Dendritic Cell-Derived Artificial Microvesicles Inhibit RLS40 Lymphosarcoma Growth in Mice via Stimulation of Th1/Th17 Immune Response

Cell-free antitumor vaccines represent a promising approach to immunotherapy of cancer. Here, we compare the antitumor potential of cell-free vaccines based on microvesicles derived from dendritic cells (DCs) with DC- and cationic-liposome-based vaccines using a murine model of drug-resistant lymphosarcoma RLS40 in vivo. The vaccines were the following: microvesicle vaccines—cytochalasin B-induced membrane vesicles (CIMVs) obtained from DCs loaded with total tumor RNA using cholesterol/spermine-containing cationic liposomes L or mannosylated liposomes ML; DC vaccines—murine DCs loaded with total tumor-derived RNA using the same liposomes; and liposomal vaccines—lipoplexes of total tumor-derived RNA with liposomes L or ML. Being non-hepatotoxic, CIMV- and DC-based vaccines administered subcutaneously exhibited comparable potential to stimulate highly efficient antitumor CTLs in vivo, whereas liposomal vaccines were 25% weaker CTL inducers. Nevertheless, the antitumor efficiencies of the different types of the vaccines were similar: sizes of tumor nodes and the number of liver metastases were significantly decreased, regardless of the vaccine type. Notably, the booster vaccination did not improve the overall antitumor efficacy of the vaccines under the study. CIMV- and DC- based vaccines more efficiently than liposome-based ones decreased mitotic activity of tumor cells and induced their apoptosis, stimulated accumulation of neutrophil inflammatory infiltration in tumor tissue, and had a more pronounced immunomodulatory activity toward the spleen and thymus. Administration of CIMV-, DC-, and liposome-based vaccines resulted in activation of Th1/Th17 cells as well as the induction of positive immune checkpoint 4-1BBL and downregulation of suppressive immune checkpoints in a raw PD-1 >>> TIGIT > CTLA4 > TIM3. We demonstrated that cell-free CIMV-based vaccines exhibited superior antitumor and antimetastatic activity in a tumor model in vivo. The obtained results can be considered as the basis for developing novel strategies for oncoimmunotherapy.


Introduction
DCs are professional antigen-presenting cells that trigger antitumor immune responses by the presentation of tumor antigens to T cells accompanied by their stimulation by membrane (co-stimulatory) and soluble (cytokines) molecules. The application of DCs loaded ex vivo with tumor antigens for the activation of antitumor immunity has been intensively investigated in experimental animals and clinical trials for more than 30 years. It has been demonstrated that antitumor DC-based vaccines are highly efficient in various murine tumor models and human tumor xenografts in immunodeficient mice, being administered as monotherapy [1,2]

Liposome-Based Vaccines
Liposome-based vaccines are lipoplexes spontaneously formed in Opti-MEM medium after total tumor-derived RNA is mixed with L or ML liposome solution at N/P ratio 4/1 (for details see [33]). Liposome-based vaccines were administered s/c in the interscapular region at a dose of 5 µg RNA in the form of lipoplexes per mouse in 0.2 mL of Opti-MEM.

Cytochalasin B-Induced Membrane Vesicle (CIMV)-Based Vaccines
CIMVs were prepared as previously described [37] with some modifications. Briefly, immature bone marrow-derived DCs or tsDCs were transfected with RNA-B16 or RNA-RLS 40 precomplexed with cationic liposomes L or ML at N/P ratio 4/1 as described above. Transfected DCs were washed with serum-free media, resuspended in serum-free IMDM supplemented with 10 µg/mL cytochalasin B (AppliChem GmbH, Darmstadt, Germany) and incubated for 30 min under SC. After that, cells were vigorously vortexed for 30 s. CIMVs were collected by sequential centrifugation at 100× g (10 min, 4 • C), 600 g (20 min, 4 • C), and 15,000× g (30 min, 4 • C). The resulting pellet was washed with PBS (MP Biomedicals, Santa Ana, CA, USA) and centrifuged at 15,000× g (30 min, 4 • C). CIMVs were resuspended in 0.1 mL of PBS and stored at −80 • C until use. CIMV concentration was measured on the basis of the total protein concentration with Qubit Protein Assay kit (Thermo Fisher Scientific, Waltham, MA, USA ) [19]. CIMV-based vaccines were administered s/c at the crest area in the interscapular region at a dose of 15 µg of CIMVs (total protein counts) per mouse in 0.2 mL of Opti-MEM.

Transmission Electron Microscopy of CIMVs
Transmission electron microscopy (TEM) of CIMVs was performed at the Center for Microscopy of Biological Subjects (Institute of Cytology and Genetics SB RAS, Novosibirsk, Russia) using a JEM1400 microscope (80 kV, JEOL, Tokyo, Japan). Negative contrast staining of CIMVs with uranyl acetate and subsequent procedures were performed as previously described [19].

Flow Cytometry Analysis of CIMVs
The CIMV surface proteins were analyzed after the absorption of CIMVs on aldehyde/sulfate latex beads (Invitrogen, USA) followed by blocking with 0.5 M glycine as described previously [19,38]. Latex beads with immobilized CIMVs were stained with monoclonal antibodies for 1 h at 4 • C. Subsequently, the latex beads were washed twice with 300 µL PBS and analyzed by flow cytometry. Three DC-specific markers involved in the formation of immunological synapse and activation of antitumor T cells were checked on the surface of DC-derived CIMVs, namely, MHC II (FITC; Sony Biotechnology, Tokyo, Japan), CD80 (APC; Sony Biotechnology, Tokyo, Japan), and CD83 (PE; BD Biosciences, San Jose, CA, USA). Flow cytometry measurements were performed using a NovoCyte 3000 flow cytometer (ACEA Biosciences, San Diego, CA, USA) using a 640 nm Ex wavelength for antibodies conjugated to APC (Em 675 ± 30 nm) and 488 nm for antibodies conjugated to FITC or PE (Em 530 ± 30 nm or 572 ± 28 nm, respectively). A forward scatter height versus forward scatter area density plot (linear scales) was used to exclude latex bead doublets. At least 30,000 events were analyzed in each experiment. All experiments were run in triplicate for statistical analysis. Data were processed with FlowJo Software (Version 10.5.3, Tree Star Inc., Ashland, OR, USA).

Tumor Transplantation and Design of Animal Experiments
Female CBA/LacSto mice of 10 to 12 weeks of age were used. Solid tumors of RLS 40 lymphosarcoma were induced by intramuscular (i/m) injection of tumor cells (10 6 ) in 0.1 mL of saline buffer into the right thighs of mice. On day 7 after tumor transplantation, mice were randomized, assigned to one of the experimental groups, and s/c vaccinated with DC-, liposome-, or CIMV-based vaccines at the crest area in the interscapular region once (denoted as (×1), on day 7) or twice (denoted as (×2), on days 7 and 14). DC-based vaccines were 10 5 of DCs transfected with RNA-RLS 40  Starting on day 9 of the experiment, tumor size was determined every 3-4 days by caliper measurements in three perpendicular dimensions. Tumor volumes were calculated as V = (π/6 × length × width × height).
On day 21 of the experiment, the mice were sacrificed. Tumor weight was determined using the equation W tumor = W right leg − W left leg , where W right leg is the weight of the right leg with tumor, W left leg is the weight of the left leg without tumor. The animal experiment was performed twice.
Tumor tissues and organs (liver, spleen, thymus) were collected for histology, immunohistochemistry, and morphometry. Peripheral blood was collected from the retroorbital sinus for serum preparation and for blood biochemistry assessment. Total RNA was isolated from splenocytes with TRIzol Reagent according to manufacturer protocol. RNA was used to analyze the expression levels of immune checkpoints and master regulators of T cell differentiation.

