Impact of adjuvants on a recombinant SARS-CoV-2 spike vaccine: Protective efficacy and recall antibody response following homologous challenge

Highlights

• •Scientific question: The continual evolution of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), compounded by immune imprinting, poses an ongoing challenge to the development of broadly protective coronavirus disease 2019 (COVID-19) vaccines. How exactly pre-existing immunity, particularly its magnitude and affinity, shapes the strength and breadth of the recall antibody response remains poorly understood.
• •Evidence Before This Study: Previous studies have shown that booster vaccinations and breakthrough infections can broaden immunity against SARS-CoV-2 variants. It is also established that extending the interval between antigen exposures or increasing antigenic distance between variants can attenuate original antigenic sin.
• •New Findings: We demonstrate that adjuvants substantially affect both the potency and epitope diversity of antibodies elicited by recombinant vaccines. Furthermore, pre-existing high-affinity antibodies mediate epitope masking of immunodominant sites, thereby shifting the immune response toward conserved, subdominant epitopes with cross-neutralizing potential.
• •Significance of the Study: The quantity and quality of pre-existing antibodies critically guide the recall antibody response toward broad neutralization and affinity maturation upon re-exposure. This insight offers a rational framework for developing next-generation vaccines against SARS-CoV-2 and other antigenically variable pathogens.

Abstract

Immune imprinting, or original antigenic sin, challenges the control of evolving severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). How the quantity and quality of pre-existing immunity modulate the recall antibody response upon re-exposure remains poorly understood. We immunized hamsters with a recombinant ancestral spike protein (S-2P) formulated with one of four distinct adjuvants to assess the impact of adjuvant-induced immunity on protective efficacy. Furthermore, we analyzed the modulatory effect of the adjuvant on the potency, breadth, and evolution of antibody responses after homologous viral challenge. We found that vaccination with Montanide ISA720-adjuvanted S-2P (S-2P:ISA720) not only induced higher initial neutralizing antibody titers and conferred stronger protection but also established a pre-existing immune state capable of cross-neutralizing the antigenically distant Omicron BA.1 variant. In contrast, aluminum hydroxide adjuvanted S-2P (S-2P:Al) elicited comparatively lower neutralizing titers and showed no cross-neutralization against BA.1. Following viral challenge, the S-2P:ISA720 group exhibited significant affinity maturation toward conserved epitopes, which markedly enhanced cross-neutralization against BA.1 without increasing neutralizing titers or affinity against the homologous strain or the antigenically related Delta variant. Conversely, the S-2P:Al group mounted a narrow, imprinting-facilitated response, characterized by boosting of strain-specific antibodies without substantial improvement in BA.1 neutralization. These findings suggest that in S-2P:ISA720-immunized animals, high-affinity antibodies mediate epitope masking of immunodominant sites, thereby redirecting responses toward subdominant conserved epitopes with cross-neutralizing potential. In conclusion, our study demonstrates that adjuvants can critically guide recall immune responses toward breadth and affinity maturation, offering a rational strategy for developing next-generation vaccines against SARS-CoV-2 and other variable pathogens.

1. Introduction

The coronavirus disease 2019 (COVID-19) pandemic, which emerged in late 2019, has had devastating impacts on global public health and socioeconomic systems. Although multiple vaccines were developed and deployed at unprecedented speed and scale, playing an important role in preventing infection and reducing severe disease [[1], [2], [3], [4]], the relentless evolution of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has led to the continual emergence of variants with significant immune evasion capabilities and enhanced transmissibility [[5], [6], [7], [8]]. These factors have sustained the pandemic despite high vaccination coverage and widespread prior infection, resulting in recurrent epidemic waves. This evolving viral landscape continues to challenge the efficacy of current vaccines and complicates the development of next-generation candidates.

In combatting the surge of rapidly evolving SARS-CoV-2 variants, strategies such as antigen updates, multivalent vaccines, and booster immunizations have been widely adopted to enhance protective antibody responses against variant strains [[9], [10], [11], [12], [13]]. Notably, “hybrid immunity”, resulting from a combination of vaccination and infection, has been shown to elicit stronger and more broad-spectrum antibody responses, significantly enhancing cross-protection against variants [[14], [15], [16], [17]]. A major complicating factor, however, is immune imprinting (or “original antigenic sin”), wherein prior exposure skews subsequent antibody responses toward immunodominant epitopes of the initial strain, potentially at the expense of responses against conserved but subdominant epitopes on new variants [[18], [19], [20], [21], [22]]. This greatly constrains the development of broad-spectrum protective immunity.

