remover Resin (Thermo Scientific, USA) and sterilized by filtration through a 0.22 µm filter (Millipore, USA). According to the E-Toxate assay (Sigma-Aldrich, USA), the level of lipopolysaccharide in the final Hsp70 preparation was lower than 0.25 U/mL. For flow cytometry and confocal experiments, Hsp70 was labeled with Alexa-555 or Alexa-488 or with NHS–biotin (Sigma, USA), according to manufacturer’s instructions.
Plasmids
Bioimaging was carried out with CT-26iRFP720 colon carcinoma cells expressing near far-red fluorescent protein (ex. 698 nm/em. 720 nm) in vivo in Balb/c and in Balb/c nude mice. The pHIV-iRFP720-E2A-Luc was purchased from Addgene, USA Packaging (psPAX2) and envelope (pMD2.G) plasmids were kindly provided by Dr. A. Tomilin (Institute of Cytology of RAS, Russia).
Cells
Mouse melanoma B16 cells were kindly provided by Prof. L. Sistonen (Turku Centre for Biotechnology, Finland), mouse colon carcinoma CT-26 cells were kindly provided by Prof. G. Multhoff (Technical University of Munchen, Germany) and HEK293FT cells were obtained from the Collection of Cell Lines of the Institute of Cytology RAS, Russia.
B16 and HEK293FT cells were grown in DMEM and CT-26 cells were grown in RPMI-1640 media supplemented with 10% heat inactivated fetal bovine serum (FBS) (HyClone, USA), 2 mM L-glutamine, 100 U/mL penicillin and 0.1 mg/mL streptomycin (PanEco, Russia) in a 5% CO2 atmosphere with 90% humidity. Viability was determined by 0.4% trypan blue exclusion.
Lentivirus production and generation of cell sublines
The HEK293FT cells were co-transfected using PEI (polyethyleneimine, linear, 25kDa, Polysciences) with a mixture of three plasmids in serum-free DMEM without antibiotics. Virus-containing supernatants were harvested after 24, 48 and 72h in complete DMEM with 10% FBS, 3 mM Sodium butyrate, 1% non-essential amino-acids (ThermoFisher, USA), 2 mM GlutaMAX (ThermoFisher, USA) and 1% sodium pyruvate (PanEco, Russia), pooled, filtered through a 0.45 µm filter and concentrated using a PEG6000/NaCl solution. The lentivirus titers were determined by counting fluorescently tagged cells according to the protocol at https://www.addgene.org/protocols/fluorescence-titering-assay/. CT-26 cells were transduced with an iRFP720 vector and were cloned in 96-well plates containing feeder of mouse peritoneal macrophages to generate individual single colonies that were analyzed further with aid of VarioSkanLUX (ThermoFischer, USA).
Analysis of Hsp70 release from B16 cells
To analyze whether B16 cells could extrude cell-Hsp70 into the extracellular milieu, we used the method described elsewhere [17, 18]. Briefly, B16 melanoma cells (1.5×107 in each probe) were incubated with 50 µg/mL biotinylated Hsp70, and after 6h the cells were washed with PBS and placed into serum-free media for the next 90 min. To elucidate which form of exo- and cell-Hsp can be released from B16 cells (i.e. in the soluble fraction or being inside/on the surface of EVs), we collected culture medium which was first centrifuged at 400 g for 5 min to remove dead cells debris, then at 20000 g for 30 min to remove big vesicles and was finally ultracentrifuged at 110000 g for 2 h and then Pellet fraction and Supernatant fraction, were used for two-step chromatography. The medium was divided and subjected to a two-step affinity precipitation, first with NeutrAvidin-agarose gel slurry (Pierce) and after the unbound fraction was incubated with ATP-agarose gel (Sigma-Aldrich) (see Fig. 1b and [18] for details). The gel probes were subjected to treatment with SDS and to western blotting. After electrophoresis in 11% polyacrylamide gel, the protein bands were transferred to a polyvinylidenedifluoride (PVDF) membrane (0,45mm, Amersham, USA) that was stained alternatively with Avidin-Peroxidase (Pierce, USA) or 3C5 anti-Hsp70 monoclonal antibody (known to recognize human, rat and mouse Hsp70 [21].
