The choice of the antigen, the host, and the production platform as well as the adjuvants used for formulation is crucial for the design of vaccines. The recent health emergency caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) causing coronavirus disease 2019 highlights the importance of vaccine development for facing a worldwide pandemic.

The SARS-CoV-2 genome encodes several structural, non-structural, and accessory proteins, being the first ones called spike (S), envelope (E), membrane (M), and nucleocapsid (N). S constitutes a 1273 amino acid glycoprotein which is heavily glycosylated and assembles into trimmers adopting a crown-like appearance. This glycoprotein binds to cellular angiotensin-converting enzyme 2 (ACE2) which acts as the cellular receptor mediating the virus entry into the cell (Arashkia et al. 2021). Since S is exposed at the virus surface, being the primary target of neutralizing antibodies, it represents an ideal candidate for the design and production of safe COVID-19 vaccines (Pino et al. 2020; Arashkia et al. 2021). It is important to consider that vaccines should elicit balanced humoral and Th1 cellular immune responses to protect against SARS-CoV-2 (Poland et al. 2020).

The choice of the bioproduction host is considered a crucial step for recombinant cell line development. Mammalian cell lines are the host of choice for the production of secreted recombinant proteins. Even though Chinese Hamster Ovary (CHO) cells are the most commonly used for the production of biopharmaceuticals, Human Embryonic Kidney (HEK) cells represent a good alternative (Pulix et al. 2021; Malm et al. 2022). Differences in the physicochemical, glycosidic, and biological properties of the proteins produced in CHO versus HEK cells have been described (Croset et al. 2012; Gugliotta et al. 2017). However, the human-derived cell line has been demonstrated to be more effective for the expression and secretion of difficult-to-express (DTE) proteins. Also, HEK cells are able to introduce human-like post-translational modifications (Malm et al. 2022).

It is known that vaccine efficacy depends not only on the antigen’s characteristics but also on the selection of the adjuvant, as it modulates the immune response. Choosing the proper adjuvant can reduce antigen concentration and the number of immunizations required for protective efficacy, therefore contributing to making vaccines cost-effective (Bonam et al. 2017). Aluminum hydroxide gel adjuvant, usually referred to as Alum (AH), is one of the first adjuvants approved by the FDA and has been used in clinical vaccines for almost a century. It usually induces classic humoral-mediated response (mainly Th2 participation) instead of cell-mediated (mainly Th1 participation) immunity, so its effect is limited (Tan et al. 2022). Alternatively, immune-stimulating complexes (ISCOMs) are cage-like structures formed spontaneously by mixing specific saponins (mostly Quil A) with cholesterol and phospholipids at a specific ratio. ISCOMs can achieve efficient antigen delivery into antigen-presenting cells, therefore inducing antigen-specific T-cell responses, long-lasting antibody responses, and balanced Th1/Th2 immunity (Bonam et al. 2017). Additionally, the stability demonstrated by ISCOMs formulations allows for long-term storage, representing an advantage for vaccine development and production (Bertona et al. 2017).

Until now, a wide variety of vaccine platforms based on inactivated viruses, protein subunits, mRNA, viral vectors, and virus-like particles (VLPs) have been developed and studied since the beginning of the COVID-19 pandemic (Kyriakidis et al. 2021).

The aim of this work was to produce, purify, and characterize a new potential recombinant protein subunit COVID-19 vaccine based on S glycoprotein in HEK cells. Two antigen formulations were prepared using ISCOMs and Alhydrogel as adjuvants, and, following mice vaccination, humoral and cellular immune responses as well as the presence of neutralizing antibodies were assessed.

Materials and methods

Sequence design and animal cell expression plasmids construction

The S ectodomain (S-ED) amino acid (aa) sequence was designed according to the information reported in the bibliography for the first virus isolate, Wuhan-Hu-1 (Pallesen et al. 2017; Kirchdoerfer et al. 2018; Wu et al. 2020; Walls et al. 2020; Amanat et al. 2020). The native signal peptide of the S protein (aa 1 to 14) was included in the construction. A bacteriophage T4 fibritin trimerization domain (FOLDON) was inserted immediately after the C-terminal region of S-ED to induce stable trimerization. A 6x-His tag was added after the C-terminal region of the FOLDON sequence to facilitate the downstream processing of the protein. A furin cleavage site between S1 and S2 domains (aa 685 and 686) was mutated to stabilize the structure. Also, two consecutive Pro were included in S2 (Lys983 and Val984 were replaced by Pro) to stabilize the native pre-fusion conformation. The S-ED sequence (accession number MT380725.1) was obtained by chemical synthesis (Gene Universal Inc., Newark DE, USA) and cloned in a plasmid optimized for mammalian cell expression, containing a CAG promoter sequence (composed of the CMV enhancer followed by the chicken β-actin promoter) and a chimeric intron. Figure 1 shows a schematic illustration of the sequence. Bacterial clones were checked by restriction analysis.