Critically, there remains a lack of systematic comparative studies investigating how pre-existing immunity, differing in both magnitude (e.g., antibody titer) and quality (e.g., epitope specific memory B cell repertoire), shapes the response to subsequent viral exposure or booster vaccination. Understanding how these features influence the recall and evolution of antibody responses is essential for overcoming immune imprinting and designing next-generation vaccines that confer broad, durable protection against diverse SARS-CoV-2 variants and other antigenically variable pathogens.

In this study, we immunized hamsters with recombinant ancestral SARS-CoV-2 Spike protein vaccines formulated with different adjuvants, establishing models with varying tiers of pre-existing immunity. Animals were then challenged with the homologous strain to evaluate protective efficacy and, most importantly, to analyze the post-challenge antibody response, including homologous and cross-variant neutralization, as well as affinity maturation. Using this controlled model, we aimed to delineate how pre-existing immunity influences the breadth and quality of recalled antibody responses.

2. Materials and methods

2.1. Cells, viruses, and proteins

HEK293T and HEK293 cells were acquired from the American Type Culture Collection (ATCC). HEK293 cells stably expressing human angiotensin-converting enzyme 2 (ACE2), designated HEK293-ACE2, were generated by transduction with a lentiviral vector encoding human ACE2, followed by selection with puromycin. Vero E6 cells were generously provided by Dr. Rong Zhang (Fudan University, China). HEK293T, HEK293-ACE2, and Vero E6 cells were maintained in Dulbecco’s Modified Eagle’s Medium (DMEM; Gibco, USA) supplemented with 10 % fetal bovine serum (FBS; Gibco, USA), 1 % penicillin–streptomycin (Thermo Fisher Scientific), 1 % L-glutamine (Thermo Fisher Scientific), and 1 % non-essential amino acids (Thermo Fisher Scientific) at 37 °C in a 5 % CO₂ atmosphere. FreeStyle™ CHO-S suspension cells (Invitrogen, USA) were cultured in EX-CELL™ 325 PF CHO Serum-Free Medium (Sigma-Aldrich) containing 1 % L-glutamine and 1 % penicillin–streptomycin. Cells were grown in nonpyrogenic, vented polycarbonate Erlenmeyer flasks (Thermo Fisher Scientific) at 37 °C with 5 % CO₂ under constant agitation at 130 revolutions Per minute (rpm). The ancestral SARS-CoV-2 strain (GenBank accession: MT622319.1) was isolated from nasopharyngeal swab specimens obtained from nucleic acid-positive patients and subsequently propagated and titrated in Vero E6 cells as previously reported [23]. Recombinant S1, S2, and receptor-binding domain (RBD) proteins derived from the ancestral, Delta, and Omicron variants of SARS-CoV-2 were procured from Sino Biological Inc. (Beijing, China).

2.2. Spike protein expression and purification

A mammalian codon-optimized gene encoding the ancestral SARS-CoV-2 spike protein ectodomain (residues 1–1211; GenBank: NC_045512.2) was synthesized by Generay Biotech (Shanghai, China). The construct included the following features to facilitate production and stabilization: a C-terminal T4 phage fibritin-derived foldon (Tfd) trimerization domain, a GSAS substitution at the furin cleavage site (replacing RRAR), two proline-stabilizing mutations (K986P/V987P) to lock the protein in its prefusion conformation, and a C-terminal polyhistidine tag for purification [24]. Following expression in CHO cells, the full-length ectodomain trimer was purified using nickel-nitrilotriacetic acid (Ni-NTA) resin, as described in our recent publication [25].

2.3. Animals

Male golden Syrian hamsters aged 6 weeks were obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). Animals were housed in individually ventilated caging systems under specific pathogen-free conditions. All experimental procedures were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of Navy Medical University and were performed in strict accordance with the Guidelines for the Administration of Laboratory Animals issued by the Chinese Ministry of Science and Technology.

