Proof-of-Concept Study of Liposomes with a Set of SARS-CoV-2 Viral Peptidic T-Cell Epitopes as a Vaccine

Abstract

Potential nonameric epitopes of CD8⁺ T lymphocytes were selected from the composition of structural, accessory, and nonstructural proteins of the SARS-CoV-2 virus (13 peptides) and a 15-mer epitope of CD4⁺ T lymphocytes, from the S-protein, based on the analysis of publications on genome-wide immunoinformatic analysis of T-cell epitopes of the virus (Wuhan strain), as well as a number of clinical studies of immunodominant epitopes among patients recovering from COVID-19 disease. The peptides were synthesized and five compositions of 6–7 peptides were included in liposomes from egg phosphatidylcholine and cholesterol (~200 nm size) obtained by extrusion. After double subcutaneous immunization of conventional mice, activation of cellular immunity was assessed by the level of cytokine synthesis by splenocytes in vitro in response to stimulation with relevant peptide compositions. Liposomal formulation exhibiting the best result in terms of the formation of specific cellular immunity in response to vaccination was selected for further experiments. Evaluation of the protective efficacy of this formulation in an infectious mouse model showed a positive trend in the frequency of occurrence of hyaline-like membranes in the lumen of the alveoli, as well as a somewhat lower severity of microcirculatory disorders. The latter circumstance can potentially help reduce the severity of the disease and prevent its adverse outcomes. A method to produce liposome preparations with peptide compositions for long-term storage is under development.

INTRODUCTION

Traditional vaccines for the treatment and prevention of infectious diseases are live attenuated or inactivated/killed pathogens. They do not need adjuvants (costimulators) as they contain not only antigens, but also other components of bacteria or viruses that effectively activate several components of the innate immune system. However, such vaccines can cause allergy-related and autoimmune reactions [1]. The so-called subunit vaccines usually contain only surface antigens (proteins or peptides) of pathogens, which reduces not only the allergenicity of vaccines, but also their immunogenicity. Packing antigens into particles comparable to viruses or bacteria in size (from hundreds of nanometers to several microns) minimizes the risks of the described adverse reactions, also protecting the antigens from premature degradation and contributing to the creation of an antigen depot [2, 3]. The particles captured by the antigen-presenting cells (APC; dendritic cells and macrophages) undergo intracellular processing. Depending on the cleavage pathway, antigens are presented on the surface of the plasmalemma for the friend-or-foe recognition by T lymphocytes in the form of epitopes, i.e., peptide fragments 9 a.a. long, complexed by class I major histocompatibility complex (MHCI) proteins, or fragments 14–20 a.a. long, complexed by class II MHC proteins (MHCII). Peptide–MHCI complexes stimulate naïve CD8⁺ T lymphocytes and convert them into cytotoxic T lymphocytes (CTLs) responsible for cellular immunity and destruction of infected cells. Peptide–MHCII complexes activate naïve CD4⁺ T lymphocytes, which proliferate and differentiate into subpopulations of the so-called helper T_H1 and/or T_H2 cells, depending on the type of infection [4]. T_H1 cells secrete various cytokines that activate and regulate CTLs. Thus, to form cellular immunity, it is necessary to stimulate both CD8⁺ and CD4⁺ T lymphocytes.

Being adjuvants as such, liposomes are of interest as antigen carriers due to their high biocompatibility and plasticity. The ability of liposomes to induce an immune response to antigens encapsulated in the internal volume or associated with the surface was first described by Allison and Gregoriadis [5]. Depending on the composition and structure, they can simultaneously activate various signal transduction pathways and induce specific T- and/or B-cell responses: antigens exposed on the surface of liposomes can stimulate B lymphocytes, causing a humoral immune response, and induce T-cell reactions; encapsulated antigens that require intracellular degradation of liposomes are able to induce CTLs. Currently, liposomal vaccines are used against influenza and hepatitis A viruses, malaria and Varicella zoster virus (Inflexal®, Epaxal®, Mosquirix®, and Shingrix®, respectively); a number of liposomal preparations are undergoing clinical trials as preventive and therapeutic vaccines against malaria, influenza, tuberculosis, HIV, and dengue fever [6–10].

The adjuvant properties of liposomes can be enhanced or directed along the path of one or another type of immune response using immunostimulants, specific ligands that cause the activation of APC receptors recognizing pathogen-associated molecular patterns (PAMPs) [11]. One of the PAMPs are unmethylated CpG motifs of bacterial DNA, which are much less common in eukaryotic chromosomes; therefore, synthetic CpG-ODNs have immunostimulatory activity. CpG motifs are recognized by the TLR9 receptor, which is expressed in the membranes of endosomal compartments of many immunocompetent cells, including B cells, monocytes, NK cells, dendritic cells, and macrophages [12, 13]. As a result, the production of proinflammatory cytokines and chemokines is stimulated and the expression of MHCII and costimulatory molecules (CD40, CD80, CD83, and CD86) is increased. Synthetic СpG-ODNs are successfully used in the design of liposomal vaccines; the immunostimulant is coencapsulated in the inner water volume of liposomes or adsorbed onto cationic liposomes [13].

Peptide vaccines are safer than vaccines based on full-length antigens or other molecules of pathogenic origin that may contain oncogenic sequences [14]. A number of preclinical studies have shown the immunogenicity of vaccines based on liposomes with encapsulated or surface-bound specific peptides [13, 15–18] (when administered together with CpG-ODN). The authors of the works [15, 16] conjugated nonameric T‑cell epitopes of the nucleocapsid [15] and nonstructural protein polyprotein 1a [16] of the SARS coronavirus (SARS-CoV) to the surface of liposomes. In the first case, four CTL epitopes restricted to HLA-A*0201 (HLA-A2 is one of the most common allelic variants of MHCI molecules) were identified using transgenic mice and a recombinant adenovirus expressing eight epitopes predicted in silico. Two of the liposomal peptides stimulated a specific T-cell response. One of these vaccines caused clearance of vaccinia virus expressing SARS-CoV epitopes in mice [15]. In the second case, out of 30 predicted peptides, eight molecules significantly stimulated the response of CTLs and were conjugated to liposomes. Of these, seven were active against the virus in vivo, and one vaccine variant induced the formation of memory cells [16]. Other authors [17] studied various methods of encapsulation of 24-mer peptides with various physicochemical properties and containing the SIINFEKL sequence (an immunodominant CTL epitope of ovalbumin) into cationic liposomes in order to achieve effective loading of liposomes. Liposomes efficiently delivered peptides to dendritic cells, which was followed by SIINFEKL-specific activation of CD8⁺ T cells. When designing a vaccine against type A influenza virus [18], ten conserved peptides of structural and nonstructural proteins were simultaneously encapsulated in liposomes, of which five nonameric epitopes were designated as promising for pandemic swine influenza H1N1 in silico. In combination with crystalline sodium urate, vaccination of piglets in an intranasal aerosol form caused an increase in the frequency of virus-specific T_H cells and memory cells; it also stimulated the response of CTLs, which ultimately led to partial protection of animals from fever and lung damage.