Histology and Immunohistochemistry
Tissue specimens (tumor nodes, liver, spleen, and thymus) were fixed in 10% neutral buffered formalin (BioVitrum, Moscow, Russia), dehydrated in ascending ethanols and xylols, and embedded in HISTOMIX paraffin (BioVitrum, Russia). The paraffin sections (5 µm) were sliced on a Microm HM 355S microtome (Thermo Fisher Scientific, Waltham, MA, USA) and stained with hematoxylin and eosin. For the immunohistochemical study, the tumor sections were deparaffinized and rehydrated. Antigen retrieval was carried out after exposure in a microwave oven at 700 W. The samples were incubated with anti-caspase-3 (ab2302, Abcam, Cambridge, UK) primary antibodies according to the manufacturer's protocol. Then, the sections were incubated with secondary horseradish peroxidase (HPR)conjugated antibodies (Spring Bioscience detection system, Spring Bioscience, Pleasanton, CA USA), exposed to the 3,3 -diaminobenzidine (DAB) substrate, and stained with Mayer's hematoxylin. All images were examined and scanned using an Axiostar Plus microscope equipped with an Axiocam MRc5 digital camera (Zeiss, Munich, Germany) at magnifications of ×100 and ×400. Five to ten random fields were studied in each specimen depending on specimen size, forming 25-50 random fields for each experimental group of mice in total. In the control group containing 10 mice, 50-100 random fields were analyzed.
The percentages of the internal metastases areas in the liver were determined relative to the total area of liver sections using Adobe Photoshop. Inhibition of metastases development was assessed using the metastasis inhibition index (MII), calculated as MII = ((mean metastasis area control − mean metastasis area experiment )/mean metastasis area control ) × 100%. MII in the tumor-bearing mice without treatment (control) was taken as 0%, and MII, reflecting the absence of metastases, was taken as 100%.
Morphometric analysis of tumor, spleen, thymus, and liver sections was performed by point counting, using a morphometric grid with 100 testing points in a testing area equal to 3.2 × 10 6 µm 2 . Morphometric analysis of tumor tissue included evaluation of the volume densities (Vv, %) of unchanged tumor tissue, inflammatory infiltration, necrosis, and numerical densities (Nv) of mitoses and caspase-3-positive cells. Morphometric analysis of the spleen included evaluation of the volume densities (Vv, %) of red pulp, white pulp, and the diameter of lymphoid follicles (µm). Morphometric analysis of the thymus included evaluation of the volume densities (Vv, %) of the cortex and medulla with subsequent calculation of cortex/medulla index. Morphometric analysis of the liver included evaluation of the volume densities (Vv, %) of normal liver tissue, dystrophy, necrosis, and numerical density (Nv) of binuclear hepatocytes.
The volume density (Vv, %) of the histological structure under study indicates the volume fraction of tissue occupied by this compartment, is determined from the testing points lying over this structure, and is calculated using the following formula: Vv = (P structure /P test ) × 100%, where P structure denotes the number of points over the structure and P test denotes the total number of test points, 100 in this case. The numerical density (Nv) of the histological structure under study indicates the number of parti- cles in the unit of tissue volume evaluated as the number of particles in the square unit, 3.2 × 10 6 µm 2 in this case.

Immune Checkpoints and Master Regulators Analysis
To analyze the type of activated T-helper immune response and the expression levels of negative and positive immune checkpoints after treatment of RLS 40 -bearing mice with different vaccines, a detection system was developed on the basis of qPCR and fluorescent probes. To design primers and probes, we utilized the RealTime PCR Tool (https:// eu.idtdna.com/scitools/Applications/RealTimePCR/, accessed on 15 July 2019). The secondary structure of the primers/probes and the self-and heterodimerization tendencies of each primer/probe set were predicted using the OligoAnalyzer Tool (https://eu.idtdna. com/calc/analyzer, accessed on 15 July 2019). Primers and probes were synthesized in the Laboratory of Biomedical Chemistry of ICBFM SB RAS (Novosibirsk, Russia).
Total RNA was isolated from the splenocytes of experimental animals using TRIzol Reagent in accordance with manufacturer recommendations. Coding DNA (cDNA) was synthesized from total RNA in 20 µL of reaction mixture containing 5 µg total RNA, 4 µL 5× RT buffer (RevertAid, Termo Fisher, Waltham, MA, USA), 200 U RevertAid (TermoFisher, USA), 1 mM dNTP mixture, and 100 µM random hexaprimer. Reverse transcription was performed at 25 • C for 10 min followed by incubation at 42 • C for 60 min. Finally, the reverse transcriptase reaction was terminated at 70 • C for 10 min.

Statistical Analysis
The data were statistically processed using one-way ANOVA with the Tukey or Fisher post hoc tests; a p value of ≤0.05 was considered to indicate a significant difference.