2.4. Vaccine formulation and immunization

The purified spike protein containing two proline-stabilizing mutations (S-2P) protein was emulsified with one of the following adjuvants: aluminum hydroxide (Croda, Denmark), Montanide™ ISA 720 VG, Montanide™ ISA 51 VG, or Sepivac™ SWE (all from Seppic, France). The resulting formulations were designated S-2P:Al, S-2P:ISA720, S-2P:ISA51, and S-2P:SWE, respectively. Each hamster received two intramuscular immunizations (100 μL each) spaced three weeks apart, containing 10 μg of S-2P protein. The adjuvant-to-antigen volume ratios were 70 % for ISA720, and 50 % for both ISA51 and SWE. The S-2P:Al formulation contained 100 μg of aluminum hydroxide per dose. Control animals were administered phosphate-buffered saline (PBS) following the same schedule.

2.5. Enzyme-linked immunosorbent assay (ELISA) for serum antibody detection

Antibody responses in hamster sera were evaluated by indirect ELISA as previously described [26], with minor modifications. In brief, 96-well high-binding plates (Corning) were coated with 1 μg/mL of recombinant SARS-CoV-2 S1, S2, or RBD protein in carbonate-bicarbonate buffer (pH 9.6) and incubated overnight at 4 °C. Plates were then blocked with 5 % non-fat milk in PBS containing 0.05 % Tween-20 (PBST) for 2 h at room temperature. After washing, serially diluted serum samples were added and incubated for 2 h at room temperature. Plates were washed extensively with PBST, followed by incubation with horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG (Thermo Fisher Scientific) for 1 h. After a final wash, tetramethylbenzidine (TMB) substrate was added, and the reaction was stopped with 2 mol (M) sulfuric acid. Absorbance was measured at 450 nm with a reference wavelength of 630 nm. The endpoint titer was defined as the highest serum dilution that yielded an optical density value greater than or equal to 2.1 times that of the negative control wells.

2.6. Determination of Serum-Specific antibody avidity index

The avidity of specific antibodies was assessed using a urea dissociation ELISA, as described for the total IgG ELISA above, with the following modifications [27]. Each serum sample was tested in duplicate. After primary antibody incubation and washing, wells were subjected to a dissociation treatment: one well received 50 µL of 8 M urea, while the parallel well received 50 µL of PBS (control). The plate was incubated for exactly 8 min at room temperature on a vertical shaker. Precise timing of this step is critical. Following incubation, the plate was washed extensively with PBST and then incubated with the secondary antibody. All subsequent steps were identical to those used for total IgG detection. The antibody avidity index was calculated as the percentage of urea-resistant antibody titer relative to the total titer, using the formula: avidity index (%) = (specific IgG titer in the urea-treated well / specific IgG titer in the PBS-treated well) × 100 %.

2.7. Serum neutralization with authentic SARS-CoV-2

Vero E6 cells were seeded into 96-well plates at a density of 1 × 10⁴ cells per well and cultured overnight. Serum samples were serially diluted twofold and mixed with an equal volume of ancestral SARS-CoV-2 virus suspension containing 0.1 multiplicity of infection (MOI). The virus-serum mixtures were incubated at 37 °C for 1 h. Subsequently, the mixtures were added to Vero E6 cell monolayers in the presence of 2 μg/mL TPCK-treated trypsin. After 24 h of incubation, cells were fixed with ice-cold methanol for 30 min at –20 °C and blocked with 3 % bovine serum albumin (BSA) in PBS. Viral infection was detected using a rabbit polyclonal antibody against SARS-CoV-2 nucleocapsid (NP) protein (Sino Biological) followed by an Alexa Fluor 488-conjugated goat anti-rabbit IgG secondary antibody (Thermo Fisher Scientific). Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich). Images were acquired using a Cytation 5 multimodal reader (BioTek Instruments, USA), and infected cells were quantified using Gen5 software (version 3.10). The neutralization titer (50 % inhibition concentration, IC₅₀) was defined as the reciprocal of the serum dilution that inhibited 50 % of infection compared to virus-only controls and was calculated using four-parameter logistic regression in GraphPad Prism version 8.0.