In connection with the pandemic of a new coronavirus infection COVID-19, the world scientific community has begun to accelerate the development of vaccines, which can activate both humoral and cellular immunity. However, data on the virulence of proteins encoded by the viral genome are still insufficient. It is important to search for the most immunogenic epitopes not only in the spike protein (S-protein) of the SARS-CoV-2 virus, but also in other membrane proteins and proteins of the viral capsid and nucleoprotein. Judging by the first published data of immunoinformatic analysis of the full-length virus genome, in the ranked list of nonameric epitopes of CTLs and 15-mer epitopes of CD4⁺ T cells (T_H cells) common for all MHC alleles and covering all predominant supertypes of HLA (human leukocyte antigen) in the global population, epitopes of protein S may be inferior in immunogenicity to the epitopes of other proteins of the virus, including the nonstructural ones [18–20]. The aim of this study was to develop a prototype vaccine based on liposomes containing a composition of T-cell epitopes from the primary structures of various proteins of the SARS-CoV-2 virus. If immunogenic epitopes that are not related to the spike protein are found, there is a prospect of developing vaccines for the treatment and prevention of COVID-19, the effectiveness of which will depend little on mutations in the viral genome.

RESULTS AND DISCUSSION

Selection of T-cell epitopes for vaccine formulations. To select peptides, we analyzed published results of genome-wide immunoinformatic analysis of T-cell epitopes of the SARS-CoV-2 virus [19–21] and the results of a number of clinical studies of immunodominant epitopes in patients recovering from COVID-19 disease [22–26]. Based on a comparison of all data, the following potential epitopes were selected and synthesized: nonameric CTL epitopes of the S-protein, envelope protein (E-protein), membrane protein (M‑protein), nucleocapsid protein (N-protein), auxiliary proteins orf6 and orf8, and nonstructural proteins orf1a, orf1b, and orf3a, as well as a 15-mer helper T cell epitope of the S-protein; 14 epitopes in total. The list of synthesized peptides is shown in Table 1.

Table 1.  .

Synthetic T-cell epitopes and their physicochemical properties

No.	Epitope	Protein	Location, a.a.	Hydro-phobicity, % a.a.*	PI**	Conserva-tion	Number of patients responded	Response rate	HLA alleles	Refs.
CTL epitopes
Structural proteins
1	VGYLQPRTF	S	267–275	44.44	9.82	1	5	0.714	MHCII coverage 100%, MHCI coverage 65%	[19–21, 24, 26]***
2	YVYSRVKNL	E	56–64	33.33	10.06	0.997	12	0.545	MHCI coverage 74%; HLA-DRB1*04:01/HLA-DRB1*11:01	[20, 25]
3	KTFPPTEPK	N	361–369	44.44	9.84	0.998	10	1	HLA-A*03:01/HLA-A*11:01	[21, 22–24, 25]
4	KAYNVTQAF	N	266–274	44.44	9.69	0.999	2	0.286	MHCI coverage 74%	[20, 22, 24]
5	ATEGALNTPK	N	134–143	40	6.92	0.997	13	0.812	HLA-A*11:01	[23, 25]
6	SPRWYFYYL	N	105–113	44.44	9.44	1	4	0.8	HLA-B*07:02	[22–24, 26]
7	GMSRIGMEV	N	316–324	44.44	6.94	0.999	12	0.522	HLA-DRB1*01:01/HLA-DRB1*04:01/HLA-DRB1*07:01/HLA-DRB1*11:01	[21, 24, 25]
8	ATSRTLSYYК	M	171–179	20	9.58	0.981	3	0.6	HLA-A*01:01	[23]
Auxiliary proteins
9	KVSIWNLDY	orf6	23–31	44.44	6.66	0.995	6	0.857	From consensus sequence of CTL/Т lymphocytes: HLVDFQVTIAEILLIIMR TFKVSIWNLDYIINLII	[19, 24, 26]
10	IQYIDIGNY	orf8	71–79	33.33	3.14	0.999	1	0.143	–	[19, 24, 26]
Nonstructural proteins
11	VLWAHGFEL	Orf1b	1708–1716	1708–1716	66.67	5.17	1	2	Coverage of 85% of the world population	[24, 26]
12	TTDPSFLGRY	Orf1a	1637–1646	1637–1646	30	6.67	0.998	15	HLA-A*01:01	[23, 24, 25, 26]
13	FTSDYYQLY	Orf3a	207–215	207–215	22.22	3.14	0.999	5	HLA-A*01:01	[23, 24, 26]
TН lymphocyte epitopes
14	SYGFQPTNGVGYQPY	S	494–508	26.67	5.83	—	—	—	A good MHCI/MHCII presenter and a predicted B-cell epitope near the S protein receptor binding domain	[20]

 * https://www.peptide2.com/N_peptide_hydrophobicity_hydrophilicity.php.  ** http://isoelectric.org/index.html. *** References to the data of clinical studies are highlighted bold.

Depending on the physicochemical properties of the peptides (Table 1), the conditions for their dissolution at high concentrations (5–7 mM, or ~7–11 mg/mL, depending on the number of peptides in the composition and molecular weight of the peptide) were chosen. The concentration range was chosen so that when mixtures of the peptides were passively encapsulated in liposomes, the loading of each of them was sufficient to exhibit potential immunogenicity. In the medium of 20 mM phosphate buffer, pH was maintained in the range from 6 to 7.2. The buffer contained an isotonic sucrose solution instead of sodium chloride, both to increase the solubility of the peptides and to allow the preparation of a long shelf-life liposomal formulation, where sucrose acts as a cryoprotectant. Peptides 6 and 11 were dissolved in the organic phase together with lipids during the formation of liposomes. Unfortunately, it was not possible to select the conditions for the dissolution of peptide 9 (KVSIWNLDY) of the orf6 auxiliary protein. According to the immunoinformatic analysis [19], this peptide belongs to the consensus sequence of the epitopes of CTLs and T_H-lymphocytes, and it was identified as immunogenic in 6 out of 7 recovered patients [24, 26].