Preparation and Characterization of Vaccines
In this study, three types of vaccines were investigated. Cellular DC-based vaccines were prepared from mouse bone marrow-derived DCs or tsDCs transfected with tumor cell-derived RNA pre-complexed with either liposomes L or ML as previously described [33,39]. Lipoplexes were pre-formed at a nitrogen-to-phosphate (N/P) ratio of 4/1 because, at this ratio, the optimal combination of transfection efficiency and DC targeting was observed [33,39]. Liposome-based vaccines were lipoplexes of the same liposomes (L or ML) with tumor RNA prepared at N/P ratio 4/1. Vesicular vaccines were cytochalasin B-induced membrane vesicles (CIMVs) prepared from DCs loaded with total tumor-derived RNA pre-complexed with liposomes L or ML. Altogether, six different vaccines were used, and each vaccine was administered once (×1) or twice (×2) ( Table 2). at this ratio, the optimal combination of transfection efficiency and DC targeting was observed [33,39]. Liposome-based vaccines were lipoplexes of the same liposomes (L or ML) with tumor RNA prepared at N/P ratio 4/1. Vesicular vaccines were cytochalasin B-induced membrane vesicles (CIMVs) prepared from DCs loaded with total tumor-derived RNA pre-complexed with liposomes L or ML. Altogether, six different vaccines were used, and each vaccine was administered once (×1) or twice (×2) ( Table 2).  1 The same designation was used for the experimental groups. 2 Color code was used throughout the article to show data corresponded to a particular group. 3 Day(s) from RLS40 lymphosarcoma transplantation.
The mean yield of DC-derived CIMVs (by total protein count) was 100-150 µg per 10 7 tsDCs, more than 10-fold higher compared with the yields of natural extracellular vesicles of DCs isolated from the conditioned medium [37].
Transmission electron microscopy revealed that DC-derived CIMVs had the same characteristic spherical structures, most of which had size ≤ 150 nm ( Figure 1A), as our previous data obtained using dynamic light scattering, nanoparticle tracking analysis, and at this ratio, the optimal combination of transfection efficiency and DC targeting was observed [33,39]. Liposome-based vaccines were lipoplexes of the same liposomes (L or ML) with tumor RNA prepared at N/P ratio 4/1. Vesicular vaccines were cytochalasin B-induced membrane vesicles (CIMVs) prepared from DCs loaded with total tumor-derived RNA pre-complexed with liposomes L or ML. Altogether, six different vaccines were used, and each vaccine was administered once (×1) or twice (×2) ( Table 2).  1 The same designation was used for the experimental groups. 2 Color code was used throughout the article to show data corresponded to a particular group. 3 Day(s) from RLS40 lymphosarcoma transplantation.
The mean yield of DC-derived CIMVs (by total protein count) was 100-150 µg per 10 7 tsDCs, more than 10-fold higher compared with the yields of natural extracellular vesicles of DCs isolated from the conditioned medium [37].
Transmission electron microscopy revealed that DC-derived CIMVs had the same characteristic spherical structures, most of which had size ≤ 150 nm ( Figure 1A), as our previous data obtained using dynamic light scattering, nanoparticle tracking analysis, and d7, d14 DC/ML(×1) served [33,39]. Liposome-based vaccines were lipoplexes of the same liposomes (L or ML) with tumor RNA prepared at N/P ratio 4/1. Vesicular vaccines were cytochalasin B-induced membrane vesicles (CIMVs) prepared from DCs loaded with total tumor-derived RNA pre-complexed with liposomes L or ML. Altogether, six different vaccines were used, and each vaccine was administered once (×1) or twice (×2) ( Table 2).  1 The same designation was used for the experimental groups. 2 Color code was used throughout the article to show data corresponded to a particular group. 3 Day(s) from RLS40 lymphosarcoma transplantation.
The mean yield of DC-derived CIMVs (by total protein count) was 100-150 µg per 10 7 tsDCs, more than 10-fold higher compared with the yields of natural extracellular vesicles of DCs isolated from the conditioned medium [37].
Transmission electron microscopy revealed that DC-derived CIMVs had the same characteristic spherical structures, most of which had size ≤ 150 nm ( Figure 1A), as our previous data obtained using dynamic light scattering, nanoparticle tracking analysis, and served [33,39]. Liposome-based vaccines were lipoplexes of the same liposomes (L or ML) with tumor RNA prepared at N/P ratio 4/1. Vesicular vaccines were cytochalasin B-induced membrane vesicles (CIMVs) prepared from DCs loaded with total tumor-derived RNA pre-complexed with liposomes L or ML. Altogether, six different vaccines were used, and each vaccine was administered once (×1) or twice (×2) ( Table 2).  1 The same designation was used for the experimental groups. 2 Color code was used throughout the article to show data corresponded to a particular group. 3 Day(s) from RLS40 lymphosarcoma transplantation.
The mean yield of DC-derived CIMVs (by total protein count) was 100-150 µg per 10 7 tsDCs, more than 10-fold higher compared with the yields of natural extracellular vesicles of DCs isolated from the conditioned medium [37].
Transmission electron microscopy revealed that DC-derived CIMVs had the same characteristic spherical structures, most of which had size ≤ 150 nm ( Figure 1A), as our previous data obtained using dynamic light scattering, nanoparticle tracking analysis, and d7, d14 Liposomal vaccines duced membrane vesicles (CIMVs) prepared from DCs loaded with total tumor-derived RNA pre-complexed with liposomes L or ML. Altogether, six different vaccines were used, and each vaccine was administered once (×1) or twice (×2) ( Table 2).  1 The same designation was used for the experimental groups. 2 Color code was used throughout the article to show data corresponded to a particular group. 3 Day(s) from RLS40 lymphosarcoma transplantation.
The mean yield of DC-derived CIMVs (by total protein count) was 100-150 µg per 10 7 tsDCs, more than 10-fold higher compared with the yields of natural extracellular vesicles of DCs isolated from the conditioned medium [37].
Transmission electron microscopy revealed that DC-derived CIMVs had the same characteristic spherical structures, most of which had size ≤ 150 nm ( Figure 1A), as our previous data obtained using dynamic light scattering, nanoparticle tracking analysis, and RNA-RLS 40 /L lipoplexes d7 5 µg RNA L(×2) duced membrane vesicles (CIMVs) prepared from DCs loaded with total tumor-derived RNA pre-complexed with liposomes L or ML. Altogether, six different vaccines were used, and each vaccine was administered once (×1) or twice (×2) ( Table 2).
The mean yield of DC-derived CIMVs (by total protein count) was 100-150 µg per 10 7 tsDCs, more than 10-fold higher compared with the yields of natural extracellular vesicles of DCs isolated from the conditioned medium [37].
Transmission electron microscopy revealed that DC-derived CIMVs had the same characteristic spherical structures, most of which had size ≤ 150 nm ( Figure 1A), as our previous data obtained using dynamic light scattering, nanoparticle tracking analysis, and d7, d14 RNA pre-complexed with liposomes L or ML. Altogether, six different vaccines were used, and each vaccine was administered once (×1) or twice (×2) ( Table 2).
The same designation was used for the experimental groups. 2 Color code was used throughout the article to show data corresponded to a particular group. 3 Day(s) from RLS40 lymphosarcoma transplantation.
The mean yield of DC-derived CIMVs (by total protein count) was 100-150 µg per 10 7 tsDCs, more than 10-fold higher compared with the yields of natural extracellular vesicles of DCs isolated from the conditioned medium [37].
Transmission electron microscopy revealed that DC-derived CIMVs had the same characteristic spherical structures, most of which had size ≤ 150 nm ( Figure 1A), as our previous data obtained using dynamic light scattering, nanoparticle tracking analysis, and RNA-RLS 40 /ML lipoplexes d7 used, and each vaccine was administered once (×1) or twice (×2) ( Table 2).
The same designation was used for the experimental groups. 2 Color code was used throughout the article to show data corresponded to a particular group. 3 Day(s) from RLS40 lymphosarcoma transplantation.
The mean yield of DC-derived CIMVs (by total protein count) was 100-150 µg per 10 7 tsDCs, more than 10-fold higher compared with the yields of natural extracellular vesicles of DCs isolated from the conditioned medium [37].
The mean yield of DC-derived CIMVs (by total protein count) was 100-150 µg per 10 7 tsDCs, more than 10-fold higher compared with the yields of natural extracellular vesicles of DCs isolated from the conditioned medium [37].
The same designation was used for the experimental groups. 2 Color code was used throughout the article to show data corresponded to a particular group. 3 Day(s) from RLS40 lymphosarcoma transplantation.
The mean yield of DC-derived CIMVs (by total protein count) was 100-150 µg per 10 7 tsDCs, more than 10-fold higher compared with the yields of natural extracellular vesicles of DCs isolated from the conditioned medium [37]. 1 The same designation was used for the experimental groups. 2 Color code was used throughout the article to show data corresponded to a particular group. 3 Day(s) from RLS40 lymphosarcoma transplantation.
The mean yield of DC-derived CIMVs (by total protein count) was 100-150 µg per 10 7 tsDCs, more than 10-fold higher compared with the yields of natural extracellular vesicles of DCs isolated from the conditioned medium [37]. 1 The same designation was used for the experimental groups. 2 Color code was used throughout the article to show data corresponded to a particular group. 3 Day(s) from RLS40 lymphosarcoma transplantation.
The mean yield of DC-derived CIMVs (by total protein count) was 100-150 µg per 10 7 tsDCs, more than 10-fold higher compared with the yields of natural extracellular vesicles of DCs isolated from the conditioned medium [37].
d7, d14 1 The same designation was used for the experimental groups. 2 Color code was used throughout the article to show data corresponded to a particular group. 3 Day(s) from RLS 40 lymphosarcoma transplantation.
The mean yield of DC-derived CIMVs (by total protein count) was 100-150 µg per 10 7 tsDCs, more than 10-fold higher compared with the yields of natural extracellular vesicles of DCs isolated from the conditioned medium [37].
Transmission electron microscopy revealed that DC-derived CIMVs had the same characteristic spherical structures, most of which had size ≤ 150 nm ( Figure 1A), as our previous data obtained using dynamic light scattering, nanoparticle tracking analysis, and TEM techniques [37]. As previously shown, lipoplexes formed by tumor RNA and liposomes L or ML at N/P = 4/1 were rather small (Ø 80-150 nm) and characterized by a polydispersity index (PDI) of 0.2, showing the formation of uniform particles [33]. DC-derived CIMVs were also characterized by small diameters (50-150 nm) and low PDI of 0.47, which indicated the formation of small and rather uniform particles similar to the lipoplexes [37]. Additionally, DC-derived CIMVs were demonstrated to express on their surface the DC-specific proteins MHC II, CD80, and CD83, essential for the formation of immunological synapses with T-lymphocytes and subsequent activation of the antitumor immune response ( Figure 1B).
The main challenge in the application of antitumor vaccines of different origins, namely, cellular, liposome-, and vesicle-based ones, in one experiment is the correlation of the dosages between different vaccine types. We adjusted the doses of vaccines based on the number of DCs or amount of tumor RNA used for vaccine preparation per mouse. DC-based vaccines were injected at a dose of 10 5 cells (1 µg tumor RNA) per mouse, in accordance with our previous data [32,33,39] and other studies [40,41]. This dose of DCs was sufficient to activate the antitumor immune response and inhibit tumor growth and metastasis in tumor-bearing mice [32,33,[39][40][41]. Liposomal vaccines containing mRNA are commonly administered to mice at a dose of 10-30 µg RNA per mouse [42][43][44][45]. Here, we used a lower dose of liposomal vaccine 5 µg RNA per mouse because it was previously shown that this dosage efficiently activates antitumor CTLs in vivo in murine melanoma B16 model [36]. Note that the amount of RNA per single dose in liposomal and DC-based vaccines differs five-fold; it is supposed that this is the minimal amount needed to overcome the loss of lipoplexes due to non-targeted delivery in vivo after s/c administration. used a lower dose of liposomal vaccine 5 µg RNA per mouse because it was previously shown that this dosage efficiently activates antitumor CTLs in vivo in murine melanoma B16 model [36]. Note that the amount of RNA per single dose in liposomal and DC-based vaccines differs five-fold; it is supposed that this is the minimal amount needed to overcome the loss of lipoplexes due to non-targeted delivery in vivo after s/c administration. We consider CIMVs as an alternative to natural extracellular vesicles (exosomes). Hence, the dosage of DC-derived CIMVs was calculated while considering both the amount of exosomes used worldwide for antitumor vaccination and the number of DCs used to prepare CIMVs to keep the dosage of DC-and CIMVs-based vaccines as close as possible. DC-derived extracellular vesicles are commonly administered to experimental animals at a dose of 40-50 µg (by total protein content) per mouse, although some researchers have used higher (100 µg per mouse [46]) or significantly lower (1-3 µg per mouse [47]) dosages. In our work, a dose of 15 µg of CIMVs (by total protein content) per mouse was used that met both criteria listed above. On the one hand, the dose of CIMVs is comparable to that of exosomes; on the other, a single dose of CIMVs was prepared from 10 6 of DCs that was only 10-fold higher than the dose of the DC-based vaccine where We consider CIMVs as an alternative to natural extracellular vesicles (exosomes). Hence, the dosage of DC-derived CIMVs was calculated while considering both the amount of exosomes used worldwide for antitumor vaccination and the number of DCs used to prepare CIMVs to keep the dosage of DC-and CIMVs-based vaccines as close as possible. DC-derived extracellular vesicles are commonly administered to experimental animals at a dose of 40-50 µg (by total protein content) per mouse, although some researchers have used higher (100 µg per mouse [46]) or significantly lower (1-3 µg per mouse [47]) dosages. In our work, a dose of 15 µg of CIMVs (by total protein content) per mouse was used that met both criteria listed above. On the one hand, the dose of CIMVs is comparable to that of exosomes; on the other, a single dose of CIMVs was prepared from 10 6 of DCs that was only 10-fold higher than the dose of the DC-based vaccine where the excess of CIMVs was required to compensate for the loss of CIMVs due to non-productive interactions with non-immunological cells and tissues upon s/c administration. This concentration of CIMVs was previously used in animal experiments by Gomzikova et al. [24].