Flow cytometry
To investigate whether exo-Hsp70 in EVs or free exo-Hsp70 collected from the B16 cell medium could re-enter other naïve B16 cells, and could extrude cell-Hsp70 onto cell surface, we incubated naïve B16 cells with Hsp70 labeled with Alexa-647 (as described above) and collected EVs and soluble Hsp70 fractions. The EV pellet was diluted in 0.5 mL of complete medium, whereas the medium containing soluble Hsp70 (10 mL) was concentrated up to final volume of 0.5 mL using the Amicon Ultra-15 Centrifugal Filters Ultracel – 30K (Merck Millipore, Ireland). Media was then transferred to naïve B16 cells and after a 6h incubation cells were washed and stained with cmHsp70.1-FITC antibody. Analysis was carried out with the aid of a CytoFLEX (Beckman Coulter, USA) flow cytometer using lasers of 488 nm and 647 nm.
Quantitative analysis of the CD9 exosome marker on the surface of the isolated EVs was carried out using an Exo-FACS ready-to-use kit for analysis of exosome markers from cell culture media (Lonza, Estonia) supplied with primary antibodies against CD9 (ab92726, Abcam, UK) and secondary Alexa488-labeled antibodies, according to the manufacturer’s recommendations. The same number of vesicles was added to each sample for flow cytometry, based on the results of measurements of particle concentration using nanoparticle tracking analysis. A sample of PBS, without any EVs, was used as a negative control for non-specific labeling. An aliquot of exosome standard was used as a positive control. Analysis was performed with a CytoFLEX instrument (Beckman Coulter, USA).
Electron microscopy
To visualize EVs, the EV-containing pellets were resuspended in 1–2 mL of ice-cold PBS and pipetted onto formvar carbon-coated copper single slot grids (Agar Scientific, Stansted, UK) and allowed to adsorb for 20 min. EVs were further fixed with cold 2.5% glutaraldehyde, stained with uranyl acetate, and observed under a Zeiss Libra 120 TEM microscope.
For immuno-electron microscopy (immunogold labeling) the EV-containing pellets were fixed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer, 2% sucrose, for 1.5 h at 4°C, postfixed in 1% osmium tetroxide in 0.1 M cacodolate buffer (pH 7.4) with 5% sucrose for 1 h at 4°C, and dehydrated in a graded ethanol series, and embedded in a mixture of Araldite and Epon. Ultrathin sections were cut using an LKB-Ultratome with a diamond knife (Diatome, Ltd., Biel, Switzerland). Sections were mounted on nickel grids and were treated with hydrogen peroxide for 20 min to loosen the resin. Then the sections were incubated with primary antibody overnight, at 4°C in a moist chamber. Subsequently, the sections were exposed to the secondary antibody for 1 h at room temperature. Finally, the sections were double-stained with uranyl-acetate and lead citrate, and examined by transmission electron microscopy (Zeiss Libra 120, Carl Zeiss, Germany). Control sections were prepared by omitting the primary antibody. The following primary antibodies were used: monoclonal anti-CD63 (diluted 1:200, ab68418, Abcam, UK) and home-made monoclonal anti-Hsp70 (Clone 8D1). The secondary antibody used was gold conjugated (10 nm) anti-mouse IgG (diluted 1:10, Sigma, USA).
Western blotting
Samples after NeutAvidin-agarose and ATP-agarose chromatography, Hsp70 in EVs and soluble Hsp70 (solHsp70) fractions and B16 cells were analyzed with Western blotting. Cells were lysed on ice in RIPA buffer (20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 0.1% Triton X-100 and 1 mM PMSF). Equal amounts of total protein (20 µg/lane) were electrophoresed with a 10% sodium dodecyl sulfate (SDS) polyacrylamide gel. Proteins were transferred to a polyvinylidene difluoride (PVDF) membrane, and non-specific binding on the membrane was blocked with 5% fat-free milk in phosphate-buffered saline (PBS). Membranes were then incubated with anti-Hsp70 monoclonal antibody (Clone 3C5). Glyceral 3-phosphate-dehydrogenase (GAPDH) was used as loading control (ab8245, Abcam, UK). To estimate the level of Hsp70 in EVs, we calculated concentration of EVs using Nanoparticle Tracking analysis EVs and used 8x109 EVs for B16 cells and 2x109 EVs for CT-26 for western blotting. To analyze the content of arginase-1 in B16 tumors, tumor samples were lysed in RIPA lysing buffer and following electrophoresis and protein transfer to PVDF membranes the membranes were incubated with an antibody to arginase-1 (AF5868, R&D Systems, USA).