Preparation of liposomes with various compositions of the peptides. Four compositions of 6 peptides were compiled to include peptide epitopes representing various viral proteins, as well as to combine epitopes of cytotoxic and helper T cells in the vaccine, in liposome formulations L₁–L₄. Peptides 6 and 11 were inserted into the lipid bilayer and the rest were encapsulated in the inner water volume. A composition of seven water-soluble peptides for encapsulation in the L₅ liposomes (Table 2) was selected based on the results of the first cycle of experiments (see the next section for details). Liposomes were formed based on egg phosphatidylcholine (ePC) and cholesterol (Chol), ePC–Chol, 67 : 33 (mol %), by extrusion. This ratio of lipid components creates a condensed rigid liposome membrane (the so-called liquid-crystalline ordered phase of the lipid bilayer, Lo [27]), which can prevent premature loss of peptides from the inner water volume.

Table 2.

Characteristics of liposomes with encapsulated peptides

Sample ID	Peptide composition	Diameter (nm) ±SD1	PDI ± SD1	Peptides outside liposomes (%) ±SD2	Liposome incorporation, %5
Experiment I
L C	–	141.4 ± 2.1	0.047 ± 0.026	–	–
L 1	1,2,3,5,11,14 (P1)	107.5 ± 0.8	0.096 ± 0.024	40.4 ± 1.73	59.6
L 2	1,2,4,6,8,14 (P2)	123.0 ± 0.6	0.076 ± 0.009	51.9 ± 3.33	48.1
L 3	5,7,10,11,12,14 (P3)	113.4 ± 1.3	0.073 ± 0.018	57.0 ± 4.03	43
L 4	3,7,10,11,13,14 (P4)	101.0 ± 0.8	0.095 ± 0.004	48.0 ± 4.93	52
Experiment II
L C	–	193.9 ± 1.4	0.092 ± 0.015	–	–
L 1	1,2,3,5,11,14 (P1)	201.9 ± 2.5	0.081 ± 0.017	66.5 ± 6.14	33.5
L 5	1,2,3,5,8,12,14 (P5)	191.7 ± 1.7	0.080 ± 0.022	47.2 ± 7.64	52.8

¹ According to measurements on a Brookhaven 90PLUS Particle Size Analyzer (Brookhaven Instruments Corp., United States). ² Calculated without taking into account peptides 6 and 11 embedded in the liposome membrane as (mass of nonincorporated peptides    in filtrates after ultrafiltration)/(mass of initially taken peptides for encapsulation into liposomes) × 100%. ³ According to the measurement by the modified Lowry method of unincorporated peptides in filtrates after ultrafiltration, n = 8. ⁴ According to the measurement of optical density at 273 nm in filtrates after ultrafiltration, n = 3–5. ⁵ Calculated as 100% – (peptides outside liposomes, %).

To reach high loading of the peptides into liposomes upon passive encapsulation, the concentration of lipids in the aqueous phase should be as high as possible. We applied the procedure described in [18], where a mixture of liposomal membrane components was lyophilized from tert-butanol and then hydrated with a minimum volume of peptide solutions, reaching a concentration of up to 200 mg/mL for lipids. After several freeze–thaw cycles, an extremely thick suspension was obtained. It was repeatedly extruded through membranes with pores of 200 and then 100 nm [18]. The effect of liposome size, which can vary from 100–150 nm to 1 µm, on the effectiveness of vaccination is multidirectional and depends on the route of administration (subcutaneous, intramuscular, or intradermal), composition of the components, etiology of infection, etc. [7–10]. Here, we chose extrusion through 200-nm pores in order to maximize liposome capacity. However, the sizes of the resulting liposomes turned out to be significantly smaller than the specified ones (Table 2, experiment I). At a lipid concentration of 200 mg/mL (after extrusion, lipid concentration was 140 mg/mL due to losses on the membrane), it is hardly possible to obtain more or less uniform monolamellar vesicles with a size of about 200 nm, due to stereometric considerations. Besides, the dissolved peptides also affected the process of vesicle formation: liposomes without peptides, control liposomes (L_C), were still slightly larger than the other samples (~141 nm versus 101–123 nm). Therefore, we modified the liposome preparation procedure adding a 2-fold dilution of the 200 mg/mL suspension immediately before the extrusion. This resulted in the sizes of liposomes becoming comparable to pore sizes of 200 nm (Table 2, experiment II).

When determining the efficiency of incorporation of the peptides into liposomes, ultrafiltration was used in the variant of stepwise diafiltration: the dispersions were concentrated approximately two times, diluted to the initial volume, concentrated again, and then the cycle was repeated twice. According to measurements of the concentration of nonincorporated peptides in filtrates using the modified Lowry procedure [28] or by optical density at the absorption peak of peptide mixtures at 273 nm, the efficiency of liposome loading was 40 to 60% of the initial amount of peptides (Table 2). Notably, the calculations were carried out without taking into account peptides 6 and 11, which make up ~1/6 of the peptide compositions, but most likely remained embedded in the liposome membrane. Thus, the loading efficiency of the L₁–L₄ liposomes should have been about 16% higher.

The low load during passive encapsulation of nanosized liposomes into the inner water volume is typical for solutions of any substances in the absence of electrostatic or any other specific interactions with the lipid matrix. For example, in [29–31], the efficiency of peptide incorporation into liposomes was 25–40%. Loading capacity limitations are associated with the small inner volume of liposomes and concentration limit of the colloid, on the one hand, and the low solubility of peptides in the aqueous phase, on the other hand. The authors of works [32, 33] registered incorporation rate of peptides exceeding 80% for liposomes about 200 nm in size judging by the fluorescence of a BODIPY-labeled peptide after dialysis of the liposomal dispersion. However, according to another work [17], various peptides were incorporated into cationic liposomes somewhat smaller than 200 nm in size with the efficiency of as little as 30–40% (despite the fact that electrostatic interactions should promote loading of corresponding peptides), as shown by HPLC analysis of peptides extracted from the liposomes. Moreover, the incorporation efficiency of fluorescently labeled analogs of the same peptides varied in the range of 9–50% (after ultrafiltration in concentrators similarly to the procedure we use here) [17]. In the work [18], incorporation efficiency of the peptides of different polarity and size, from nonameric to 34-mers, into liposomes made of soy lecithin and cholesterol with a diameter of 130–140 nm ranged from 54 to 92%, according to flash chromatography (nonincorporated peptides in the eluate were determined by absorbance at 223 nm). Judging by the published data, different methods for analyzing the efficiency of liposome loading with peptides can give conflicting results. When subunit vaccines are produced, they are usually not separated from nonincorporated antigens (proteins, peptides) in order to avoid loss of the target material due to sorption during purification by gel filtration or ultrafiltration/diafiltration and further need to concentrate the dispersion [18, 29–33]. Technological in terms of losses, tangential filtration is hardly applicable when working with analytical amounts of solutions, as in our case [34]. Therefore, we did not separate nonincorporated peptides from liposomes in order to obtain a significant biological effect from encapsulation in liposomes of any of the peptide compositions.