DC-Derived CIMVs Efficiently Primed Antitumor Cytotoxic T-Cells In Vivo
The potential of DC-, liposome-, and CIMV-based vaccines to activate antitumor CTLs in vivo was assessed. Because RLS 40 lymphosarcoma is a tumor that is maintained mainly in ascites and poorly handled in culture, we decided to perform experiments in a melanoma B16 model. For this purpose, splenocytes were isolated from healthy C57Bl/6 mice 7 days after s/c vaccination with the respective vaccine and restimulated for 6 days ex vivo with the lysate of B16 cells. Restimulated splenocytes (effector cells) were co-cultured with B16 cells (target cells) at effector-to-target cell ratio 20/1, and cytotoxic activity of splenocytes was monitored in real-time mode using the xCELLigence system (ACEA Biosciences, USA) ( Figure 2). It was shown that all tested vaccines stimulated antitumor CTLs that efficiently lysed tumor cells, showing up to a 30-fold decrease in the cell index of melanoma B16 cells compared with the control (Figure 2). Among tested vaccines, the liposome-based one was a slightly less efficient activator of antitumor CTLs compared with DC-and CIMVbased vaccines. Thus, DC-, liposome-, and CIMV-based vaccines under the study were demonstrated to efficiently activate antitumor CTLs in vivo.
the excess of CIMVs was required to compensate for the loss of CIMVs due to non-productive interactions with non-immunological cells and tissues upon s/c administration. This concentration of CIMVs was previously used in animal experiments by Gomzikova et al. [24].

DC-Derived CIMVs Efficiently Primed Antitumor Cytotoxic T-Cells In Vivo
The potential of DC-, liposome-, and CIMV-based vaccines to activate antitumor CTLs in vivo was assessed. Because RLS40 lymphosarcoma is a tumor that is maintained mainly in ascites and poorly handled in culture, we decided to perform experiments in a melanoma B16 model. For this purpose, splenocytes were isolated from healthy C57Bl/6 mice 7 days after s/c vaccination with the respective vaccine and restimulated for 6 days ex vivo with the lysate of B16 cells. Restimulated splenocytes (effector cells) were co-cultured with B16 cells (target cells) at effector-to-target cell ratio 20/1, and cytotoxic activity of splenocytes was monitored in real-time mode using the xCELLigence system (ACEA Biosciences, USA) ( Figure 2). It was shown that all tested vaccines stimulated antitumor CTLs that efficiently lysed tumor cells, showing up to a 30-fold decrease in the cell index of melanoma B16 cells compared with the control (Figure 2). Among tested vaccines, the liposome-based one was a slightly less efficient activator of antitumor CTLs compared with DC-and CIMV-based vaccines. Thus, DC-, liposome-, and CIMV-based vaccines under the study were demonstrated to efficiently activate antitumor CTLs in vivo. Control-B16 cells incubated in the absence of splenocytes (gray line). Healthy-B16 cells incubated with splenocytes from healthy mice (blue line). ML, CIMV/ML, and DC/ML corresponded to B16 cells incubated with splenocytes isolated from mice vaccinated with ML, CIMV/ML, or DC/ML vaccines, respectively (yellow, magenta, and indigo lines, respectively). The viability of B16 cells was assessed in real-time mode using xCELLigence (ACEA Biosciences, San Diego, CA, USA). Data presented as mean ± SD. Data were statistically analyzed using one-way ANOVA with post hoc Tukey test; p-value indicates a statistically reliable difference:** p < 0.01. Control-B16 cells incubated in the absence of splenocytes (gray line). Healthy-B16 cells incubated with splenocytes from healthy mice (blue line). ML, CIMV/ML, and DC/ML corresponded to B16 cells incubated with splenocytes isolated from mice vaccinated with ML, CIMV/ML, or DC/ML vaccines, respectively (yellow, magenta, and indigo lines, respectively). The viability of B16 cells was assessed in real-time mode using xCELLigence (ACEA Biosciences, San Diego, CA, USA). Data presented as mean ± SD. Data were statistically analyzed using one-way ANOVA with post hoc Tukey test; p-value indicates a statistically reliable difference: ** p < 0.01.