Fig. 1
figure 1

Schematic representation of S proteins. A Wild type S protein. B Soluble S-ED stabilized in its pre-fusion conformation through furin cleavage site (aa 685–686) and two consecutive Pro (K983P and V984P) mutations. sp, signal peptide (aa 1–14); RBD, receptor binding domain (aa 329–522); S1, S1 domain; S2, S2 domain; TM, transmembrane domain; CT, cytoplasmic domain; FOLDON, T4 fibritin trimerization domain; 6x-His tag, hexahistidine tag

The expression plasmid was amplified and purified from competent E. coli cells. Sequences were verified by dideoxy DNA sequencing (Macrogen, Seoul, South Korea).

Cells and virus

Suspension growth adapted HEK293 cells (sHEK) were established in our laboratory by gradual adaptation of adherent HEK293T/17 cells (ATCC) from adherent growth medium [Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco) containing 10% fetal bovine serum (FBS, Gibco) and 2 mM glutamine (DMEM10)] to suspension growth medium [EX-CELL®293 serum-free medium (SAFC Bioscience) supplemented with 6 mM glutamine (Gibco)]. Production in 1-L bioreactors (BIOSTAT® Qplus, Sartorius) was performed using production medium (CD BHK-21 Production Medium, Thermo Fisher).

Suspension cells were cultured in Erlenmeyer flasks (Corning, USA) agitated at 140 rpm using a shaking incubator cabinet (CERTOMAT®CT plus, Sartorius). Adherent cells were cultured in T-flasks and multiwell plates (Greiner). Suspension and adherent cells were maintained at 37 °C with 5% CO2.

Vero E6 cells (ATCC CRL-1586) were cultured in DMEM10 containing 100 IU.mL−1 penicillin and 100 μg.mL−1 streptomycin (Gibco). HEK-hACE2 cells were cultured in DMEM10.

SARS-CoV-2 WT (B.1, D614G) was isolated from nasopharyngeal specimens at the laboratory and adapted to grow in Vero E6 cultures.


Rabbit polyclonal anti-S-ED serum and biotinylated rabbit polyclonal anti-S-ED antibodies used in this work were produced and characterized in our laboratory by immunization of rabbits with highly purified S-ED molecule, using standard protocols (Harlow and Lane 1988).

Cell line generation

sHEK cells were cultured in adherent medium for 48 h before starting the transfection protocol. Once they adopted the adherent condition, 2.105 cells.mL−1 were seeded in 6-well plates and incubated at 37 °C and 5% CO2. After 24 h, transfection complexes were prepared and incubated with the cells. For the generation of the transfection complexes, LipofectamineTM 3000 Transfection Reagent (Invitrogen) and circular plasmids were mixed following the manufacturer’s instructions. After transfection, cells were cultured under zeocin selection pressure (200 μg.mL−1) until they reached a monolayer, and no dead cells were observed.

The recombinant cell line was immediately transferred to production medium, and the expression of S-ED was confirmed by SDS-PAGE/western blot.

Scale up and production of S-ED

Cells were initially grown in Erlenmeyer flasks, and then they were transferred to a 1-L bioreactor operated in perfusion mode (Biostat Q Plus, Sartorius). The culture parameters were controlled online, maintaining an agitation rate of 140 rpm, a pH of 7.2, and a pO2 of 30% of air saturation. Temperature varied between 37 and 33 °C. Sample collection was performed daily in order to determine cell density, cell viability, glucose, and lactate levels. Also, S-ED production was monitored by SDS-PAGE/western blot and sandwich ELISA (sELISA).

Purification of SARS-CoV-2 S-ED from culture supernatants

S-ED was purified from clarified culture supernatants by immobilized metal affinity chromatography (IMAC) using highly cross-linked agarose with a covalently-coupled chelating group (IMAC Sepharose 6 Fast flow, Cytiva), which was previously activated with NiCl2. Culture supernatant was conditioned to 20 mM imidazole, 0.5 M NaCl, pH 7.4, and applied to the matrix after equilibration with 8 column volumes (CV) of binding buffer. Following two washes with different imidazole concentrations (20 mM, 8 CV and 50 mM, 10 CV), S-ED was eluted with 10 CV of 250 mM imidazole, 0.5 M NaCl in phosphate buffer pH 7.4. Fractions containing the protein were pooled and diafiltered against phosphate-buffered saline (PBS, 10 mM Na2HPO4, 1.8 mM KH2PO4, 140 mM NaCl, 2.7 mM KCl, pH 7.4).