Evaluation of the efficacy of peptide compositions and liposomal formulations. Since the selection of peptides was based on known human T-cell epitopes, it was necessary to evaluate the possibility of their recognition by mouse T-cells. For this purpose, splenocytes from intact mice were incubated with peptide compositions P₁–P₄ for 7 days. The spleen cells contain approximately 20–25% CD4⁺ T cells, 10–13% CD8⁺ T cells, 35–40% B cells, 4–7% natural killer cells, and 5–8% macrophages. Macrophages are APCs and can present antigens to T cells. There are few specific clones among T cells. Therefore, the analysis was performed on one million splenocytes in each well in 24 replicates, which allows the estimation of the frequency of specific clones among 24 million cells. Figures 1a–1d show data on the recognition of P₁–P₄ peptide compositions by splenocytes of intact mice. A similar number of cells without antigens was used as a control. When adjusted for the proliferation of splenocytes in the absence of peptides, 15, 5, 13, and 13 positive clones were identified in groups P₁–P₄, respectively (Figs. 1a–1d). These data suggest that some of the selected peptides can be recognized by mouse T cells. Accordingly, we used the resulting formulations to immunize mice.

Fig. 1.

Characteristics of the immune response to compositions of P₁–P₄ peptides and their L₁–L₄ liposomal formulations. (a–d) Spontaneous proliferation (MTT) of splenocytes (10⁶/well) of intact mice in response to compositions P₁ (a), P₂ (b), P₃ (c), and P₄ (d); mean + standard deviation of splenocyte proliferation without the addition of antigens is marked with a line; n = number of positive clones. (e) In vitro production of IFN-γ by splenocytes of mice immunized with various drugs in response to specific antigens P₁–P₄. Significant differences (p < 0.05, Mann–Whitney) are marked in parentheses.

To assess the immunogenicity of peptide compositions P₁–P₄ and liposomal formulations L₁–L₄, the preparations were injected into the pad of the hind paw of C57BL/6 mice twice with an interval of three weeks. One week after the last immunization, spleens of the animals were taken and splenocytes were cultivated with the corresponding peptide compositions (P₁–P₄) for 24 h. In Fig. 1e, L_C corresponds to the control group, which was administered with liposomes that did not carry any peptides, but contained CpG-ODN. Since such liposomes can stimulate innate immunity, the analysis of IFN-γ production was carried out during cultivation with all pools of P₁–P₄ peptides. Indeed, in response to immunization with L_C, splenocytes responded with the production of IFN-γ, but there was no specific response to the peptides. Additionally, the study included a group immunized with a sucrose-based buffer, which was used to formulate liposomes and peptide solutions (Fig. 1e). In this group, IFN-γ production was significantly lower (data not shown). In groups of mice immunized with peptide compositions and liposomal formulations with peptides, antigen-specific production of IFN-γ was observed upon immunization with peptide compositions P₂, P₃, P₄, and all liposomal formulations. Maximal responses were observed in groups P₂, P₃, L₁, and L₄. However, in the L₄ group, there was significant variability in the response to the buffer, which resulted in no significant differences with the response to the peptides. In general, the results obtained indicate the formation of specific cellular immunity in response to vaccination with liposomal formulations L₁–L₄, as well as peptide compositions P₂–P₄ (Fig. 1e)

As demonstrated previously, subcutaneous vaccination of transgenic mice against the SARS-CoV coronavirus using liposomes with covalently linked peptides—HLA-A*0201-restricted nonameric CTL epitopes of the virus N-protein [15] and polyprotein 1a (pp1a) [16]—induced the formation of antigen-specific CD8⁺ T cells. Also, when transgenic mice were vaccinated using liposomes with encapsulated nonameric T epitopes of the hepatitis C virus and CpG-ODN, the authors observed about 2000 pg/mL IFN-γ production in the ELISPOT test [33]. In the study of a candidate peptide vaccine against influenza type A in pigs [18], the level of IFN-γ produced by the fraction of peripheral blood mononuclear cells reached 1000–1500 pg/mL (in the ELISA test). In this regard, the results obtained for the L₁ formulation can be considered comparable and promising for immunization in terms of inducing T-cell response.

For further work, the L₁ formulation was selected, for which a significant difference was found between the control and the in vitro response to the antigen. In the next cycle of experiments, peptides from compositions P₂ and P₃ were selected and added to L₁ to prepare a new formulation, L₅.

When compiling a new peptide composition P₅, the following changes were introduced to P₁: instead of peptide 11, peptides 8 and 12 (components of compositions P₂ and P₃ in the first cycle of experiments, Fig. 1e) were introduced. Thus, in the L₅ liposomal formulation, no peptides were included in the lipid bilayer, and the liposome loading efficiency was 52.8% (Table 2).

Peptide recognition analysis by intact splenocytes was performed to formulation P₅, as well as P₁, as a control. The results confirmed previously obtained data on the compatibility of some of the selected peptides with MHC molecules in mice (Figs. 2a, 2d).

Fig. 2.

Characteristics of the immune response to the compositions of peptides P₁ and P₅ and liposomal formulations L₁ (a–c) and L₅ (d–f). Spontaneous proliferation of splenocytes (10⁶/well) of intact mice in response to compositions P₁ (a) and P₅ (d); mean + standard deviation of the splenocyte proliferation value without antigens is marked with a line; positive clones are marked in parentheses. Concentration of IL-6 and TNF-α in sera of mice immunized with L_C, L₁, or P₁ (b) and L_C, L₅, or P₅ (e). Proliferation of splenocytes obtained from mice immunized with L_C, L₁, or P₁ (c) and L_C, L₅, or P₅ (f) in response to individual peptides. Significant differences (p < 0.05, Mann–Whitney) are marked in parentheses.