Effect of Vaccination on RLS 40 Lymphosarcoma Growth and Metastasis Development In Vivo
The experimental setup for investigating the antitumor potential of the designed vaccines is depicted in Figure 3A. RLS 40 cells (1 × 10 6 ) were implanted i/m into the right thigh of CBA mice. After that, on day 7, animals were randomized, assigned to one of the experimental groups (n = 10), and vaccinated s/c with DC-, liposome-, or CIMV-based vaccines. Five animals from each group under study (see Table 2 for details) received a second dose of vaccine on day 14 ( Figure 3A). Altogether, 12 experimental and one control (non-treated) groups were included in the experiment. Tumor size was monitored every 3-4 days, followed by calculation of tumor volumes ( Figure 3B). Mice were euthanized on day 21 of tumor growth, followed by the determination of tumor weight as well as collection of tumor nodes and organs for subsequent histological analysis ( Figure 3C). perimental groups (n = 10), and vaccinated s/c with DC-, liposome-, or CIMV-based vac-cines. Five animals from each group under study (see Table 2 for details) received a second dose of vaccine on day 14 ( Figure 3A). Altogether, 12 experimental and one control (nontreated) groups were included in the experiment. Tumor size was monitored every 3-4 days, followed by calculation of tumor volumes ( Figure 3B). Mice were euthanized on day 21 of tumor growth, followed by the determination of tumor weight as well as collection of tumor nodes and organs for subsequent histological analysis ( Figure 3C).  It is obvious from Figure 3B that all types of tested vaccines caused 4-5-fold inhibition of the growth of primary tumor nodes compared with the control group. However, differences in the tumor volume between the experimental groups were statistically insignificant. For example, under a single vaccination schedule ( Figure 3B, upper left panel) DC-based vaccine DC/ML and both non-cellular vaccines (lipoplexes RNA-RLS 40 /ML and CIMVs derived from loaded DCs; CIMV/ML) exhibited similar inhibition of tumor growth at the end of the experiment. Interestingly, some faint differences in the tumor suppression rate appeared between the vaccines on the 15th day after tumor transplantation: DC/ML looked slightly less efficient than ML and CIMV/ML. In the case of double vaccination with the same vaccines, CIMV/ML again seems somewhat more efficient, since no tumor growth was observed during the entire period of observation, while in the case of DC/ML and ML vaccines, some tumor growth was observed starting on day 15 of the experiment ( Figure 3B). Interestingly, preparation of antitumor vaccines using DC-targeted mannosylated liposomes ML provides almost no advantages over the application of nontargeted cationic liposomes L ( Figure 3B). We previously observed this fact, and it can be explained by the superior transfection efficiency of liposomes L, especially in vitro [32,33]. However, in vivo, single and double vaccination of mice with vaccine L led to vigorous inhibition of tumor growth, which was even more visible in the case of single vaccination ( Figure 3B, low left panel); in the case of DC/L and CIMV/L vaccines, visible growth of tumor was observed starting on day 12 of observation, while in the group L vaccine, the tumor growth was not observed. One of the possible explanations is that, being positively charged, L vaccine caused non-specific activation of the immune system that resulted in tumor growth inhibition. Similar effects of cationic liposomes on the immune system are well-documented [48,49].
Thus, two schedules of vaccination of tumor-bearing mice were used in the study, including (1) single vaccination on day 7 after the tumor transplantation and (2) double vaccination on day 7 followed by booster vaccination on day 14. Both regimens of vaccination were demonstrated to be efficient and resulted in significant retardation of RLS 40 growth compared with the control group ( Figure 3B). Some advantages of double vaccination were observed only for the groups vaccinated with DC/L vaccine, in which double vaccination was much more effective compared with a single dose. Similarly, double vaccination with CIMV/ML was also more effective than a single dose.
It was shown that tumor weight of vaccinated mice decreased 4-8-fold in all groups compared to the control ( Figure 3C). These data correlated well with the results of vaccineinduced retardation of tumor growth ( Figure 3B,C). Furthermore, statistically significant differences between the tumor weights in different groups were observed ( Figure 3C). The more pronounced effect of vaccination on tumor weight was observed in the groups double vaccinated with DC/L, CIMV/ML, and once vaccinated with CIMV/L. In these groups, no tumor growth, as well as minimal (as low as 0.1-0.2 g) tumor weight, were registered. In other groups, tumor weights decreased after vaccination but not significantly.
Thus, DC-, liposome-, and CIMV-based vaccines significantly retarded the growth of the primary tumor node. However, in the total application of the DC-targeted mannosylated liposomes, ML instead of L and administration of booster vaccination did not significantly improve the antitumor effect of the vaccines.
Given the very efficient inhibition of RLS 40 tumor growth in vaccinated groups of mice, we further analyzed the effect of vaccination on the development of internal metastasis in the liver: the target organ of lymphosarcoma RLS 40 . Histologically, liver metastases of the lymphosarcoma RLS 40 were represented predominantly by rounded foci with unclear boundaries consisting of the large monomorphic atypical lymphoid cells, which were comparable with the cells forming the primary tumor node. The histological structure of the primary tumor node of RLS 40 lymphosarcoma has been previously described in detail [50,51]. The areas occupied by metastases were determined relative to the total area of the liver section and expressed with the Metastases Inhibition Index (MII) (for details, see Section 2). As depicted in Figure S1 Figure S1). The anti-metastatic effect of DC/L and DC/ML vaccines ranged from 73.8% to 94% and tended to increase with twofold immunization (Supplementary Figure S1). After administration of L or ML lipoplexes loaded with RNA-RLS 40 , MII reached 93.1% for L and 99.2% for ML, thus showing that ML vaccine was somewhat more effective; however, no statistically significant differences between single and twofold immunizations were found (Supplementary Figure S1). After administration of CIMV obtained from L/RNA-RLS 40 -or ML/RNA-RLS 40 -loaded DCs, MII reached 87.6-99% and tended toward a slight decrease with double immunization (98-99% for single and 91.8-87.6% for double immunization) (Supplementary Figure S1). Thus, there were no statistically significant differences concerning the antimetastatic activity between the groups; however, cell-free vaccines (L, ML, CIMV/L, and CIMV/ML) were more effective than DC-based vaccines, and booster immunization was more effective in the case of DCs and less effective in the case of CIMV-based vaccines. The use of DC-targeted or non-targeted liposomes for delivery of tumor-derived RNA to DCs did not affect the antimetastatic activity of different types of vaccines.
A histological study of RLS 40 lymphosarcoma tissue revealed that vaccines caused changes in the histological characteristics of the primary tumor node manifested in the increase of foci of necrosis in the central part of the tumor node and inflammatory infiltration located on the border between the unchanged and necrotic tissue and was represented mainly by neutrophils with slight lymphocyte/macrophage admixture. Morphometric analysis revealed that the volume density of necrotic changes in the control RLS 40 tumors was 4.8 ± 0.5% (Figure 4, Supplementary Table S1). The administration of vaccines led to a slight increase in necrotic decay; however, this parameter was two-fold increased in the groups that received DC/ML(×2) and ML(×2) vaccines: the volume density of necrosis was 9 ± 1.5% and 10.2 ± 4.1%, respectively (Figure 4, Supplementary Table S1). The volume density of inflammatory infiltration in the control RLS 40 tumors was 6.2 ± 1.5%, and vaccination with DC/L(×2), DC/ML(×2) and CIMV/ML(×2) led to 3.5-, 2.2-, and 2-fold increase of this parameter compared to the control group (Figure 4, Supplementary  Table S1).     Table 3), however, single immunization with these types of vaccines did not induce apoptosis in tumor tissue ( Table 3). As for liposome-based vaccines, both single and double vaccinations with these vaccines were ineffective with respect to mitotic activity and apoptosis induction in tumor tissue (Table 3).

Liver Toxicity of Vaccines in RLS40-Bearing Mice
Cationic liposomes accumulate and undergo biotransformation in the liver, causing toxic effects. To assess the damaging potential of vaccines under the study, morphological changes in the liver tissue and biochemical parameters of the peripheral blood, showing hepatic destruction, were assessed.
We demonstrated that the vaccines under study did not induce significant hepatotoxicity. Indeed, the blood biochemistry of mice with RLS40 showed that tumor development led to a 1.7-fold increase in the AST level compared with the healthy animals (Supplementary Table S2). The values of the other parameters under study (ALT, ALK, and total protein) in tumor-bearing mice did not differ from the healthy level (Supplementary  Table S2). Single and double vaccination did not increase parameters reflecting the cytolysis of hepatocytes (ALT and AST) compared to the control but led to 1.4-1.9-fold increase in the parameter reflecting the cholestasis (ALK), especially during boosted treatment (Supplementary Table S2). Protein synthetic liver function was not affected in any experimental groups (Supplementary Table S2).
Liver morphometry of the RLS40-bearing mice demonstrated that tumor development was accompanied by the expansion of destructive changes in the liver parenchyma

Liver Toxicity of Vaccines in RLS40-Bearing Mice
Cationic liposomes accumulate and undergo biotransformation in the liver, causing toxic effects. To assess the damaging potential of vaccines under the study, morphological changes in the liver tissue and biochemical parameters of the peripheral blood, showing hepatic destruction, were assessed.
We demonstrated that the vaccines under study did not induce significant hepatotoxicity. Indeed, the blood biochemistry of mice with RLS40 showed that tumor development led to a 1.7-fold increase in the AST level compared with the healthy animals (Supplementary Table S2). The values of the other parameters under study (ALT, ALK, and total protein) in tumor-bearing mice did not differ from the healthy level (Supplementary  Table S2). Single and double vaccination did not increase parameters reflecting the cytolysis of hepatocytes (ALT and AST) compared to the control but led to 1.4-1.9-fold increase in the parameter reflecting the cholestasis (ALK), especially during boosted treatment (Supplementary Table S2). Protein synthetic liver function was not affected in any experimental groups (Supplementary Table S2).
Liver morphometry of the RLS40-bearing mice demonstrated that tumor development was accompanied by the expansion of destructive changes in the liver parenchyma (Supplementary Figure S2, Table S3): the volume densities of dystrophy and necrosis were increased by the factors of 1.4 and 4.2, respectively, compared with the healthy mice (Sup-

Liver Toxicity of Vaccines in RLS40-Bearing Mice
Cationic liposomes accumulate and undergo biotransformation in the liver, causing toxic effects. To assess the damaging potential of vaccines under the study, morphological changes in the liver tissue and biochemical parameters of the peripheral blood, showing hepatic destruction, were assessed.
We demonstrated that the vaccines under study did not induce significant hepatotoxicity. Indeed, the blood biochemistry of mice with RLS40 showed that tumor development led to a 1.7-fold increase in the AST level compared with the healthy animals (Supplementary Table S2). The values of the other parameters under study (ALT, ALK, and total protein) in tumor-bearing mice did not differ from the healthy level (Supplementary  Table S2). Single and double vaccination did not increase parameters reflecting the cytolysis of hepatocytes (ALT and AST) compared to the control but led to 1.4-1.9-fold increase in the parameter reflecting the cholestasis (ALK), especially during boosted treatment (Supplementary Table S2). Protein synthetic liver function was not affected in any experimental groups (Supplementary Table S2).
Liver morphometry of the RLS40-bearing mice demonstrated that tumor development was accompanied by the expansion of destructive changes in the liver parenchyma (Supplementary Figure S2, Table S3): the volume densities of dystrophy and necrosis were increased by the factors of 1.4 and 4.2, respectively, compared with the healthy mice (Supplementary Figure S2, Table S3). Vaccination with DC/L(×2) and L(×2) was the most damaging and was accompanied by an increase in the total destructive changes in the liver by

Liver Toxicity of Vaccines in RLS40-Bearing Mice
Cationic liposomes accumulate and undergo biotransformation in the liver, causing toxic effects. To assess the damaging potential of vaccines under the study, morphological changes in the liver tissue and biochemical parameters of the peripheral blood, showing hepatic destruction, were assessed.
We demonstrated that the vaccines under study did not induce significant hepatotoxicity. Indeed, the blood biochemistry of mice with RLS40 showed that tumor development led to a 1.7-fold increase in the AST level compared with the healthy animals (Supplementary Table S2). The values of the other parameters under study (ALT, ALK, and total protein) in tumor-bearing mice did not differ from the healthy level (Supplementary  Table S2). Single and double vaccination did not increase parameters reflecting the cytolysis of hepatocytes (ALT and AST) compared to the control but led to 1.4-1.9-fold increase in the parameter reflecting the cholestasis (ALK), especially during boosted treatment (Supplementary Table S2). Protein synthetic liver function was not affected in any experimental groups (Supplementary Table S2).
Liver morphometry of the RLS40-bearing mice demonstrated that tumor development was accompanied by the expansion of destructive changes in the liver parenchyma (Supplementary Figure S2, Table S3): the volume densities of dystrophy and necrosis were increased by the factors of 1.4 and 4.2, respectively, compared with the healthy mice (Supplementary Figure S2, Table S3). Vaccination with DC/L(×2) and L(×2) was the most damaging and was accompanied by an increase in the total destructive changes in the liver by 1.6-and 1.5-fold compared with control and by 3.6-and 3.3-fold compared with healthy mice, respectively (Supplementary Figure S2, Table S3). All other types of vaccines did 2.1 ± 0.7* 5.8 ± 1.4 * * Differences from the control mice were statistically significant at p ≤ 0.05. The largest alterations are indicated in bold.