2.8. Production and neutralization assay of SARS-CoV-2 pseudotyped viruses

Lentiviral particles pseudotyped with SARS-CoV-2 spike proteins were produced by co-transfecting HEK293T cells with a packaging plasmid, a transfer plasmid encoding enhanced green fluorescent protein (EGFP) under the control of a cytomegalovirus (CMV) promoter, and a plasmid expressing the spike protein of either the Delta or Omicron BA.1 variant using Lipofectamine 2000 (Thermo Fisher Scientific). The spike expression construct contained a C-terminal 19-amino acid deletion to enhance pseudovirus incorporation. At 48 h post-transfection, supernatants were collected, clarified by centrifugation at 1,000 × g for 10 min, and filtered through a 0.45 μm polyethersulfone membrane. Aliquots were stored at –80 °C until use. For neutralization assays, sera were serially diluted and incubated with pseudoviruses at 37 °C for 1 h. The mixtures were then added to HEK293-ACE2 target cells (1 × 10⁴ cells per well) seeded in 96-well plates at a MOI of 0.1. After 4 h, the medium was replaced with fresh DMEM containing 2 % FBS. At 48 h post-infection, EGFP-positive cells were quantified by fluorescence imaging using a Cytation 5 system. The 50 % inhibitory dilution (ID₅₀) was determined by nonlinear regression analysis using GraphPad 5 Prism.

2.9. Virus challenge in hamsters

Hamsters were anesthetized by inhalation of isoflurane and inoculated intranasally with 3 × 10⁴ plaque-forming units (PFU) of the ancestral SARS-CoV-2 strain in a 50 μL volume, as previously described [23]. Body weight was monitored daily for two weeks post-infection. On day 4 post-infection, a subset of animals was humanely euthanized for collection of nasal turbinates and lung tissues. Tissues were snap-frozen in liquid nitrogen and stored at –80 °C for subsequent viral load and cytokine analysis.

2.10. Quantification of viral RNA and inflammatory cytokine expression

Total ribonucleic acid (RNA) was extracted from homogenized tissue samples using TRIzol Reagent (Invitrogen) according to the manufacturer’s instructions. Complementary deoxyribonucleic acid (cDNA) was synthesized from 1 μg of total RNA using the PrimeScript RT Master Mix (Takara Bio, Japan). Quantitative real-time polymerase chain reaction (qRT-PCR) was performed using TB Green Premix Ex Taq II (Takara Bio) on a QuantStudio 5 Real-Time PCR System (Applied Biosystems). The following primer sequences were used in Table 1.

Table 1.

Primer sequences of quantitative real-time polymerase chain reaction.

Gene	Forward (5′-3′)	Reverse (5′-3′)
SARS-CoV-2 N	AAGGCGTTCCAATTAACACCA	TGCCGTCTTTGTTAGCACCA
Hamster β-actin	AGCAGTCTGTTGGAGCAAGC	TCTAGGGAATTGGGGTGGCT
Hamster IFN-γ	ATGACCTTCAGGTTCAGCGG	AGCACCGACTTCTTTTCCGT
Hamster IP-10	TTTATACGTCGGCCTATGGC	TAGTAGAGTTGGGGACTCTTGT
Hamster MX2	CGCAGACATACCAGAAGATGACA	TTCTCCTCCCCTTGTACTATGGC

Gene expression levels were normalized to β-actin, and relative quantification was performed using the comparative threshold cycle (2^–ΔΔCT) method.

2.11. Histopathological analysis

Lung tissues were fixed in 4 % paraformaldehyde for 72 h, processed through a graded ethanol series, and embedded in paraffin. Sections (3–4 μm thickness) were cut and stained with hematoxylin and eosin (H & E) using standard protocols. Histopathological images were captured using a Cytation 5 imaging system.

2.12. Statistical analysis

All statistical analyses were performed using GraphPad Prism version 8.0. Data are presented as mean ± standard error of the mean (SEM). Differences between two groups were analyzed using an unpaired or paired two-tailed Student’s t-test, as appropriate. Multiple group comparisons were conducted by one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test. A P-value of less than 0.05 was considered statistically significant (*P < 0.05, **P < 0.01, ***P < 0.001). Non-significant differences are indicated as “ns” (P ≥ 0.05).

3. Results

3.1. Vaccine preparation, hamster immunization, and virus challenge

The prefusion-stabilized spike trimer of ancestral SARS-CoV-2 (designated S-2P) was expressed in CHO cells and purified by affinity chromatography. The purified S-2P was analyzed by both native and denaturing SDS–PAGE, which confirmed its trimeric state [25].

The S-2P protein was formulated with aluminum hydroxide (Croda, Denmark) or with the Seppic adjuvants Montanide ISA 720, ISA 51, and Sepivac SWE (Seppic, France). The resulting formulations, designated S-2P:Al, S-2P:ISA720, S-2P:ISA51, and S-2P:SWE, were used for immunization.