The resulting formulations were then used to immunize mice. Sera were collected one week after the second immunization and the level of cytokines in the blood was assessed using multiplex analysis, including 10 cytokines (IFN-γ, interleukins (IL) 2, 4, 6, 9, 10, 13, 17, 22, and tumor necrosis factor-alpha (TNF-α)). Among the 10 cytokines, only three were registered in serum: IFN-γ, IL-6, and TNF-α. Blood levels of IFN-γ were low (not shown). Immunization with P₁ and L₁ resulted in an increase in both IL-6 and TNF-α (Fig. 2b) compared with the blood levels in intact mice and the control group L_C. Immunization with L₅ also caused an increase in the levels of IL-6 and TNF-α (Fig. 2e), but to a lesser extent than in the case of L₁.

Splenocytes of immune mice were stimulated in vitro with individual peptides. Peptides 1, 2, 3 and 5, common to all preparations, were recognized by splenocytes after immunization with liposomes L₁ (Fig. 2b), L₅ and a composition of peptides P₅ (Fig. 2e). After immunization with the L₅ liposomes, peptides 8, 12 and 14 were also recognized. In general, when immunized with liposomal formulations, the T-cell response was formed more efficiently than when immunized with peptide compositions.

The level of cytokines in the supernatants of splenocytes of immune mice upon stimulation with individual peptides was analyzed using the multiplex method. Similar to blood serum, out of 10 factors, only three were identified: IFN-γ, IL-6, and TNF-α. The overall level of IFN-γ was significantly lower than in the first experiment, which can be explained by the difference in the methods of determination. In the first experiment, an enzyme immunoassay system was used to determine the concentration of IFN-γ; in the second experiment, multiplex cytometric analysis was used. IL-6 production did not correlate with antigen-specific response (data not shown). The production of IFN-γ and TNF-α is shown in Fig. 3.

Fig. 3.

In vitro production of IFN-γ (a, b) and TNF-α (c, d) by splenocytes of mice immunized with control L_C liposomes, P₁ and P₅ peptide compositions, and L₁ and L₅ liposomal formulations in response to specific peptides. Significant differences (p < 0.05, Mann–Whitney) are marked in parentheses.

In the group of mice immunized with L₁ liposomes, T cells produced IFN-γ and TNF-α in response to peptides 1, 2, 5, and 14 and in the L₅ group, to all peptides from the L₅ formulation except for peptide 2 (Fig. 3).

The highly hydrophobic peptide 11 from the nonstructural protein Orf1b (Table 1) was selected based on the results of clinical studies, although the frequency of immune responses in the multiplex test for the identification of antigenic specificity of T-cell receptors was not so high (two responses out of seven) [24, 26]. Judging by the production of IFN-γ by splenocytes of immunized conventional mice, not all compositions with peptide 11 were immunogenic (e.g., L₃, Fig. 1e). In addition, this peptide did not cause a significant increase in splenocyte proliferation compared to the other peptides of the P₁ composition, in contrast to peptide 2 of the viral envelope protein E (Fig. 2c), and was not recognized by T cells in the multiplex analysis (Figs. 3a, 3c). Nevertheless, when peptide 11 was replaced by two hydrophilic peptides (peptide 8 from M-protein and peptide 12 from nonstructural protein Orf1b), the level of IL-6 and TNF-α cytokines in the sera of immunized mice decreased (compare Figs. 2b, 2e). Therefore, the L₁ liposomal formulation was chosen to study the protective efficacy in the hACE2 mouse infection model.

Assessment of the protective efficacy of formulation L₁ in an infectious mouse model. To evaluate the protective effect of the candidate vaccine, we used genetically modified mice of the C57BL/6JTgTn (CAG-human ACE2-IRES-Luciferase-WPRE-polyA) line (abbreviated, hACE2) carrying a humanized receptor of angiotensin-converting enzyme-2 (ACE2), which ensures the penetration of SARS-CoV-2 coronavirus, as well as NL63 and SARS-CoV viruses, into the cell. This mouse model can be used in studies of the protective efficacy of drugs intended for the prevention and treatment of coronavirus infection COVID-19 [35].

In the present work, male hACE2 mice were immunized with L₁ liposomes (n = 6) and control L_C liposomes (n = 4) twice with an interval of 21 days. Two weeks after the second immunization, the mice were challenged with a lethal dose (3 log PFU) of SARS-CoV-2 (Wuhan strain).

The mean time to death in the groups did not differ. After infection with the virus, all animals died within 4–5 days (the control group was immunized with buffer, n = 3). The lack of differences in lifespan in male mice of different groups is probably due to the initially high dose of infection with SARS-CoV-2.

Lungs were harvested from all mice and histological analysis was performed. The results of a pathomorphological study of the lungs of infected mice are presented below. In the control group without treatment, in the lungs of hACE2 mice infected with the Wuhan strain of SARS-CoV-2 at a dose of 3 log PFU, a diffuse pronounced plethora of large vessels, as well as vessels of the microvasculature, was noted with symptoms of stasis and erythrocyte sludge (Figs. 4a, 4b). Typically, in the walls and lumen of the alveoli, moderate diffuse mononuclear infiltration with an admixture of a few neutrophils was noted. In the lumen of individual alveoli in different parts of the organ, hyaline-like membranes and transudate were found. In the lumen of the majority of the bronchi, attention was drawn to the pronounced desquamation of the respiratory epithelium with the initial destruction of the integrity of the bronchial wall.

Fig. 4.

Histology of the lungs of hACE2 transgenic mice infected with the Wuhan SARS-CoV-2 strain at a dose of 3 log PFU without treatment (a, b) and against the background of immunization with control L_C liposomes (c, d) and formulation L₁ (e, f). Diffuse plethora of large vessels with symptoms of stasis and erythrocyte sludge is marked by arrows; desquamation of the respiratory epithelium with the initial destruction of the integrity of the bronchial wall is marked with asterisks. The scale corresponds to 200 (a, c, e) and 50 (b, d, f) µm. Staining with hematoxylin and eosin.

In animals immunized with the L_C liposomes, no transudate was found in the pulmonary acini in any of lung lobes (Figs. 4c, 4d). In addition, three out of four animals did not have pronounced microcirculatory disorders and plethora of large vessels of the lungs. In two out of four male mice of this group, a decrease in the degree of alteration of the respiratory epithelium of the bronchi was noted. Hyaline-like membranes in this group of animals occurred in three out of four cases with a mean score of 1.25 versus 2.00 in untreated male mice. At the same time, the severity of mononuclear and neutrophilic infiltration of the walls and lumen of the alveoli was comparable to that among animals that did not receive treatment and preliminary immunization.