Liver Toxicity of Vaccines in RLS 40 -Bearing Mice
Cationic liposomes accumulate and undergo biotransformation in the liver, causing toxic effects. To assess the damaging potential of vaccines under the study, morphological changes in the liver tissue and biochemical parameters of the peripheral blood, showing hepatic destruction, were assessed.
We demonstrated that the vaccines under study did not induce significant hepatotoxicity. Indeed, the blood biochemistry of mice with RLS 40 showed that tumor development led to a 1.7-fold increase in the AST level compared with the healthy animals (Supplementary Table S2). The values of the other parameters under study (ALT, ALK, and total protein) in tumor-bearing mice did not differ from the healthy level (Supplementary Table S2). Single and double vaccination did not increase parameters reflecting the cytolysis of hepatocytes (ALT and AST) compared to the control but led to 1.4-1.9-fold increase in the parameter reflecting the cholestasis (ALK), especially during boosted treatment (Supplementary  Table S2). Protein synthetic liver function was not affected in any experimental groups (Supplementary Table S2).
Liver morphometry of the RLS 40 -bearing mice demonstrated that tumor development was accompanied by the expansion of destructive changes in the liver parenchyma (Supplementary Figure S2, Table S3): the volume densities of dystrophy and necrosis were increased by the factors of 1.4 and 4.2, respectively, compared with the healthy mice (Supplementary Figure S2, Table S3). Vaccination with DC/L(×2) and L(×2) was the most damaging and was accompanied by an increase in the total destructive changes in the liver by 1.6-and 1.5-fold compared with control and by 3.6-and 3.3-fold compared with healthy mice, respectively (Supplementary Figure S2, Table S3). All other types of vaccines did not enhance the destructive effect of tumor growth on the liver parenchyma of tumor-bearing mice (Supplementary Figure S2, Table S3).

Immunomodulatory Effect of Different Types of Vaccines in RLS 40 -Bearing Mice
Histologically, the spleen of control RLS 40 -bearing mice had a typical structure and was comparable with the spleen of healthy mice: the white pulp was developed moderately and consisted of lymphoid follicles, most of which were isolated from each other ( Figures 5 and 6, Supplementary Table S4). The structural organization of the thymus in the control group also did not differ from that of healthy mice and was represented equally by cortex and medulla (Figures 5 and 6, Supplementary Table S4). The morphometric study of the spleen and thymus revealed that the tumor progression had no immunomodulatory effects: morphofunctional parameters of the spleen and thymus did not differ from those of healthy animals ( Figures 5 and 6, Supplementary Table S4).
Both single and double vaccination of tumor-bearing animals activated the immune response, which was expressed in an increase in the volume density of the white pulp and the diameter of lymphoid follicles in the spleen as well as in an increase in the volume density of the thymus cortex and cortex-medulla index in the thymus (Figures 5 and 6, Supplementary Table S4). The most efficient immunomodulators were DC-and CIMVbased vaccines, particularly during boosted treatment. Administration of DC/ML (×2) led to a 1.6-and 1.3-fold increase in the volume density of the white pulp and the diameter of the lymphoid follicles of the spleen, respectively, and a 1.4-and 2.3-fold increase in the volume density of the cortex and cortex-medulla index in the thymus, respectively, compared to healthy and untreated tumor-bearing mice ( Figure 6, Supplementary Table S2). The administration of CIMV/ML (×2) led to a more significant stimulation of the spleen and thymus: the white pulp and lymphoid follicles of the spleen were increased by a factor of 1.7 and 1.4, respectively, and the thymus cortex and cortical-cerebral index were increased by a factor of 1.7 and 11.8, respectively, compared with healthy and control animals ( Figure 6, Supplementary Table S2). L and ML liposome-based vaccines activated the immune organs to a lesser extent. Thus, double immunization with CIMV-based vaccines demonstrated the most pronounced immunomodulatory effects. Table S2). Protein synthetic liver function was not affected in any experimental groups (Supplementary Table S2).

(Supplementary
Liver morphometry of the RLS40-bearing mice demonstrated that tumor development was accompanied by the expansion of destructive changes in the liver parenchyma (Supplementary Figure S2, Table S3): the volume densities of dystrophy and necrosis were increased by the factors of 1.4 and 4.2, respectively, compared with the healthy mice (Supplementary Figure S2, Table S3). Vaccination with DC/L(×2) and L(×2) was the most damaging and was accompanied by an increase in the total destructive changes in the liver by 1.6-and 1.5-fold compared with control and by 3.6-and 3.3-fold compared with healthy mice, respectively (Supplementary Figure S2, Table S3). All other types of vaccines did not enhance the destructive effect of tumor growth on the liver parenchyma of tumorbearing mice (Supplementary Figure S2, Table S3).

Immunomodulatory Effect of Different Types of Vaccines in RLS40-Bearing Mice
Histologically, the spleen of control RLS40-bearing mice had a typical structure and was comparable with the spleen of healthy mice: the white pulp was developed moderately and consisted of lymphoid follicles, most of which were isolated from each other ( Figures 5 and 6, Supplementary Table S4). The structural organization of the thymus in the control group also did not differ from that of healthy mice and was represented equally by cortex and medulla ( Figures 5 and 6, Supplementary Table S4). The morphometric study of the spleen and thymus revealed that the tumor progression had no immunomodulatory effects: morphofunctional parameters of the spleen and thymus did not differ from those of healthy animals ( Figures 5 and 6, Supplementary Table S4).
Both single and double vaccination of tumor-bearing animals activated the immune response, which was expressed in an increase in the volume density of the white pulp and the diameter of lymphoid follicles in the spleen as well as in an increase in the volume density of the thymus cortex and cortex-medulla index in the thymus ( Figures 5 and 6, Supplementary Table S4). The most efficient immunomodulators were DC-and CIMVbased vaccines, particularly during boosted treatment. Administration of DC/ML (×2) led to a 1.6-and 1.3-fold increase in the volume density of the white pulp and the diameter of the lymphoid follicles of the spleen, respectively, and a 1.4-and 2.3-fold increase in the volume density of the cortex and cortex-medulla index in the thymus, respectively, compared to healthy and untreated tumor-bearing mice ( Figure 6, Supplementary Table S2). The administration of CIMV/ML (×2) led to a more significant stimulation of the spleen and thymus: the white pulp and lymphoid follicles of the spleen were increased by a factor of 1.7 and 1.4, respectively, and the thymus cortex and cortical-cerebral index were increased by a factor of 1.7 and 11.8, respectively, compared with healthy and control animals ( Figure 6, Supplementary Table S2). L and ML liposome-based vaccines activated the immune organs to a lesser extent. Thus, double immunization with CIMV-based vaccines demonstrated the most pronounced immunomodulatory effects.    Dash lines represent the levels of the control group. Data are presented as mean ± S.E.M. Data were statistically analyzed using one-way ANOVA with post-hoc Fisher test. Statistically reliable differences between the groups: * p < 0.05, ** p < 0.005.