SARS-CoV-2 S-ED characterization

Protein measurement: absorbance at 280 nm

The concentration of purified S-ED samples was estimated spectrophotometrically using absorbance at 280 nm (149,100 M−1 cm−1).

Protein measurement: sandwich enzyme-linked immunosorbent assay (sELISA)

Ninety-six-well plates were coated with 100 μL per well of 100 ng of rabbit polyclonal anti-S-ED serum in 50 mM carbonate/bicarbonate buffer (pH 9.6) and incubated overnight (ON) at 4 °C. After blocking 1 h at 37 °C with 2% (w/v) non-fat milk in PBS, plates were incubated with 1:2 serial dilutions of S-ED internal standard or samples (culture supernatants, harvests, different purification fractions) for 1 h at 37 °C. At the same time, the following controls were tested in each run: (1) a reactive blank and (2) a sample blank. Then, plates were incubated with biotinylated rabbit anti-S-ED antibodies diluted 1:1500 for 1 h at 37 °C. Afterwards, streptavidin–horseradish peroxidase (HRP) conjugate (Sigma) diluted 1:2.000 was added to the wells. Finally, plates were incubated for 13 min with substrate solution (0.5 mg.mL−1 o-phenylenediamine, 0.5 μl.mL−1 H2O2 30 vol in 50 mM phosphate citrate buffer). Absorbance was measured at 492 nm with a microtiter plate reader (LabSystems Multiskan, Thermo Fisher Scientific). Between every step, plates were washed six times with 0.05% (v/v) Tween-20 in PBS (PBS-T). Dilutions of tested samples and antibodies were prepared in PBS-T containing 0.2% (w/v) non-fat milk.


S-ED electrophoresis profile was analyzed using a 12% (w/v) polyacrylamide resolving gel and 5% (w/v) stacking gel. Separation was performed at 200 V for 65 min. Samples were mixed with 4× sample buffer (2% (w/v) SDS, 10% (v/v) glycerol, 0.05% (v/v) bromophenol blue, 50 mM Tris–HCl pH 6.8), in reducing (5% (v/v) beta-mercaptoethanol) and non-reducing conditions.

Gels were stained with Coomassie blue and destained with a solution containing 15% (v/v) methanol and 10% (v/v) acetic acid.

For western blot analysis, proteins were transferred to polyvinylidene fluoride (PVDF) membranes (Bio-Rad Laboratories Inc.). Membranes were blocked with Tris-buffered saline (TBS, 50 mM Tris–HCl, pH 7.5, 150 mM NaCl), 5% (w/v) non-fat milk, and incubated with an anti-SARS-CoV-2 S neutralizing monoclonal antibody (mAb) (Creative Diagnostics) diluted 1:1000 in TBS containing 0.05% (v/v) Tween 20 (TBS-T), 0.5% (w/v) non-fat milk. HRP-conjugated rabbit anti-mouse (Invitrogen) diluted 1:750 was used as a secondary antibody, and the reaction was visualized using a chemiluminescence reagent (Pierce® ECL Western Blotting Substrate, Thermo Scientific).

The relative quantification of bands was performed using the free-image processing Image J software (Schindelin et al. 2012).

Intrinsic fluorescent spectroscopy

Spectrofluorimetric measurements were performed using a Perkin-Elmer LS-55 luminescence spectrometer equipped with a Xenon discharge lamp, Monk-Gillieson type monochromators, and a gated photomultiplier. The acquisition of data was performed using the FL Winlab v4.00.03 software supplied by Perkin Elmer. Emission spectra were recorded from 290 to 500 nm in 1 nm steps with excitation at 275 nm.

Urea and DTT treatment

Purified S-ED was treated with 6 M urea for 1 h at 37 °C. Besides, its reduced form was prepared by incubating the samples with 100 mM DTT for 1 h at 37 °C. Aliquots of freshly stocked solution of urea and DTT were always used for the treatment.

PNGase F treatment

Five micrograms of purified S-ED were mixed with 75 IU of PNGase F (New England Biolabs, UK) in a total volume of 20 μL in denaturing conditions and incubated overnight at 37 °C, following the enzyme manufacturer’s specifications. Samples were analyzed by SDS-PAGE, using a 10% (w/v) polyacrylamide resolving gel, followed by Coomassie blue staining to verify the removal of glycans of the samples, as described above.

S-ED specific binding with HEK293T cells expressing hACE2

The specific interaction between S-ED and ACE2 receptor was determined using HEK-ACE2 cells. HEK293T cells were used as a negative interaction control.