Golden Syrian hamsters, which represent a natural infection model for SARS-CoV-2 that recapitulates key disease features including pulmonary inflammation and weight loss [28], were employed to evaluate antibody-mediated immune responses and the protective efficacy of the vaccine candidates. Animals were randomly assigned to five groups (n = 5 per group) and immunized intramuscularly twice at a 3-week interval with one of the vaccine formulations or PBS (Fig. 1A). Six weeks after the initial immunization, all hamsters were challenged intranasally with ancestral SARS-CoV-2. Blood samples were collected at weeks 2, 5, and 9 after the primary immunization for serological analysis.

Fig. 1.

Schematic timeline of hamster immunization and challenge experiments. A) Hamsters were immunized with S-2P formulated with one of four different adjuvants (n = 5 per group) under a prime-boost regimen and subsequently challenged intranasally with ancestral SARS-CoV-2. Body weight was monitored daily. Blood samples were collected at weeks 2, 5, and 9 after the primary immunization for serological analysis. B) Hamsters were immunized with S-2P adjuvanted with one of two different adjuvants (n = 3 per group) and challenged intranasally with ancestral SARS-CoV-2. On day 4 post-challenge, animals were euthanized, and nasal turbinates and lung tissues were collected for quantification of viral nucleic acid load, inflammatory cytokine mRNA expression, and histopathological evaluation. Abbreviations: SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; PBS, phosphate-buffered saline; dpi, days post-infection; mRNA, messenger ribonucleic acid; S-2P, spike protein containing two proline-stabilizing mutations.

For histopathological evaluation, viral load quantification, and cytokine profiling, three additional groups, specifically S-2P:Al, S-2P:ISA720, and PBS control (n = 3 per group), were included (Fig. 1B). On day 4 post-challenge, hamsters were anesthetized via isoflurane inhalation and euthanized. Nasal turbinates and lung tissues were collected for further analysis.

3.2. Antibody responses and protective efficacy of S-2P vaccines with different adjuvants

By week 2 post-booster immunization, S-2P:ISA720 elicited the highest IgG antibody titers against the S1 and S2 subunits, as well as the most potent neutralizing antibody responses among all tested formulations, representing 4.6-, 4.5-, and 6.7-fold increases relative to the S-2P:Al group, respectively (Fig. 2A–C). The other three vaccine candidates induced neutralization activities at comparable but lower magnitudes. Notably, neutralization activity was detectable as early as day 14 after the primary immunization in all five hamsters immunized with S-2P:ISA720 under the specified dilution conditions. In contrast, neutralization was observed in only one animal from the S-2P:ISA51 group, and no activity was detected at this time point in sera from either S-2P:Al or S-2P:SWE groups.

Fig. 2.

Antibody response and protective efficacy of adjuvanted ancestral SARS-CoV-2 S-2P vaccines. Hamsters (n = 5 per group) received immunization with different adjuvanted S-2P at weeks 0 and 3. Sera were collected at weeks 2 and 5 (week 3 post-booster immunization). Serum anti-S1 and anti-S2 IgG antibodies were determined using ELISA. the dashed lines represent the LLOD. Three weeks post-boost immunization, animals were challenged with ancestral SARS-CoV-2 and monitored for clinical outcomes. A) Serum Anti-S1 IgG endpoint titers at week 5 (The LLOD is 200-fold dilution). B) Serum anti-S2 IgG endpoint titers at week 5 (The LLOD is 200-fold dilution). C) Serum neutralizing antibody titers (NT₅₀) against authentic ancestral SARS-CoV-2 at weeks 2 and 5 (The LLOD is 100-fold dilution). D) Daily body weight changes post-challenge. Abbreviations: LLOD, lower limits of detection; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; PBS, phosphate buffered saline; S-2P, spike protein containing two proline-stabilizing mutations.

Following challenge with ancestral SARS-CoV-2, PBS-control hamsters exhibited clinical signs and sustained progressive weight loss through day 6, after which gradual recovery occurred (Fig. 2D). In comparison, hamsters immunized with S-2P:ISA720 showed no appreciable weight loss beyond day 1 post-challenge, with weight profiles comparable to those of uninfected controls. Although other vaccinated groups exhibited moderately delayed weight gain, none showed significant weight loss characteristic of SARS-CoV-2 infection (Fig. 2D).