Immunologic Characteristics of Antitumor Immune Responses Caused by Vaccination
Antitumor and antimetastatic activity of the vaccines (Figure 3 and Supplementary Figure S1) was mediated by stimulation of adaptive immunity, including T lymphocytes, that was corroborated by our data from CTL assays for DC- [33], liposome- [36], and CIMVsbased ( Figure 2) vaccines as well as morphological changes in the spleen and thymus of experimental animals revealed by morphometry ( Figures 5 and 6, Supplementary Table S4).
To study the characteristics of antitumor responses activated by the vaccines in detail, analysis of the expression of immune checkpoints (ICs) in immune cells and polarization of T helper responses were performed.
Among stimulatory ICs, moderate upregulation of CD27 and CD28 expression was observed in the groups once injected with DC-based and liposomal vaccines, whereas double vaccination resulted in downregulation of the expression of these ICs ( Figure 7A,B). CIMV-based vaccines tended to slightly upregulate the expression of CD27 ( Figure 7A) and reduce the CD28 expression 2-3-fold compared with the control (Figure 7A,B). Analysis of the expression of master-regulators controlling T-helper differentiation revealed that increase in Tbet expression was observed in all experimental groups that received a single vaccination ( Figure 8A) and did not decline to the control level in the Surprisingly, CD40L expression was strongly downregulated in all experimental groups compared with the control ( Figure 7C). Generally, CD40L is expressed mostly on activated CD4 + T-helpers and, on the one hand, promotes co-stimulatory activity of APC that is crucial for generating CD8 + T cell immunity [52]. On the other hand, CD40L is essential for inducing Th2 cells and the subsequent activation of humoral immunity [53,54]. Hence, vaccination-induced downregulation of CD40L may result from decreased activation levels of CD4+ T cells associated with low tumor burden in vaccinated groups. Furthermore, CD40L expression is characterized by fast kinetics and up-regulated within hours after stimuli and declines to a basal level during the next 24 h [52]. However, in this study, CD40L levels were evaluated at the end of the experiment, at least 7 days after the booster vaccine administration. However, a more likely explanation of the drop in CD40L expression levels is that vaccines under the study mediated the switch from the Th2-mediated humoral immune response observed in tumor-bearing animals (control) to Th1 immunity and antitumor CTL development. This statement was confirmed by analyzing the expression levels of Th-specific master regulators (see below).
The expression of 4-1BBL, an antigen-presenting cell-derived immunostimulating checkpoint, was upregulated in splenocytes of mice once challenged with DC-and liposomebased vaccines and with CIMV/L vaccine ( Figure 7D Regarding negative immune checkpoints, all tested vaccines significantly downregulated the expression of PD-1, one of the most critical inhibitory molecules suppressing T cell-mediated antitumor response ( Figure 7E). DC-and liposomal-based vaccines resulted in a two-fold decrease in CTLA4 expression compared with the control, whereas CIMVbased vaccines had no effects on the expression of this factor ( Figure 7F). Nevertheless, CIMV-based vaccines efficiently downregulated the expression of another co-inhibitory factor, TIGIT, regardless of the number of vaccine administration ( Figure 7G), whereas DCand liposome-based vaccines efficiently downregulated TIGIT expression only in the case of double vaccination ( Figure 7G). Downregulation of TIM3 expression was observed after double vaccination with DC-and liposome-based vaccines and CIMV/ML vaccine. A single administration of any vaccines under study did not affect TIM3 expression ( Figure 7H).
Analysis of the expression of master-regulators controlling T-helper differentiation revealed that increase in Tbet expression was observed in all experimental groups that received a single vaccination ( Figure 8A) and did not decline to the control level in the groups DC/L, ML, and CIMV/ML after a double vaccination, indicating the activation of the Th1 response. Th2-related transcription factor GATA3 was strongly downregulated in all groups of vaccinated mice ( Figure 8B). The expression level of the RORγ master regulator inducing the development of Th17 immune response was shown to be significantly upregulated, by 4-11-fold, compared with the control after both single and double vaccination with all types of tested vaccines, excluding DC/L(×2) and CIMV/L(×2). In all cases, single vaccination was more efficient than double one in stimulation of RORγ expression ( Figure 8C). Foxp3 transcription factor is responsible for developing T regulatory cells and its level represents the extent of suppression of T cell response. It was demonstrated that the levels of Foxp3 expression were not increased after vaccination in all groups compared with the control (Figure 8D), indicating the lack of induction of T regulatory cells. It is worth mentioning that up to a 2-fold decrease in Foxp3 expression levels was observed after vaccination in some groups (see Figure 8D).
Thus, we showed that RLS 40 progression in mice led to the formation of a predominant Th2-mediated pro-humoral immune response that was confirmed by the activation of GATA3 with low levels of T-bet and RORγ and a high level of CD40L immune checkpoints related to B cell development and activation. Vaccination of RLS 40 -bearing mice with DC-, liposome-, or CIMV-based vaccines resulted in switching to cellular immunity by activation of Th1/Th17 cells as well as the induction of positive immune checkpoint 4-1BBL and downregulation of suppressive immune checkpoints in a raw PD-1 >>> TIGIT > CTLA4 > TIM3. All these immunological alterations caused significant retardation of tumor growth and inhibition of metastasis development.
( Figure 8C). Foxp3 transcription factor is responsible for developing T regulatory cells and its level represents the extent of suppression of T cell response. It was demonstrated that the levels of Foxp3 expression were not increased after vaccination in all groups compared with the control (Figure 8D), indicating the lack of induction of T regulatory cells. It is worth mentioning that up to a 2-fold decrease in Foxp3 expression levels was observed after vaccination in some groups (see Figure 8D). Data were statistically analyzed using one-way ANOVA with post-hoc Tukey test. Statistically reliable differences between the groups: * p < 0.05, ** p < 0.01.
Thus, we showed that RLS40 progression in mice led to the formation of a predominant Th2-mediated pro-humoral immune response that was confirmed by the activation of GATA3 with low levels of T-bet and RORγ and a high level of CD40L immune  40 -bearing mice treated with DC-, liposome-, and CIMV-based vaccines determined by qPCR and TaqMan probes. The expression level of gene-specific product was normalized to the expression level of β-actin. The expression levels of genes were presented on the 21st day after tumor transplantation. Data are presented as mean ± S.E.M. Data were statistically analyzed using one-way ANOVA with post-hoc Tukey test. Statistically reliable differences between the groups: * p < 0.05, ** p < 0.01.