For flow cytometry analysis, 2 × 105 HEK293T or HEK-ACE2 cells were harvested and washed with PBS. Cells were centrifuged at 100 g for 10 min and resuspended in 200 μL of 100 μg.mL−1 S-ED in PBS. Incubation was performed during 45 min at room temperature (RT). Negative controls for both cell lines were incubated only with PBS. Then, cells were harvested and washed again with PBS. After centrifugation, cells were resuspended in 200 μL of a 1:1000 dilution of the polyclonal antibody anti-S-ED. After a 30-min incubation at RT, cells were washed and harvested. The cell pellet was resuspended using 200 μL of Alexa Fluor 488-conjugated goat anti-rabbit antibody (Thermo Fisher Scientific) and incubated at RT in the dark for 30 min. Finally, cells were washed, resuspended in 200 μL of PBS, and evaluated using a GUAVA EasyCyte cytometer (Millipore). The data was analyzed using Cytoflow software, 1.2 version (©Copyright Massachusetts Institute of Technology).

For fluorescence, microscopy analysis cells were analyzed under adherent conditions, following the protocol described by Rodríguez et al. (2022).

Transmission electron microscopy

S-ED was analyzed by negative staining electron microscopy. Ten microliters of purified S-ED (50 μg.mL−1) was deposited in a formvar/carbon-coated 200-mesh copper grid and incubated for 10 min. Excess was removed with filter paper, and the grid was then negatively stained with 10 μL of 1% (w/v) phosphotungstic acid for 1 h. Sample was examined using a transmission electron microscope (JEM-2100 plus, JEOL) at 100 kV, and images were visualized and analyzed using ImageJ software.

Mice immunization


Eight-week-old BALB\c mice were obtained from Centro de Medicina Comparada (ICIVET-CONICET-UNL, Argentina) and housed in a temperature-controlled room at 23 °C with a 12-h light/dark cycle. Standard food and water were available ad libitum. The immunization protocols were approved and supervised by the Advisory Committee on Ethics and Security of the School of Biochemistry and Biological Sciences, Universidad Nacional del Litoral, according to international guidelines (“Guide for the Care and Use of Laboratory Animals”, Eighth Edition National Research Council 2011).

Adjuvants and formulation

Five micrograms of S-ED were formulated with LipoSap® adjuvant (Lipomize S.R.L., Argentina) at a final concentration of 100 μg.mL−1 of saponin (S-ED/LipoSap). This adjuvant is called ISPA (immune-stimulating particles) and consists of cage-like particles containing Quil-A® as an immune response stimulator (Bertona et al. 2017). Also, the same amount of S-ED was formulated with aluminum hydroxide gel (Aldhydrogel® 2%, Brenntag, Germany) at a final ratio of 10 μg Al+3:1 μg antigen (S-ED/Alum).

Immunization schedule

Mice were inoculated intramuscularly (i.m.) with a first dose of the corresponding formulation at day 0. Group I (n=9) received 100 uL of the S-ED/LipoSap formulation, while group II was inoculated with the same volume of S-ED/Alum preparation. A second dose was applied on day 21. On day 155 (134 days after the second dose), animals received a booster inoculation. Blood samples were collected on days 36, 114, 148, and 171 after the first immunization and preserved at −20 °C until analysis. On day 171, animals were sacrificed, and spleens were obtained.

Determination of antibody levels in plasma

Anti-RBD IgG levels were determined by indirect ELISA (iELISA). Ninety-six-well plates were coated with 150 ng of RBD in 50 mM carbonate/bicarbonate buffer (pH 9.6) overnight at 4 °C. After blocking with PBS, 2% (w/v) non-fat milk for 1 h at 37 °C, plates were incubated with 2-fold serial dilutions of plasma samples for 1 h at 37 °C. Then, plates were incubated with 1:2000 diluted HRP-conjugated rabbit anti-mouse IgG (Invitrogen) for 1 h at 37 °C. Finally, the assay was revealed as described above, and absorbance was measured at 492 nm with a microtiter plate reader. Between every step, plates were washed six times with PBS-T. Dilutions of tested samples and antibodies were prepared in PBS-T containing 0.2% (w/v) non-fat milk. Titer was calculated as the reciprocal of plasma dilution yielding an OD higher than the cut-off value, which was determined as the mean plus 3 SD of negative controls (basal mice plasma).