Hsp70-ELISA
Hsp70-ELISA was employed to assess the Hsp70 concentration during biological sample evaluation. Sample preparation: The samples for further analysis were prepared as follows: (i) B16 cells (untreated and incubated with rHsp70), as well as CT-26, and CT-26 cells incubated with rHsp70 for 6 h, were washed with cold PBS three times and fresh serum-free medium was added for the next 90 min. The EV fraction was isolated by a series of sequential centrifugations of the conditioned media as described above. To dissolve the EVs for Hsp70 measurement, 100 µl of 0.2 M HCl (pH 3.0) were added to the dry EV pellet following the final centrifugation and the pH of the mixture was immediately altered to pH 7.5 with Tris-HCl (pH 10.0); (ii) The supernatant after the last centrifugation, containing Hsp70 in soluble form; (iii) B16 cells, untreated and incubated with rHsp70, as well as CT-26 and CT-26 cells incubated with rHsp70 (3.5x106 cells/sample) were lysed in RIPA sample buffer, three times freeze-thawed, ultrasonicated and centrifuged (5000 g).
To evaluate Hsp70 content in samples 100 µL of affinity-purified homemade polyclonal anti-Hsp70 antibody (1.5 mg/mL) (see Supplementary Figure S2) was placed into wells of 96-well plate (Corning, USA) in 20 mM Borate buffer (pH 8.0), supplemented with 0.15 M NaCl, and incubated for 20 h in a humid chamber. After washing with Buffer A (20 mM Borate buffer (pH 8.0), 0.15 М NaCl, 0.05% Tween-20) the wells were loaded with 100 µL of the appropriate samples as indicated above, and with the rHsp70 titers used for calibration; all probes were mixed with 100 µL of Buffer A. Plates were further incubated for 1 h at 37oC on a shaker and after washing in Buffer A, biotinylated anti-Hsp70 polyclonal antibody (0.01 µg/mL) was added with an incubation of 1 h at 37oC on a shaker. After washing StreptAvidin-Peroxidase was added for an additional 1 h. The staining reaction was performed with tetramethylbenzene in citrate buffer (pH 4.5) and the intensity of staining was measured with a Varioscan (Thermo Fisher, USA). All results were obtained in triplicate.
Cytotoxicity assay
The cytotoxicity assay was performed with the aid of xCELLigence Real-Time Cell Analyzer DC equipment (Hoffmann-La Roche, Switzerland). This impedance-based assay carries out label-free, real-time high-throughput analysis of cell growth and lymphocyte-mediated cytotoxicity [22]. Since NK cells and other lymphocytes are non-adhesive, they do not possess impedance [23], therefore a change in electrical signal relates only to adherent tumor cells. To evaluate the sensitivity of B16 cells (incubated with Hsp70) to cytotoxic lymphocytes, intact B16 or CT-26 cells were seeded in the wells of an E-plate at concentrations of 104 cells/well and on the next morning rHsp70 (50 µg/mL), or the EVs fraction (concentration of 2.5x1012 EVs/mL), or soluble Hsp70 extruded from B16 cells (60 ng/mL), were added to intact B16 cells. 6 h later, effector cells isolated from the spleen of C3HA mice were added to the same wells at a ratio of 100:1 (effector:target) and recording was carried out over the next 20 h.
In experiments evaluating the cytotoxic effect of lymphocytes derived from tumor-bearing animals, we washed out cells from the spleens of mice inoculated with B16 cells, with or without EVs and then divided them into two parts: (i) total lymphocyte fraction, (ii) lymphocytes that were incubated with DynabeadsFlowComp™ Mouse CD8 (#1146D, Invitrogen, USA) in order to collect CD8 + cells. Cells were washed out from the beads according to manufacturer’s instructions and used as effector cells in the ratio (100:1).