Upon immunization of animals with the L₁ preparation, no pronounced microcirculatory disorders in the lungs were detected in 2/3 of the cases; the average score for this characteristic corresponded to 3.17 versus 4.00 in animals without therapy (Figs. 4d, 4e). Hyaline-like membranes were also less common among animals treated with L₁ (1.33 vs. 2.00, respectively). However, according to other characteristics, no differences from the “background” course of the disease were observed.

Thus, the administration of liposomes, both control and peptide-containing, improved the state of lung function, although it did not prevent the death of the animals, which is associated with the use of the lethal dose of the virus. A certain protective effect of liposomes, as particles comparable in size to viruses, is apparently associated with the activation of the innate immune system. Since the main task was the induction of adaptive immunity with the help of specific peptides, the lack of the effect is associated with a rapidly occurring infectious process. The lower effect of the peptide formulation of liposomes compared to control liposomes may be associated with the activation of adaptive immunity, which consumes the body’s energy resources. To identify the role of adaptive immunity, sublethal doses of the virus should be used in the future.

Liposomal formulations of long-term storage. Formulations of L₁ and L₅ were investigated in terms of the possibility to obtain preparations with long shelf-life. Liposomal dispersions were lyophilized and then rehydrated with an appropriate volume of water. Data on the size of liposomes before lyophilization and after reconstitution, as well as on the content of peptides therein, are presented in Table 3.

Table 3.

Characteristics of liposome samples prior to and after lyophilization

Sample	Prior to lyophilization			After lyophilization and rehydration
	Diameter (nm) ±SD1	PDI ± SD1	Liposome incorporation, %2	Diameter (nm) ± SD1	PDI ± SD1	Liposome incorporation, %2
L C	211.3 ± 2.1	0.103 ± 0.022	–	187.9 ± 0.6	0.093 ± 0.028	–
L 1	220.6 ± 1.7	0.114 ± 0.029	33.53	188.9 ± 1.7	0.099 ± 0.028	49.23
L 5	212.0 ± 1.3	0.088 ± 0.026	52.8	185.3 ± 1.5	0.070 ± 0.022	51.9

¹ According to measurements on a Brookhaven 90PLUS Particle Size Analyzer (Brookhaven Instruments Corp., United States). ² Calculated as 100% – (peptides outside liposomes, %). According to the measurement of optical density at 273 nm in filtrates after    ultrafiltration. ³ Calculated without taking into account peptide 11 embedded in the liposome membrane.

Neither of the liposome formulations underwent significant changes of size after lyophilization. The content of encapsulated peptides in sample L₁ increased significantly. We hypothesized that such a change may be associated with the disruption of the integrity of the liposome membrane during lyophilization/rehydration followed by redistribution of peptides between liposomes and the solution of unencapsulated peptides. The structure of liposomes was analyzed using cryogenic transmission electron microscopy (Fig. 5).

Fig. 5.

Cryoelectronic micrographs of control L_C liposomes and liposomal formulations of peptides L₁ and L₅ before (a) and after lyophilization and rehydration (b). Bar 50 nm.

Indeed, in sample L₁, during lyophilization/rehydration, small particles, most likely, precipitates of hydrophobic peptide 11, are segregated from the liposomes (Fig. 5b, center). Possibly, this peptide was initially present in the lipid bilayer in the form of aggregates, which precipitated as rather large formations observed in the images obtained with an electron microscope. For the L₅ liposomes, which contain only water-soluble peptides, separate nonvesicular particles were not observed. That is, lyophilization in this case is a suitable method for long-term storage of vaccine constructs.

EXPERIMENTAL

Materials and reagents. The peptides were obtained by solid-phase synthesis using the Fmoc/tert-butyl strategy on a trityl chloride–polystyrene polymer as described previously [36]. Oligonucleotide CpG-ODN 1826 (TCCATGACGTTTCCTGACGTT), specific for TLR9 mice, was kindly provided by Dr. V.A. Gushchin (N.F. Gamaleya National Research Center for Epidemiology and Microbiology). Phosphatidylcholine from egg yolk (ePC, Lipoid E PC S) and cholesterol (Chol) produced by Lipoid GmbH (Heidelberg, Germany), USP qualification (United States Pharmacopeia), were used; sucrose and ethylenediaminetetraacetic acid (EDTA) were from Panreac (USP, Spain); Sepharose CL-4B (Pharmacia, United States); Na₂HPO₄, NaH₂PO₄, and KH₂PO₄ ACS qualification (Helikon, Russia); other reagents manufactured by Sigma and Flow Laboratories (United States) were used. Solvents were purified by standard methods. Evaporation was carried out under vacuum at temperatures not exceeding 40°C.

Preparation of the liposomal formulations of peptides. Individual peptides in the form of trifluoroacetates were dissolved (except for peptides 6 and 11) in phosphate buffer with isotonic sucrose solution PB-Suc, pH 7.2 (6.25 mM Na₂HPO₄, 1.3 mM NaH₂PO₄, 1.2 mM KH₂PO₄, 1 mM EDTA, 240 mM sucrose, H₂O_dd). Then, solutions of mixtures of peptides were prepared, where the final concentration of each of the peptides was 1 mM (see Table 2 for compositions) and, if necessary, titrated with 1 N NaOH to pH ~6.3–7.0. Solutions were frozen in liquid nitrogen (–196°C) and stored at –20°C until use.

Egg phosphatidylcholine and cholesterol were dissolved in tert-butanol at a molar ratio of 67 : 33. In the case of formulations L₁, L₂, L₃, and L₄, 0.5 μmol of peptides 11, 6, 11, and 11, respectively, were added to the tert-butanol solution of lipids. The solutions were frozen and lyophilized for 12 h at a pressure of ~3 Pa (INEY 4 freeze dryer; UPS RAS, Russia). Lyophilized lipid–peptide mixtures were hydrated with solutions of peptide compositions for 2 h with occasional shaking. After hydration, the suspension of liposomes, 200 mg/mL lipids, was subjected to 10 cycles of freezing (–196°C)–thawing (water bath, +40°C)–shaking (Vortex FV-2400, Biosan, Latvia), diluted two times with PB-Suc buffer, and extruded 10 times sequentially through Whatman Nuclepore polycarbonate membrane filters (Cytiva, United States) with pore sizes of 400 and 200 nm on a Mini-extruder (Avanti Polar Lipids, United States). The concentration of ePC in liposomal dispersions was determined using an enzymatic colorimetric method (Phospholipids kit, Sentinel Diagnostics, Italy): 3 µL of sample and 150 µL of the working solution (phospholipase D, >1500 U/L; choline oxidase, >7500 U/L; 4 -aminoantipyrine, 1.2 mM; peroxidase, >7000 U/L; TES buffer, 50 mM, pH 7.6; hydroxybenzoic acid, 12 mM; EDTA, 1.3 mM; sodium azide, <0.1%) were added to the well of a 96-well plate, incubated at 37°C for 10 min, and optical density at 540 nm was measured using a Multiscan FC microplate photometer (ThermoFisher Scientific, United States). The amount of ePC in the samples was determined from the calibration curve for dispersions of ePC in PB-Suc. To obtain control liposomes L_C, a lyophilized mixture of ePC–Chol (67 : 33, mol) was hydrated with PB-Suc buffer and extruded as described above.