Ex vivo immune stimulation of spleen-derived MNCs

Sixteen days after the third vaccine dose, mice were euthanized. The abdominal area was sterilized with 70% alcohol. Spleen was removed through an incision. Connective tissues were removed, and the spleen was dissected, placed in a petri dish, and rinsed twice with PBS. For tissue homogenization, the spleen was placed on a sterile 60-mesh stainless steel sieve (pore size 250 μm) onto a collector tube, perfused with culture medium, teased apart using micro-dissecting scissors, and gently pushed through the sieve. The final volume of the splenic suspension was 6 mL. The splenic suspension was layered onto 2 mL Ficoll-Paque, and mononuclear cells (MNCs) were recovered at the interface following centrifugation at 400 × g for 30 min. MNCs were washed twice with PBS. After centrifugation at 200 g for 10 min, the pellet was resuspended in RPMI medium supplemented with 1 mM sodium pyruvate, 2 mM glutamine, and 10% (v/v) FBS. After 24 h, MNCs from each mouse were counted and seeded at a density of 4×106 cells/well in 24-well plates. MNCs were stimulated either with 5 μg of purified S-ED protein, 7 μg of lectins (Phaseolus vulgaris Phytohemagglutinin (PHA-M), Sigma-Aldrich) (positive control), or RPMI medium (negative control). Conditions were assayed in triplicate wells. After 72 h of stimulation, supernatants were harvested and stored at −80 °C until quantification of IL-4 and IFN-γ cytokines by Mouse IL-4 and Interferon gamma ELISA Kits (Abcam).

SARS-CoV-2 neutralization assay

Protocols involving manipulation of SARS-CoV-2 were approved by the Biosecurity Officials and performed under BSL-3 containment at the Universidad Nacional de San Martín according to international guidelines. Vero E6 cells were seeded in 96-well plates at a density of 1.5 × 104 cells per well in DMEM10 and incubated for 24 h at 37 °C and 5% CO2. SARS-CoV-2 WT (B.1, D614G, GISAID Accession ID EPI_ISL_15806335) was preincubated with serially diluted immunized mice sera for 1 h at 37 °C from 1:128 or 1:512. A 300 Median Tissue Culture Infectious Dose (300 TCID50) of the virus was used. Each serum dilution was tested by duplicates. Then, virus-serum mixture was added onto Vero E6 cells in a final volume of 100 μL in DMEM 2% FBS. After 72 h at 37 °C and 5% CO2, cultures were fixed with formaldehyde 10% at 4 °C for 24 h and stained with crystal violet solution in methanol. The cytopathic effect (CPE) of the virus on the cell monolayer was assessed visually, and neutralization titer (NT) was defined as the inverse of the highest serum dilution without any CPE.


Recombinant cell line generation

In the present study, sHEK cell lines producing S-ED were obtained after transfection and zeocin selection (sHEK Tf S-ED). The S-ED expression was confirmed by SDS-PAGE/western blot (Fig. S1).

sHEK Tf S-ED was gradually adapted to production medium. This process took approximately 15 days, until normal daily growth was observed. In order to verify S-ED expression in this adapted cell line, SDS-PAGE/western blot was performed using a neutralizing monoclonal antibody anti-SARS-CoV-2 S protein (data not shown). In this way, the identity of the S-ED protein could be confirmed, since it was recognized by the mentioned antibody. Also, as expected, it presented a molecular mass higher than 150 kDa, similar to previous reports for analogous sequences (Amanat et al. 2020).

S-ED production in 1 L bioreactors

sHEK Tf S-ED cell line was cultured in production medium using 1 L bioreactors operated in perfusion mode during 10 days. The perfusion rate varied between 0.25 and 1 reactor volumes per day, as the process demanded. The culture was inoculated at a cell density of 1 × 106 cells.mL−1 and reached a maximum concentration of 1.3 × 107 cells.mL−1. Glucose and lactate levels remained above 0.4 g.L−1 and under 1 g.L−1, respectively, during the whole process. Parameters such as temperature and perfusion rate were controlled to assure proper conditions for cell growth and viability.

S-ED expression was assessed along the production process. Bioreactor samples from day 3 to 10 were analyzed by SDS-PAGE/western blot (Fig. 2A). A total volume of 10.5 L was harvested.

Fig. 2
figure 2

S-ED production and purification. A Production of S-ED in 1 L bioreactors. Samples from day 3 to 10 (lanes 1–8) were analyzed by SDS-PAGE/western blot using an anti-SARS-CoV-2 S neutralizing mAb for detection. Lane 9, Precision Plus ProteinTM All Blue (Bio-Rad). B S-ED purification by IMAC. Fractions corresponding to different stages of S-ED purification were subjected to electrophoresis under reducing conditions and stained with Coomassie blue. Lane 1, Precision Plus ProteinTM All Blue (Bio-Rad); lane 2, clarified harvest; lane 3, flowthrough (FT); lane 4, first wash; lane 5, second wash; and lanes 6–13, different fractions corresponding to elution step

Soluble SARS-CoV-2 S-ED purification

Clarified crude supernatant containing approximately 2 mg of S-ED was loaded onto an IMAC resin, as S-ED included a C-terminal 6x-His tag. S-ED was efficiently bound to the resin, since no leakage of the protein was observed during the loading and first washing step. A 5% loss was observed during the second wash step; nonetheless, it helped to improve the purity profile. S-ED eluted with a recovery of 95%, and, as judged by SDS-PAGE analysis, it exhibited high purity levels (> 95%) (Fig. 2B).