Animal experiments
All in vivo experiments were carried out following the requirements of the Institute of Cytology of Russian Academy of Sciences Ethic Committee (Identification number F18-00380).
Female C57BL/6 mice (for subcutaneous B16 melanoma tumor formation), male BALB/c and female BALB/c nude mice (for subcutaneous CT26 tumor formation) were used in this study. C57BL/6 and BALB/c mice were purchased from Biomedical Technology Research Center (Russia), Balb/c nude – from National Research Nizhny Novgorod State University by N.I. Lobachevsky (Russia). The C57BL/6 mice were subcutaneously injected with 106 B16 cells, Balb/c - with 2x105 CT26iRFP720 cells/mouse and /Balb/c nude with 5x105 cells. The cells were mixed with EVs, harvested from either untreated cells (EVs-CNTR) or from cells previously incubated with 50 µg/mL rHsp70 (EVs-Hsp70). The number of EVs in the inoculum of the tumor cells was equal to 1x1011/mouse.
Survival curves were established according to the method of Kaplan and Meier and compared using a generalized Wilcoxon’s test. Tumor growth rate was estimated by weighing tumors taken from control and treated animals on day 19 after engrafting; sera and spleens were collected on the same day. Tumors were photographed, weighed and fixed in 4% formalin. Sera were frozen at -80oC before analysis while spleens were used immediately. CT-26 iRFP720-injected mice were subjected to bioimaging with the use of the IVIS Spectrum imaging system (Perkin-Elmer, UK) on day 19.
To estimate the specific cytotoxic activity, the splenocytes of mice from all experimental groups were used. For the precise analysis of the total or specific CD8 + cell response, we firstly isolated spleen cells from animals belonging to appropriate treatment groups and further divided each into two groups. One group was incubated with Dynabeads FlowComp™ Mouse CD8 (Invitrogen, USA) to isolate CD8 + cells and the other comprised the total lymphocyte fraction. These cell populations were used as effector cells, which were added to B16 or CT-26 cells at a ratio of 100:1. Cell viability was analyzed using the xCELLigence equipment as described above.
Immunohistochemistry
To reveal the possible infiltration of B16 tumors with pro-tumor M2 macrophages we prepared serial 7 µm frozen section from tumors isolated from B16-bearing mice and probed the slices with an antibody to arginase-1 (AF5868, R&DSystems, USA) followed by anti-sheep antibody (ab6747, Abcam, UK) and then anti-rabbit antibody conjugated with Alexa488 (Invitrogen, USA). Immunofluorescence images were captured with a Leica TCS SP5 confocal system (Leica Microsystems, Heidelberg, Germany).
Cytokine measurement in mouse sera
IL-10, IFN-γ and TNF-α measurements were performed with the aid of the following kits: Mouse IL-10 ELISA (#88-7105-88, Invitrogen, USA), Mouse IFN-γ ELISA (#88-7314-88, Invitrogen, USA), and Mouse TNF-α ELISA (#88-7324-88,Invitrogen, USA) according to the manufacturer’s instructions. All the probes were triplicated.
Nanoparticle Tracking Analysis (NTA)
The size of EVs and their concentration were determined using the NTA NanoSight® LM10 (Malvern Instruments) analyzer, equipped with a blue laser (45 mW at 488 nm) and a C11440-5B camera (Hamamatsu Photonics K.K., Japan). Recording and data analysis were performed using the NTA software 2.3. To optimize the measurement mode, the samples of isolated vesicles were diluted 1:100, 1:1000, or 1:10,000 by PBS. In the selected dilution, each sample was measured in triplicate. The following parameters were evaluated during the analysis of recordings monitored for 60 s: the average hydrodynamic diameter, the mode of distribution, the standard deviation, and the concentration of vesicles in the suspension.
Statistical analysis
All numerical results are reported as the mean ± standard error of the mean (SEM) and represent data from a minimum of three independent experiments. Quantitative analysis was performed with the use of Graph Pad Prism 8.0 (Graph Pad Software Inc). The one-way ANOVA test followed by Dunnett’s multiple comparison test (which compares the mean of each condition with control) was used. Differences were considered statistically significant at p < 0.05.