Discussion
Cell-free antitumor vaccines have many advantages over cellular vaccines, such as prolonged storage without loss of immunotherapeutic activity, insensitivity to the immunosuppressive tumor microenvironment, and potentially better standardization. Therefore, the development of anticancer cell-free vaccines has great potential.
Previously, we developed DC-based vaccines composed of murine BM-DCs loaded with total tumor RNA by using cationic liposomes L or DC-targeted mannosylated liposomes ML that exhibited high antimetastatic activity in a murine melanoma B16-F10 model in vivo [32,33,36]. In the present study, these DC-based vaccines were applied for the treatment of drug-resistant lymphosarcoma RLS 40 and demonstrated superior antitumor and antimetastatic activity (Figure 3, Supplementary Figure S1). It is interesting that the change in the route of administration of DC vaccine from intravenous [32,33] to subcutaneous did not affect the antitumor activity of the vaccine.
The main result of this study is the strong antitumor and antimetastatic activity of cellfree liposome-and DC-derived CIMV-based vaccines in a murine RLS 40 lymphosarcoma model in vivo, with the efficiency being similar to that of DC-based anticancer vaccines.
The antitumor potential of liposomal vaccines, more commonly called mRNA vaccines, is under intensive investigation, especially after the success of mRNA vaccines developed for the prevention of SARS-CoV-2. Antitumor liposomal vaccines are generally composed of cationic liposomes/polymers and mRNAs encoding particular tumor-associated antigens (TAA), such as TRP2 [55,56], MUC1 [57], cytokeratin 19 [58], E7 [56,59], neoantigens [60], etc. or model antigen OVA [43,56,59,61,62]. These mRNA vaccines efficiently delivered TAA-encoding mRNA to DCs and subsequently activated antitumor CD8 + and CD4 + T-cell immune responses, reduced tumor-associated immunosuppression, and finally significantly decreased tumor growth and metastasis in different murine tumor models [43,[55][56][57][58][59][60][61][62]. Our liposomal vaccines were composed of cholesterol/spermine-based cationic liposomes L or DC-targeted mannosylated liposomes ML complexed with total tumor-derived RNA (source of multiple tumor antigens). Their significant potential to deliver RNA in murine DCs ex vivo and activate antitumor CTLs in vivo was demonstrated in our previous studies [32,33,36]. Here, we demonstrated that liposomal vaccines under the study had great antitumor potential in vivo, significantly suppressing RLS 40 lymphosarcoma growth (Figure 3), inhibiting hepatic metastasis ( Figure S1), and activating antitumor immunity (Figures 6-8, Table S4). Note that there were no significant differences between the in vivo antitumor activity of the vaccines based on L-and DC-targeted ML liposomes (Figures 3 and S1), except for lower hepatotoxicity of ML liposomes (Table S3). Indeed, in our previous work, we demonstrated an equal efficiency of core L liposomes and their derivative ML liposomes to deliver RNA in murine DCs ex vivo [33], which resulted from the basic superior transfection activity of core L liposomes.
Another type of cell-free vaccine studied in this work use artificial vesicles prepared from loaded DCs using cytoskeleton-disrupting agent cytochalasin B. Being analogous to natural extracellular vesicles, CIMVs represent submicrometer-sized vesicles surrounded by plasma membranes comprising functional cell surface receptors on the outer vesicle surface and cytosolic nucleic acids and proteins in the vesicle interior [63]. Currently, biological activities of membrane vesicles obtained from cytochalasin-B-treated cells have been actively investigated. A significant contribution to this area was made by the laboratory of Prof. Rizvanov A.A. of Kazan Federal University, Russia [24,25,27,28,[64][65][66][67]. To date, delivery efficiency, angiogenic, and immunomodulatory properties of CIMVs obtained from human [19,25,37,[64][65][66] and mouse [24] mesenchymal stem cells, human neuroblastoma SH-SY5Y [28,67], prostate cancer PC3 [27], and HeLa cells [68] have been investigated. CIMVs were demonstrated to be non-toxic, have no fusion specificity to the cells in vitro, and possess angiogenic activity. Immunomodulatory activities of CIMVs depended on the presence of immune-related molecules on their surface or inside the vesicles. Indeed, CIMVs derived from human mesenchymal stem cells overexpressed IL-2 [25] or TRAIL, while PTEN and IFN-β1 [65] were demonstrated to stimulate the proliferation of T lymphocytes in vitro. Furthermore, CIMVs obtained from HeLa cells transduced with genes encoding tumor-associated antigens significantly activated and maintained the antigen-specific stimulation and expansion of CAR-T lymphocytes [68]. Nevertheless, the immunostimulatory properties of DC-derived CIMVs have been insufficiently studied. Recently, Oshchepkova et al. have demonstrated that DC-derived CIMVs could transfer nucleic acids to the cells without any fusion specificity [37]. In the present study, we demonstrated for the first time the high antitumor and immunostimulatory activity in vivo of CIMVs obtained from murine DCs. We showed that CIMVs prepared from DCs loaded with total tumor RNA can directly prime antitumor CTLs ex vivo ( Figure 2) and activate potent antitumor immune response in vivo (Figures 3, 7, 8 and S1), which is similar to or even higher than that of DC-based vaccines. Now consider common and different features of immune responses activated by different types of the vaccines under the study. It was demonstrated that DC-and CIMV-based vaccines being administered subcutaneously had comparable potential to stimulate highly efficient antitumor CTLs in vivo, whereas liposomal vaccines were weaker CTL inducers ( Figure 2). Previously, we have shown that CTLs activated in vivo by intravenously administered liposomal vaccines lysed tumor cells with the same efficiency as DC-based vaccines [36]. Undoubtedly, the route of administration of the vaccines affects their activity. Hence, intravenous inoculation of liposomal vaccines should potentially be preferred over subcutaneous one. Despite different levels of CTL induction, the antitumor efficiencies of the different vaccines were similar: tumor volumes/weights ( Figure 3) and the amounts of liver metastasis ( Figure S1) were significantly decreased, regardless of the vaccine type. However, different details of immune responses activated by the vaccines and their diverse effects on tumor tissue were observed. Indeed, immunostimulating properties of DC-and CIMV-based vaccines had common patterns, whereas liposomal vaccines were slightly different from them, namely, in weaker activation of lymphoid organs (minimal volume of white pulp in spleen and cortex in thymus especially for L liposomes ( Figure 6, Table S4)) and lower volumes of inflammatory infiltration in tumor tissues (Table S1). Additionally, a greater number of mitotic tumor cells and a lesser activation of apoptosis in tumor tissue were observed in the case of double vaccination with liposomal vaccines (Table 3). Nevertheless, the changes in the expression levels of immune checkpoints triggered by liposomal vaccines were conversely similar to those of DC-based vaccines, whereas the effects of CIMV-based vaccines differed (Figure 7). It was demonstrated that CIMV-based vaccines more efficiently activated the expression of 4-1BBL and CTLA4 and suppressed the expression of CD28, TIGIT, and TIM3 compared with DC-and liposomal vaccines. Despite these indicated differences, all the vaccines under the study induced Th1/Th17 immune response, with a more polarized Th17 response in the case of CIMV-based vaccines ( Figure 8).
Interestingly, the booster vaccination did not improve the overall antitumor efficacy of the vaccines under the study (Figure 3, Table S1). A similar absence of the advantages of booster DC-based was previously observed [39]. However, some differences in the histological characteristics of tumor nodes and the details of activated immune responses were found between single and double vaccination regimens. Thus, double immunization with all tested vaccines more efficiently activated proliferation of lymphocytes in lymphoid organs of experimental animals compared with single vaccination ( Figure 6, Table S4). Moreover, the number of vaccinations affects the expression of immune checkpoints in splenocytes. In most cases, double vaccination with DC-based and liposomal vaccines resulted in down-regulation of both positive and negative immune checkpoints compared with a single dose, whereas booster immunization with CIMV-based vaccines, in contrast, up-regulated the expression of CD27 and 4-1BB-L (positive) as well as CTLA-4 and TIM3 (negative) immune checkpoints (Figure 7). We can assume that the missing effect of the booster vaccination can be associated with the untimely restimulation of the immune response. Probably, a 7-day interval between the vaccines is not optimal and should be reduced to 3-4 days. This problem needs further investigation.
The main positive immunologic result of the vaccination of tumor-bearing mice with vaccines under the study was the activation of Th1/Th17 immune response. Generally, antitumor DC-based vaccines frequently activate the Th1/Th17 response, which correlates well with the strong antitumor activity of the vaccines [69][70][71][72][73]. It is known that Th17 cells modulate inflammation, are negatively correlated with Tregs [74,75], and play a pivotal role in the activation of tumor-specific CD8+ T cells [76]. Adoptive transport of tumor-specific Th17 cells was demonstrated to retard tumor growth and improve survival in murine tumor models [75,77]. Thus, protective inflammation elicited by Th17 cells promotes the activation of antitumor immunity and helps Th1 to stimulate tumor-specific CD8+ T cells. Interestingly, CIMV-based and liposomal vaccines also activated Th1/Th17 cells, indicating apparently similar mechanisms of action of the vaccines to DCs. We suppose that the source of tumor antigens to load DCs determines to a greater extent the activation of the Th17 cells after the vaccination. Indeed, a whole tumor cell set of antigens was used both in our study (total tumor RNA) and in other studies (tumor cell lysates [69,73], irradiated tumor cells [71,72], DC/tumor cell hybrids [70]) where Th17 response was polarized after the vaccination. It is possible that self-antigens contained in such sources of antigens can trigger autoimmune reactions that play an important role in the activation of Th17 cells.
Thus, we have demonstrated that non-cellular vaccines based on CIMVs or liposomes have a high antitumor potential and can efficiently activate antitumor immunity that results in significant retardation of tumor growth and inhibition of metastasis development. Among the vaccines under the study, CIMV-based vaccines seem to be the most promising and provide the basis for developing a novel highly efficient antitumor immunotherapeutic approach.

Conclusions
In this study, we demonstrated that cell-free CIMV-based vaccines exhibited superior antitumor and antimetastatic activity in a tumor model in vivo. CIMV-based vaccines stimulated CTLs in vivo that highly efficiently lysed tumor cells. The antitumor activity of the different types of the vaccines was similar: they exhibited similar ability to decrease the sizes of primary tumor nodes and the number of metastases regardless of the vaccine type in the case of single vaccination, and the booster vaccination did not improve the overall antitumor efficacy. CIMVbased vaccine application resulted in switching of humoral immune response formed upon RLS 40 progression in mice to cellular immunity by activation of Th1/Th17 cells as well as induction of positive immune checkpoint 4-1BBL and downregulation of suppressive immune checkpoints in a raw PD-1 >>> TIGIT > CTLA4 > TIM3.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/pharmaceutics14112542/s1, Figure S1: The effect of different types of the vaccines on the liver metastases in RLS 40 -bearing mice; Table S1: Morphometry of RLS 40 lymphosarcoma tumor tissue of mice without treatment (control) and after vaccines administration; Figure S2: The effect of vaccines on liver in RLS 40 -bearing mice; Table S2: The effect of the vaccines on blood biochemistry of RLS 40 -bearing mice; Table S3: Liver morphometry of RLS 40 lymphosarcoma-bearing mice without treatment and after vaccine administration; Table S4: Spleen and thymus morphometry of RLS 40 lymphosarcoma-bearing mice without treatment (control) and after vaccine administration.