Soluble SARS-CoV-2 S-ED characterization

As can be seen in Fig. 3A, in non-reducing conditions, the S monomer migrated with an apparent molecular mass higher than 150 kDa, which was also higher than its theoretical weight of 139 kDa calculated from its amino acid sequence. Besides, in both conditions but after PNGase F treatment, a shift in the mobility of the band was observed as a consequence of N-glycan’s removal. However, the resulting molecular mass was still higher than the theoretical one, suggesting the existence of O-glycosylation.

Fig. 3
figure 3

S-ED physicochemical and morphological characterization. A SDS-PAGE analysis of S-ED treated (+) and untreated (−) with PNGase F under denaturing conditions, with (+) and without (−) heat treatment and in the absence (−) and presence (+) of a reducing agent (β-ME). Lane 1 shows the PNGase F band pattern (36 KDa), and lane 2, Precision Plus ProteinTM All Blue (Bio-Rad). B Steady-state fluorescent profile (at 275 nm excitation) of S-ED (■), S-ED treated with 6 M Urea (●), and S-ED treated with 100 mM DTT (▲). C Transmission electron micrograph of negatively stained S-ED. Scale bar: 200 nm

Purified S-ED samples were also evaluated throughout steady-state fluorescence of aromatic amino acids to provide tertiary structural information, since the fluorescence of tryptophan (Trp), mainly, is highly sensitive to its local environment. Untreated S-ED sample exhibited an emission maximum at 342.5 nm. A 6 M urea-treated sample showed a red-shifted fluorescence profile (350 nm) and an increased emission intensity that would reflect the unfolding of the protein. The reduced form exhibited a lower redshift (348 nm) and a decrease in the emission intensity (Fig. 3B). Those changes in the Trp emission profile are also consistent with protein unfolding.

TEM image (Fig. 3C) revealed the presence of structures similar to trimerized S-ED exhibiting the expected size (14 ± 3 nm, n=60).

S-ED binding to the receptor hACE2 present in recombinant HEK293T cells was confirmed by flow cytometry (Fig. 4A) and fluorescence microscopy (Fig. 4B).

Fig. 4
figure 4

ACE2 receptor binding assay. HEK293T-ACE2 cells were incubated with purified S-ED for 45 min at RT. After washing, cells were incubated with rabbit polyclonal anti-S-ED antibodies. After 30 min at RT, cells were washed and incubated with Alexa Fluor 488-conjugated goat anti-rabbit antibodies and incubated at RT in the dark for 30 min. Then, cells were analyzed by flow cytometry (FC). For fluorescence microscopy analysis, nuclei were stained with Hoechst. The same protocol was performed with HEK293T WT cells as negative control. A Flow cytometry analysis. B Fluorescence microscopy. Scale bar: 25 μm

Mice immunization

The immunogenicity of recombinant S-ED formulated with different adjuvants was evaluated in mice, following the schedule described in “Materials and methods” (Fig. 5A). Remarkably, IgG levels in mice inoculated with LipoSap formulation were between 5.9 and 10 times greater than the ones obtained with Alhydrogel formulation (p<0.01) for the different sampling times. Besides, the analysis of the collected plasma samples showed a gradual decrease in anti-RBD IgG levels until day 148 post-prime immunization with both adjuvants (Fig. 5B). Interestingly, after a booster at day 155, IgG titers increased significantly (p<0.01), reaching an 8.3-fold increase for Alum and a 4.2-fold increase for LipoSap (Fig. 5C).

Fig. 5
figure 5

Mice immunization with S-ED formulated with different adjuvants. A Schematic representation of immunization schedule and splenocytes proliferation assay. Five micrograms per dose of S-ED formulated with Alum or LipoSap were intramuscularly injected in BALB/c mice at day 0 and 21, followed by a booster inoculation 134 days later (155 days after prime inoculation). Blood samples were collected on days 36, 114, 148, and 171 post-prime immunization and IgG anti-RBD antibodies title was determined by indirect ELISA. On day 171, animals were sacrificed, and the spleens were obtained. The mononuclear cells (MNCs) were isolated from spleen suspensions by Ficoll-Paque density gradient and stimulated with 5 μg of S-ED. After 72 h, supernatants were harvested, and IFN-γ and IL-4 were quantified by ELISA. B IgG anti-RBD antibodies quantified by indirect ELISA. Antibody titers were calculated as the end-point plasma dilution yielding an OD higher than the cut-off value, which was calculated as the mean + 3 SD of the OD of pre-immune plasma. Dots are means ± SEM. C IgG anti-RBD titer pre- and post-booster immunization. Bars are means ± SEM. **p<0.01, Mann-Whitney U test. D Secreted IFN-γ and E IL-4, by MNCs derived from spleens. In both cases, positive controls obtained after MNCs stimulation with lectins are indicated as (+). Bars are means ± SEM of IFN-γ and IL-4 in pg.mL−1 after subtracting the amount in medium-stimulated cells. **p<0.01, ***p<0.001, Mann-Whitney U test

Furthermore, splenocytes from mice immunized with either adjuvant were able to produce IFN-γ and IL-4 in response to S-ED stimulus, suggesting the presence of antigen-triggered immune responses related to Th1 and Th2 profiles. Spleen MNCs from mice immunized with LipoSap secreted significantly higher levels of IFN-γ and IL-4 (p<0.01 and p<0.001, respectively) than the ones detected for mice immunized with Alum (Fig. 5D).