The size of liposomes was determined by dynamic laser light scattering using a Brookhaven 90PLUS Particle Size Analyzer (Brookhaven Instruments Corp., United States). At least three measurements of diluted liposome dispersions (50 µg lipids/mL PBS) were performed using a He–Ne laser, λ 633 nm, at an angle of 90°, three cycles of 1 min. Liposomal formulations at the concentrations used for vaccinations (approximately 40 mg/mL in terms of total lipids) were stable for at least 3 weeks at 4–8°C.

Determination of the amount of unencapsulated peptides. Samples of liposomal formulations L₁ and L₅ (500 μL, ~8.5 mg/mL of lipids) were diluted 2-fold with PB-Suc buffer and placed into Vivaspin 2 concentrators (300  000 MWCO, Sartorius, Germany) prewashed with H₂O_dd water and centrifuged for 35 min at 2300 rpm, ~1000 g (CM-6M, ELMI, Latvia). Then, the optical density was measured in the eluted volume of the liquid (~400–500 μL) at 273 nm (SF-2000, OKB-Spektr, Russian Federation). The described steps of washing and measurements were carried out three times. The molar extinction coefficients of the peptide mixtures calculated based on Tyr absorbance were 1536 and 1646 M^–1 cm^–1 for compositions P₁ (without the hydrophobic peptide 11) and P₅, respectively. The amount of free peptides in three filtrates was summed up.

When determining the amount of peptides in washings according to Lowry [28], in samples L₁–L₄, solutions of mixtures of the corresponding peptides (P₁–P₄) were prepared to construct a calibration curve. When standard solutions of albumin were used for this purpose, overestimated (up to three times) values of peptide concentrations were obtained.

Cryogenic transmission electron microscopy. Samples were prepared using copper support meshes with holes in an amorphous carbon film (Lacey C only, 01895-F/01896-F, Ted Pella, Redding, CA, United States) hydrophilized in a glow discharge using a PELCO easiGlow setup (Ted Pella, Redding, CA, United States) under the following conditions: sample processing time, 25 s; current strength, 0.20 mA; and residual pressure in the chamber, 0.26 mbar; 3 μL of the sample was applied to the grid and, using a Vitrobot Mark IV automated system (Thermo Fisher Scientific, Waltham, MA, United States), the excess solution was removed with filter paper for 2.5 s at chamber humidity of 95–100% and a temperature of 4°C, which was followed by vitrification. The samples were examined using a Titan Krios 60-300 cryogenic transmission electron microscope (Thermo Fisher Scientific, Waltham, MA, United States) equipped with a Falcon II direct electron detection device and a CEOS Image Corrector running the EPU software. Basic data acquisition parameters: accelerating voltage 300 kV, nominal magnification 37 000, exposure time 4 s, and defocusing from –3 to –5 µm.

Immunization of mice. Female mice of the C57BL/6 line weighing 18–20 g were obtained from the Stolbovaya (Moscow oblast) nursery. They were kept under conventional conditions without restriction in water and food. All studies and procedures for the routine care of animals were carried out in accordance with the International Guidelines for Biomedical Research on Animals at the Gamaleya Research Center for Epidemiology and Microbiology, Ministry of Healthcare of the Russian Federation, protocol of the Biomedical Ethics Committee No. 5 dated 19.03.2021, and at the Shemyakin–Ovchinnikov Institute of Bioorganic Chemistry, Russian Academy of Sciences, Protocol of the Institute Committee for the Care and Use of Animals No. 325 dated 24.05.2021.

Formulations L₁–L₅, peptide compositions P₁–P₅, and control PB-Suc solution were injected into the pads of hind paws, 50 μL each, twice with an interval of 21 days. Total content of the peptides in a dose was 120–130 μg. All preparations, except for the PB-Suc control, contained the CpG-ODN immunostimulator, 75 μg. The number of mice in groups was n = 7. A week after the second immunization, blood was taken from the orbital sinus of mice under isoflurane anesthesia into heparinized tubes, plasma was isolated and stored frozen until the analysis. Mice were sacrificed by cervical dislocation, and spleens were collected under sterile conditions.

MTT analysis of splenocyte proliferation. The spleens of intact or immune mice were homogenized in saline and centrifuged at 1000 rpm for 7 min; the sediment was treated with 0.83% ammonium chloride to lyse the red blood cells, washed in saline solution twice by centrifugation, and transferred to a culture medium based on RPMI-1640 with the addition of 7% fetal calf serum and penicillin–streptomycin–glutamine (PanEco, Moscow, Russian Federation). Splenocytes were spread on flat-bottomed 96-well plates in the amount of 1 million/well in 200 μL of the medium. Antigens were added to the wells in the amount of 20–30 µg. The plates were incubated for 72 h. During the last 3 h, 10 μL of MTT (5 mg/mL) was added to each well. After incubation, the culture medium was removed and 100 μL of DMSO was added to each well. The plates were incubated with shaking for 15 min to dissolve the formazan. Optical density was measured on a Titertek spectrophotometer (Great Britain) at 540 nm. The results were analyzed using the Excel package (Microsoft). Data are reported as optical density (OD).

Analysis of spontaneous response to peptides. To assess the recognition of selected peptides by mouse splenocytes, 24 or 48 replicas of splenocytes obtained from one mouse were incubated with antigens (mixtures of peptides P₁–P₅) at a concentration of 160 μg/mL and without antigens (control) as described above.

