HEK293T and Huh7 cell lines were cultured in Dulbecco’s Modified Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% non-essential amino acids (NEAA). For HBV neutralization assays, HepG2hNTCP cells (provided by Professor Stephan Urban, German Center for Infection Research, University of Heidelberg) were cultured as previously described (5). HepG2.2.2.15 cells, which are transfected stably with two copies of the HBV genome (gift from Dr. David Durantel, INSERM U871, Lyon, France), were utilized for WT HBV stock production, as reported earlier (15). All cell lines were maintained at 37°C with 5% CO₂ in an incubator with a humidified atmosphere. Stocks of HBV with a clinically relevant VEM, namely the G145R mutation within the antigenic domain of the S protein, were obtained by transfecting Huh7 cells with pGEM-4Z-HBV 1.3 G145R (5) using Lipofectamine 3000 (Invitrogen) according to the manufacturer’s protocols. Cell media were collected between days 7 and 12 post-transfection, concentrated via ultracentrifugation on a 20% sucrose layer. The viral stocks were then further purified to eliminate subviral particles (SVPs) via sucrose gradient ultracentrifugation as described in Wettengel et al. (16). Fractions containing HBV and lacking SVPs were determined via real-time qPCR and ELISA, respectively, and then quantified using real-time qPCR to determine viral titers (17).

The HBV-S/preS1^16-42 antigen was produced in CRISPR/Cas9-edited N. benthamiana lines with knockouts in their α (1, 3)-fucosyltransferase and β (1, 2)-xylosyltransferase genes (FX-KO) through vacuum-based agroinfiltration, as previously described (7). The S/preS1^16-42 antigen- encoding gene was introduced into the expression plasmid vector pEAQ-ht-DEST1 using Gateway cloning as described in a previous publication (6). The S/preS1^16-42 plasmid vector was transformed into Agrobacterium tumefaciens LBA4404 via electroporation. Plant leaves were infiltrated with Agrobacterium suspension and subsequently harvested 7 or 8 days after infiltration (7).

Leaves harvested from FX-KO N. benthamiana plants expressing the S/preS1^16-42 antigen were ground in liquid nitrogen and homogenized in 5 volumes of lysis buffer (0.15 M NaCl, 20 mM Na2HPO4, 20 mM sodium ascorbate, 0.1% Triton X-100, pH 7) containing a cocktail of protease inhibitors (Santa Cruz Biotechnology), as previously described (7).

Lysates were then concentrated via ultracentrifugation on a 20% sucrose layer, for 5 h at 30,000 rpm (SW 32 Ti, Beckman Coulter) and then further separated via ultracentrifugation on a 15-60% sucrose gradient at 32,000 rpm (SW 42Ti, Beckman Coulter) for 18 h to isolate high-molecular weight SVPs. Collected fractions were tested for the presence of HBV antigens via ELISA using the Monolisa HBsAg Ultra kit (BioRad). Positive fractions were then pooled, dialyzed against PBS and incubated with 1% activated carbon (Carbocit^®, S.C. Biofarm S.A.) for 3 h at 4°C with head-over-tail agitation, to remove smaller protein impurities, and then further purified via size-exclusion chromatography on the CaptoCore 400 resin (Cytiva), by gravitational flow, as previously described (7). Flow-through fractions containing S/preS1^16-42 SVPs were identified via ELISA as described above, pooled and dialyzed against PBS 0.01X and then SpeedVac-concentrated at 35°C for 5 h (Thermo Scientific-Pierce). The antigen was then quantified by western blot and Coomassie staining under denaturing conditions, as previously described (4).