SARS-CoV2 neutralization assay

The neutralization capacity of the antibodies elicited by the LipoSap formulation was confirmed by a neutralization assay against the ancestral SARS-CoV-2 virus.

Neutralization titer (NT) was assessed at different time points during 178 days post-prime immunization. As shown in Fig. 6, an initial mean titer of 7680 was obtained at 35 days after prime immunization. NT tended to decrease towards day 150. However, the administration of a booster dose at day 155 caused a significant increase in NT, reaching a 12.8-fold increase in the neutralization capacity (p<0.01).

Fig. 6
figure 6

Neutralization titer (NT) against ancestral SARS-CoV-2 virus. The neutralization titer was determined for plasma collected from mice immunized with S-ED formulated with LipoSap. NT was defined as the inverse of the highest plasma dilution without any cytopathic effect (CPE). Samples with NT below 8 (limit of detection, LOD) were set at this value. Dots are means ± SEM. *p<0.05, **p<0.01. Kruskal-Wallis with Dunn’s post-test


The importance of vaccines to fight against and avoid viral illnesses has been historically demonstrated. Indeed, the negative impact that the worldwide COVID-19 pandemic (caused by SARS-CoV-2) had on health and economy highlighted the importance of vaccine development for facing emergencies. The S protein of SARS-CoV-2 constitutes an ideal antigen since it is a determinant of cell entry (through RBD interaction with ACE2 receptor) and tropism. Also, S presents epitopes that are responsible for triggering protective B- and T-cell responses (Arashkia et al. 2021). Furthermore, it has been already proved that S represents a robust antigen that elicits more potent immune responses compared to RBD alone (Mandolesi et al. 2021).

Previous studies carried out with first-generation vaccines (dead or attenuated virus) for SARS-CoV or MERS-CoV demonstrated that some adverse effects may appear after immunization, such as increased infectivity and immunopotentiation. The close genetic and pathogenic relationship of those viruses with SARS-CoV-2 suggests that second or third-generation vaccines would be better alternatives. Thus, the development of new-generation vaccines is considered a priority in order to fight against the emergent variants of concern. Despite the fact that mRNA vaccines have been successfully developed, several difficulties related to their cost, storage, and transportation have been reported (Mekonnen et al. 2022). Recombinant protein subunit vaccines have been demonstrated to be safe and effective to develop immune responses. For instance, these types of vaccines have already been developed and approved for human papillomavirus (HPV), hepatitis B virus (HBV), and influenza A virus. To date, several protein subunit vaccine candidates against SARS-CoV-2 have been approved or are in the last stages of clinical trials. Such is the case of NVX-CoV2373 (marketed as Novavax), the first recombinant subunit vaccine with emergency use authorization (EUA) by FDA for the USA and with WHO-EUA for 38 other countries (FDA 2022). There are also 12 different subunit vaccines based on spike or RBD sequence which have been currently approved in one or more countries (Bayani et al. 2023). However, it is important to consider that the efficacy of a recombinant protein subunit vaccine can be greatly enhanced by the co-administration of an adjuvant (Mandolesi et al. 2021; Martínez-Flores et al. 2021). In a pandemic situation, when the fast development of vaccines is determinant, the use of adjuvants is of particular importance since it reduces the amount of protein that is necessary to formulate a vaccine dose (Liang et al. 2021).

Herein, we describe the development of a sHEK cell line producing S-ED by several cationic lipid transfections followed by antibiotic selection. Large viral proteins such as S-ED are considered DTE proteins and thus the choice of the host for production is crucial to reach good productivity. As we have mentioned before, HEK cells are preferred over CHO-K1 cells as they can support the expression of DTE proteins both in transient and stable conditions (Schwarz et al. 2020). Despite many authors have already described the production, purification, and characterization of S expressed in HEK cells as a potential candidate for the design of COVID-19 vaccines, most of them relied on transient production processes which present many disadvantages related mainly to the inconsistency between batches (Cai et al. 2020; Esposito et al. 2020; Tee et al. 2020; Wrapp et al. 2020; Herrera et al. 2021; Stuible et al. 2021). Stable cell lines are preferred for large-scale production since they provide recombinant proteins with a consistent quality (O’Flaherty et al. 2020). Only CHO recombinant cell lines have been developed and described for the production of S protein (Tee et al. 2020; Pino et al. 2020; Johari et al. 2021; Liang et al. 2021). To our knowledge, this is the first work reporting results of S-ED stably expressed in sHEK cells.