Analysis of the antigen-specific response after immunization of mice with preparations P₁–P₄ and L₁–L₄ (experiment I). The concentration of IFN-γ produced by splenocytes in response to stimulation with the P₁–P₄ peptide compositions was determined using the interferon-gamma release assay (IGRA) in accordance with the described method [37]. Splenocytes were seeded at a density of 10⁷ cells/mL in 100 µL of RPMI-1640 growth culture medium supplemented with 2 mM L-glutamine, 10% fetal calf serum, 1× antimycotic antibiotic, and 0.05 mM 2-mercaptoethanol (PanEco, Moscow, Russian Federation). Cells were incubated in a CO₂ incubator at 37°C, 5% CO₂, and 100% humidity for 1.5–2 h. Antigen mixtures P₁–P₄ were added to the wells at a concentration of 10 μg/mL, concanavalin A (10 μg/mL) was used as a positive control, and PB-Suc, as a negative control. Each experimental group was stimulated with the relevant peptide composition, the PB-Suc and L_C control groups were stimulated with each of the peptide compositions. Each group was also challenged with positive and negative controls. Stimulation was carried out in two repeats.

The culture plates were incubated in a CO₂ incubator for 20 h, the culture medium was taken, and the concentration of IFN-γ (pg/mL) was determined therein using a commercial IFN gamma Mouse ELISA Kit (Thermo Fisher, United States) according to the manufacturer’s instructions.

Analysis of the antigen-specific response to peptides (experiment II). To assess the response to peptides by proliferation and production of cytokines, splenocytes of immunized mice of each group (n = 7) were stimulated with peptides comprising corresponding formulation at a dose of 25–45 μg. Experiments were performed in triplicate for each peptide. After 24 h, 50 µL were taken from the wells, transferred to new plates, the plates were stored at –60°C until the analysis. Splenocytes from intact mice, mice immunized with PB-Suc buffer, or control L_C liposomes were used as controls. Proliferation analysis was performed 72 h later as described above.

Complex analysis of cytokines (experiment II). A standard panel of magnetic beads for the analysis of mouse cytokines IFN-γ, IL 2, 4, 6, 9, 10, 13, 17, 22, and TNF-α (Biolegend, San Diego, CA, United States) was used to analyze proteins in plasma blood and splenocyte supernatants according to the manufacturer’s protocol using a MACSQuant Tyro Sorter (Miltenei, Germany) flow cytometer.

Statistical analysis was performed using the Excel software and Student’s t-test. Differences at p < 0.05 were considered statistically significant.

Verification of the protective efficacy of the L₁ formulation. We used male humanized mice C57BL/6-TgTn(CAG-humanACE2-IRES-Luciferase-WPRE-polyA) aged 14–16 weeks. The animals were kept under standard conditions at the Laboratory Animal Nursery of the Institute of Bioorganic Chemistry, Russian Academy of Sciences (Unique scientific facility, Bio-Model, Institute of Bioorganic Chemistry, Russian Academy of Sciences), which has the AAALACi international accreditation. All experiments and manipulations were approved by the Institute Committee for the Care and Use of Animals (No. 757/22 dated 17.02.22). All animals were checked for the presence of the target hACE2 gene, the expression of which was analyzed by PCR-RT. Animals were immunized twice (on days 1 and 21) by injection of 100 µL (2 × 50) of PB-Suc buffer (n = 3), control L_C liposomes (n = 4), and L₁ preparation supplemented with 75 µg CpG-ODN (n = 6). According to the dynamic light scattering data, the size of the L_C liposomes was 207.9 ± 5.7 nm (PDI 0.096 ± 0.042), L₁ liposomes were 207.5 ± 1.8 nm in diameter (PDI 0.093 ± 0.021).

Infection of mice with the SARS-CoV-2 virus. Thirty-six days after the first immunization, all C57BL/6-TgTn(CAG-humanACE2-IRES-Luciferase-WPRE-polyA) mice were transferred to a specialized ABSL-3 level laboratory (Sergiev Posad), where they were infected with the Wuhan SARS-CoV-2 strain by intranasal administration of the SARS-CoV-2 virus at a dose of 3 log PFU in a volume of 20 µL of saline [38]. In mice that died during the observation period, the cardiopulmonary complex was extirpated and, after two weeks of inactivation in 10% formalin solution, was sent for histological examination.

Histology. Lungs fixed in 10% neutral formalin solution were washed in tap water, dehydrated in ascending concentrations of ethyl alcohol, and embedded in paraffin. Paraffin sections 4–5 μm thick, stained with hematoxylin and eosin, were examined using the AxioScope.A1 (Carl Zeiss, Germany) conventional light microscopy. Microphotographs of the histological preparations were obtained using a high-resolution Axiocam 305 (Carl Zeiss, Germany) color camera and ZEN 2.6 lite (Carl Zeiss, Germany) software. Evaluation of the severity of certain pathological signs was carried out on a 5-point scale, where 0 indicates no sign of infection (within the normal range); (1) minimal severity; (2) weak; (3) medium (moderate); (4) pronounced; and (5) severe [39].

CONCLUSIONS

Subcutaneous vaccination with natural lipid liposomes was shown to improve lung function, apparently due to activation of innate immunity, but failed to protect against a lethal dose of the virus. Liposomes carrying a set of T-cell epitopes of the SARS-CoV-2 virus can serve as a vaccine prototype, at least for the prevention of a chronic course and moderate severity of an infectious disease. When choosing epitopes of cytotoxic and helper T lymphocytes for the preparation of peptide compositions, the results of clinical analyzes of convalescent patients should be taken as a basis, and, secondarily, immunoinformatics data should be taken into account. Immunodominant epitopes calculated in silico often occupy the first positions simply because of their high hydrophobicity, while their presentation by the MHCI or MHCII complexes can be hindered. In addition, hydrophobic peptides must be included in the lipid bilayer, which creates certain difficulties in the formation of liposomes, and can also lead to the destruction of the latter during lyophilization due to the disruption of the liquid crystal structure of the membrane. Optimization of the method to obtain liposomes with T-cell epitopes suitable for long-term storage will be the subject of our further research.

FUNDING

The study was supported by the Russian Foundation for Basic Research (project no. 20-04-60478).

COMPLIANCE WITH ETHICAL STANDARDS

Conflict of interest. The authors declare that they have no conflicts of interest.

Statement on the welfare of animals. The conduct of this study on laboratory animals was approved by the Biomedical Ethics Committee of the Gamaleya Research Center for Epidemiology and Microbiology, Ministry of Healthcare of the Russian Federation, and the Institute Committee for the Care and Use of Animals of Shemyakin–Ovchinnikov Institute of Bioorganic Chemistry, Russian Academy of Sciences.