Protein samples were heat-denatured by boiling for 10 min in Laemmli buffer and then were separated via sodium dodecyl sulfate 12% polyacrylamide gel electrophoresis (SDS-PAGE) and subjected to either Coomassie staining (using ReadyBlue Protein Gel Stain from Sigma-Aldrich) or transferred onto nitrocellulose membranes. After blocking the membranes in 10% skimmed dry milk in PBS for 1 h, the membranes were incubated with mouse anti-preS1 antibodies (sc-57762, Santa Cruz Biotechnology, 1:1,000) overnight at 4°C, followed by a 1 h incubation at room temperature (RT) with HRP-conjugated mouse-IgGκ binding protein (sc-516102, Santa Cruz Biotechnology, 1:10,000). Protein visualization was accomplished using an enhanced chemiluminescence assay (ECL, Thermo Scientific-Pierce). Protein densitometry analysis was performed using Image J software from the National Institutes of Health, and antigen concentrations were estimated through a standard curve generated with commercial L protein processed under the same experimental conditions (preS1 antigen, Beacle).

Female NSG mice (stock number 005557) were purchased from Jackson Laboratory (Bar Harbor, ME). All animals were housed in specific pathogen-free (SPF) conditions and used at the age of 8-10 weeks. Animal experiments were conducted in accordance with national and institutional guidelines for animal care and were approved by the Institute Animal Ethics Committee and by the National Authority ANSVSA nr. 488/22.01.2020.

hPBMCs were obtained from the peripheral venous blood of three healthy volunteers after obtaining informed consent. The cells were isolated via Ficoll-Hypaque density gradient centrifugation (1.077 g/mL density, Merck, Darmstadt, Germany). The isolated hPBMCs, suspended in PBS, were injected at 1×10⁷ cells/dose, intraperitoneally (ip) into NSG mice.

Groups of 5-7 NSG mice were reconstituted with hPBMCs of each donor depending on the available cell number. The success of the transplant was demonstrated by human lymphocytes phenotyping via flow-cytometry using mice peripheral blood collected on the fourth day after transplantation. hPBMCs-engrafted mice were immunized by intramuscular injection with 30 µg/dose of HBV-S/preS1^16-42 antigen purified from FX-KO N. benthamiana adjuvanted with Addavax (InVivoGen) (n=11) or with Addavax only as control (n=8). The immunization protocol consisted of a prime immunization, the fifth day after hPBMCs engraftment, followed by two boosters at 2-week intervals. The mice were monitored daily and their body weight was verified twice per week. One week after the last boost (day 35), spleens and serum samples were collected ( Table 1 ). Mice showing clinical deterioration before the last boost were removed from the study. Serum samples were processed for detection of human antibodies by ELISA and neutralization analysis, and spleen cells were used for evaluation of cellular immune response by Luminex multiplex assays and FACS analysis.

Peripheral blood and spleens were obtained from hPBMCs-engrafted mice and were prepared as single-cell suspensions. Human immune cell populations were detected using mAbs specific for the following human antigens: CD3-PE (clone HIT3), CD14-FITC (clone HCD14), and CD19-APC (clone 4G7). Isotype antibodies were used as negative controls and reported values were corrected using these isotype control values.

To exclude murine host cells, anti-mouse Ly-6G/Ly-6C Pacific Blue (Gr-1) (clone RB6-8C5) monoclonal antibody staining was performed in all experiments. All antibodies were purchased from BioLegend (San Diego, CA). Specific mAbs were added to the samples and incubated for 30 min at 4°C. Stained peripheral blood samples were then washed and red blood cells were lysed using BD FACS Lysing solution (BD Biosciences) followed by analysis via flow-cytometry. At least 50 000 events were acquired on the FACSCanto II instrument (BD Biosciences). Data analysis was performed using the BD FACSDiva Software v.6.1.2.

Blood samples were collected under anesthesia from individual mice by orbital puncture at the end of the experiment. Blood sera were obtained from clotted blood following centrifugation at 14,000 × g for 10 min at RT then stored at -80°C until further use for antibody titration. The antigen-specific human IgG (hIgG) and IgM (hIgM) were detected via ELISA as previously described (5), by using a suspension of UV-inactivated HBV in PBS (containing 0.4 μg/mL of HBsAg) for plate coating and anti-human IgG H&L (HRP) (ab6858) or anti-human IgM mu chain (HRP) (ab97205), as secondary antibodies. Statistical analysis was performed by using the Wilcoxon rank-sum test.