It is largely known that differences in cell lines, culture medium, and production conditions can affect the glycosylation pattern, resulting in differences in the apparent molecular weight of the recombinant proteins. Previous reports revealed that the S protein is highly glycosylated with 22 potential N-linked glycosylation sites and five potential O-glycosylation sites (Schaub et al. 2021). Differences in glycosylation macro and microheterogeneity have already been described for S produced in HEK and CHO cells (Wang et al. 2020).

In this work, we found that, on the one hand, S-ED exhibited a higher apparent molecular mass than the theoretically calculated one when analyzed in SDS-PAGE assays. This result was expected since HEK cells can introduce post-translational modifications such as glycosylation, which increases the molecular mass. Moreover, the recombinant protein revealed the expected morphology as assessed by TEM. Similar images were previously shown and published by other authors (Walls et al. 2020; Amanat et al. 2020; McCallum et al. 2020; Pino et al. 2020; Thépaut et al. 2021). On the other hand, PNGase F treatment of S-ED produced a mobility shift, indicating the presence of N-glycans attached to the protein. However, the electrophoretic profile showed that the treated samples still presented a molecular mass higher than theoretical one, suggesting either an incomplete N-deglycosylation, the existence of O-glycosylation, or a combination of both. This result is in agreement with reports of mass spectrometry analyses of recombinant S purified from FreeStyle 293-F cells that confirm a strong occupation of the O-glycosylation site at Thr323 and trace levels of O-linked glycans at Thr325 (Schaub et al. 2021). Further glycosylation studies will be necessary to elucidate N- and O-glycan occupancy and structures of S-ED produced in this work. Importantly, the interaction of S-ED with hACE2 receptor present in recombinant HEK293T cells was verified, confirming its in vitro functionality.

Although many approaches have been reported regarding the generation of subunit vaccine candidates, only a few reached the clinical trials stage, indicating difficulties in defining an appropriate immunological product where a synergistic effect between antigens and adjuvants is desired. Considering that subunit vaccines are poorly immunogenic, the choice of the adjuvant is crucial for its design and further assessment of its effectiveness (Mekonnen et al. 2022).

In this work, we aimed to compare the immunogenicity of S-ED formulated with Alum or a saponin immune-stimulating complex (LipoSap) as adjuvants. The quantification of anti-RBD IgGs in plasma samples collected at different time points revealed a stronger humoral response elicited by LipoSap formulation compared with Alum. This result is in agreement with previous reports that demonstrate a higher specific antibody titer for the cage-like particle adjuvant compared with the traditional Alum (Bertona et al. 2017; Fontana et al. 2020; Lupi et al. 2022). As expected for a candidate vaccine, IgG titer increased significantly after a booster.

Many reports suggest that both humoral and cell-mediated immune responses may confer optimal protection against SARS-CoV-2. Thus, ideal candidate vaccines should be able to develop both responses (Liang et al. 2021; Wørzner et al. 2021; Coria et al. 2022). In the present study, both IL-4 and IFN-γ cytokines were detected after S-ED stimulation of MNCs from mice groups immunized with S-ED/Alum and S ED/LipoSap formulations, being cytokine levels significantly higher for the LipoSap group. It has been previously described that ISCOMs improve endocytosis by promoting the maturation of antigen-presenting cells (APC), a mechanism that is needed for efficient T-helper and CTL induction (Lövgren Bengtsson et al. 2011; Bertona et al. 2017). This aids the activation of CD4+ T cells and CD8+T cells and, as a consequence, the secretion of cytokines such as IL-4 and IFN-γ, explaining the better results obtained for the S-ED/LipoSap formulation. Previous findings reported by Carmen et al. (2021), for a S ferritin nanoparticle (SpFN) showed a better response for the formulation adjuvanted with saponin QS-21 compared to the one including AH. Particularly, a robust induction of IFN-γ-producing CD8+ T cells was exhibited (Carmen et al. 2021). This result is in agreement with the high IFN-γ concentration achieved in our work. This cytokine is produced as a consequence of a viral infection and presents a synergistic action with type I interferons.

The results presented in this work demonstrate that the S-ED/LipoSap formulation constitutes a potential vaccine for COVID-19. Also, it can be used as a heterologous booster for other vaccines. Furthermore, this formulation can be easily updated for the production of other SARS-CoV-2 variants of concern, taking advantage of the stable production process designed and reported in the present work.