Splenocytes isolated from immunized mice were seeded in 96-well culture plates at 0.5×10⁶ and 0.1×10⁶ cells/well in RPMI-1640 media containing 10% fetal calf serum (FCS) and 2-mercaptoethanol (50 µM, Sigma-Aldrich). Triplicate samples were stimulated with UV-inactivated HBV (10 µg/mL HBsAg). Unstimulated cells were used as control samples for the baseline response and cell functionality was assessed by stimulation with Concanavalin A (ConA 5 µg/mL, Sigma-Aldrich). The cultures were incubated for 36 h at 37°C, 5% CO₂. The supernatant was then removed and stored at -80°C for later use. Cellular immune responses were analyzed by detecting cytokine secretion using a human Luminex-based bead multiplex assay (LXSAHS-14, R&D Systems) according to the manufacturer’s instructions. We selected a panel of 14 cytokines secreted by different subclasses of immune cells (IP-10, IL-23, MIP-1α, IFN-γ, IL-2, IL-10, IL-17, TNF-α, MCP-1, IL-1β, IL-4, IL-6, IL-13, IL-5). Statistical analysis was performed by using the Wilcoxon rank-sum test.

HepG2hNTCP cells were plated in 48-well plates one day before HBV infection. Sera collected from immunized mice at day 35 post-immunization were diluted 1:20 in complete DMEM media with 4% polyethylene glycol (PEG, Sigma Aldrich) and incubated for 1 h at 37°C with WT HBV and the VEM HBV-SG145R (100 genome equivalents/cell). The pre-inoculum was then used to infect the plated HepG2hNTCP cells. Control conditions included cell treatment with 1 µM Myrcludex B (Pepscan), a well-established HBV entry inhibitor (18), for 3 h before virus inoculation, as well as cells infected with untreated HBV. At 16 h post-infection, cells were thoroughly washed with PBS and incubated with complete DMEM media supplemented with 2.5% DMSO (AppliChem). The medium collected between days 7-11 post-infection was used to quantify HBsAg secretion through the Monolisa HBs Ag-Ab PLUS kit (BioRad). The inhibition of HBV infection by the immune sera was expressed as a percentage relative to infectivity values in the presence of pre-immune sera at the same dilution. Statistical analysis was conducted using the Mann-Whitney U test in GraphPad Prism version 6.

The S/preS1^16-42 chimeric antigen ( Figure 1 ) was produced by transient expression in CRISPR/Cas9-edited N. benthamiana plants with a humanized N-glycosylation pathway (FX-KO) (7). To obtain sufficient quantities of antigen for the immunization experiments, 50 g of FX-KO N. benthamiana leaves expressing the S/preS1^16-42 antigen were lysed and the resulting extracts were concentrated via 20% sucrose-bed ultracentrifugation to isolate assembled SVPs. SVPs were further purified via sucrose gradient ultracentrifugation and gel filtration of antigen-positive fractions on CaptoCore 400 as previously described (7). Samples were concentrated and analyzed by SDS-PAGE followed by either Coomassie staining, or western blot and detection with anti-preS1 antibodies ( Figure 2 ). These analyses revealed the presence of a major, broad band at ~25 kDa, representing the monomeric antigen, and of weaker, higher molecular weight bands corresponding to protein dimers and oligomers. Total protein and S/preS1^16-42 quantification indicated an antigen purity of ~80%.

To verify hPBMCs engraftment, we analyzed human leukocytes in the peripheral blood of recipient mice 4 days after intraperitoneal injection of 1×10⁷ hPBMCs (day -1). A gating strategy using anti-mouse Ly-6G/Ly-6C staining was employed to exclude murine cells. As shown in Figure 3 , human CD3+ T cells were detected in all mice, with percentages ranging from 1 to 3%. However, human CD19+ B cells and CD14+ monocytes were not detected by flow-cytometry analysis in transplanted mice. A table detailing CD19+ and CD14+ cell percentages is shown in the supplementary ( Supplementary Table S1 ). These results suggest that at least part of the hPBMCs were successfully engrafted into NSG mice.

Secondary lymphoid tissues, including the spleen, are sites of antigen presentation and lymphocyte activation and are consequently critical for antigen-specific immune response evaluation. Therefore, we analyzed the presence of human CD3+ T cells in the spleen of hPBMCs-engrafted mice on day 35 after first immunization. The same gating strategy was applied for splenocytes as for peripheral blood cells to exclude murine cells. Immunophenotyping analysis revealed high engraftment rates of human CD3+ T cells in the spleen of humanized mice while human B cells and monocytes were barely detected ( Figure 4 ).

The humoral immune response of NSG mice was investigated following three immunizations with the S/preS1^16-42 antigen administered at 14-day intervals. As clinical deterioration was observed in these mice, serum samples were collected one week after the last boost. The presence of antigen-specific hIgG and hIgM was determined by ELISA assay using plates coated with inactivated HBV as previously described (5). At 35 days post-immunization elevated hIgG and hIgM titers were detected in both the adjuvant control group (Addavax) and the S/preS1^16-42 antigen group. Higher hIgG and hIgM responses were observed in the antigen immunization group ( Figure 5 ). Due to the inherent heterogeneity of the humoral immune response previously observed in BALB/c mice (7), the genetic variability between the human donors and the small number of mice per group, this response was not statistically significant ( Figure 5 ). The hIgG and hIgM endpoint titers for each individual mouse are presented in the supplementary section ( Supplementary Table S2 ). The lack of significant differences between the two groups may also be explained by the properties of the used adjuvant Addavax, which is a squalene-based oil-in-water nano-emulsion that can activate the immune response in a nonspecific way (19).

To evaluate antigen-specific cellular immune response of hPBMCs-transplanted NSG mice, isolated splenocytes were stimulated ex-vivo for 36 h with UV-inactivated HBV and secretion of various human cytokines was measured in cell culture supernatants. We found that antigen stimulation induced only a slight increase in IFN-γ levels in HBV-S/preS1^16-42 antigen-immunized groups compared to adjuvant-control groups with no statistical difference between the two immunization groups or as compared with the unstimulated cells ( Figure 6 ). Notably, from the 14 cytokines tested, except IFN-γ, all other tested cytokines had minimal to negligible secretion levels. However, the functional activity of the isolated splenocytes was highlighted by in vitro cytokine secretion in response to polyclonal ConA stimulation, which was used as a positive control (data not shown).

Finally, to more accurately determine whether immunization with the S/preS1^16-42 antigen in NSG mice could give rise to a neutralizing antibody response, we evaluated the capacity of sera to protect from HBV infection in vitro. WT-HBV and a vaccine-escape variant containing the SG145R mutation were obtained as previously described (5) and purified. HepG2hNTCP cells were then infected with an inoculum containing virus (100 Geq/cell) and pre- or post-immune sera (1:20 dilution) from NSG mice vaccinated with antigen or adjuvant. Pre-treatment with Myrcludex B was used as a positive control of specific inhibition of viral entry. As shown in Figure 7 a strong neutralization ability of the immune sera from NSG mice immunized with the S/preS1^16-42 antigen (of about 50%) was observed against both WT-HBV ( Figure 7A ) and VEM-HBV ( Figure 7B ), when compared to the group immunized with only adjuvant. This suggests that the antibodies induced by S/preS1^16-42 immunization in NSG mice are specific and potently neutralizing, which is consistent with the immunogenicity profile shown by this antigen in other mouse models (5). Notably, we have previously observed a higher HBV neutralization activity of ~ 75%, for sera from BALB/c mice immunized with the same FX-KO N. benthamiana – derived S/preS1^16-42 antigen (7). However, this effect is likely due to the significantly higher IgG titers triggered by vaccination in BALB/c mice, as compared to the NSG animal model.