Collectively, these results indicate that all four experimental vaccines elicited protective antibody responses and conferred effective protection against homologous viral challenge.

3.3. Vaccination significantly reduces viral loads but does not confer sterilizing immunity

To further assess vaccine efficacy, nasal turbinate and lung tissues were collected on day 4 post-infection from hamsters immunized with S-2P:Al, S-2P:ISA720, or PBS (n = 3 per group). Consistent with the differential protection observed, immunization with S-2P:Al and S-2P:ISA720 resulted in 1.4- and 9.6-fold reductions in viral load in nasal turbinates, and 12.7- and 24,962.6-fold reductions in the lungs, respectively, compared to PBS controls (Fig. 3A).

Fig. 3.

Assays for viral nucleic acid load, inflammatory cytokine mRNA expression, and histopathological examination. Hamsters (n = 3 per group) immunized with S-2P:ISA720, S-2P:Al, PBS control and uninfected hamster were sacrificed at 4 days post-infection for analysis of viral nucleic acid load, inflammatory cytokine mRNA expression, and histopathological alterations. A) SARS-CoV-2 nucleocapsid nucleic acid load in nasal turbinates and lungs. B) Inflammatory cytokine mRNA levels in lungs. C) H&E-stained lung sections at 4 dpi. Representative photomicrographs of left whole-lung sections (10 ×, left panels) from three hamsters per group, with magnified images (100 ×, right panels) showing selected areas from the far-right section. Abbreviations: SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; PBS, phosphate buffered saline; mRNA, messenger ribonucleic acid; H & E, hematoxylin and eosin; dpi, days post-infection.

Analysis of inflammatory cytokine mRNA levels in lung tissues revealed that SARS-CoV-2 infection strongly upregulated the expression of MX2, CXCL10, and interferon-gamma (IFN-γ). Immunization with S-2P:ISA720 significantly suppressed the upregulation of these cytokines, reducing their levels by 31.6-, 25.9-, and 93.1-fold, respectively (Fig. 3B). In comparison, S-2P:Al immunization resulted in more modest reductions of 3.4-, 4.3-, and 2.7-fold, respectively (Fig. 3B).

Histopathological examination of lung sections via H&E staining indicated that SARS-CoV-2 challenge induced severe pneumonia in control animals, characterized by alveolar septal thickening, diffuse inflammatory infiltration, and extensive hemorrhage. Vaccinated animals exhibited markedly attenuated lung pathology, with the S-2P:ISA720 group demonstrating nearly complete protection and significantly less damage than the S-2P:Al group (Fig. 3C).

Together, these findings indicate that although the vaccines did not confer sterilizing immunity, they significantly suppressed virus infection and ameliorated virus-induced pathology. The superior control of viral load and inflammation in the S-2P:ISA720 group underscores the role of adjuvant in enhancing vaccine efficacy. This pattern is consistent with the typical immune profile induced by intramuscular immunization, which typically induces robust systemic immunity but limited mucosal protection within the respiratory tract.

3.4. Divergent antibody response patterns following homologous virus challenge

To evaluate the impact of breakthrough infection on the magnitude and breadth of antibody responses, we measured neutralizing antibodies and RBD-specific IgG in serum samples collected at week 5 (pre-challenge) and week 9 (post-challenge) from hamsters immunized with S-2P:Al or S-2P:ISA720, as well as from PBS-control animals after challenge. These antibodies were assessed against the ancestral strain, Delta, and Omicron BA.1 variants.

Sera from convalescent unvaccinated (PBS-control) hamsters effectively neutralized ancestral SARS-CoV-2 and Delta pseudovirus, with titers comparable to those induced by S-2P:Al and S-2P:ISA720 vaccination. However, these sera exhibited minimal neutralization against Omicron BA.1, with only one sample showing low-level activity (Fig. 4A–C).

Fig. 4.

Analysis of antibody levels pre- and post-challenge. Serum samples were collected from immunized hamsters (n = 5 per group; S-2P:ISA720, S-2P:Alum, or PBS control) at week 5 (pre-challenge) and week 9 (3 weeks post-challenge). Serum RBD IgG antibodies and neutralizing antibodies were measured (the LLOD is 100-fold dilution). A-C) Neutralizing antibodies: Against authentic ancestral SARS-CoV-2 (A), Delta (B), and Omicron (C) pseudotyped viruses. D-F) Anti-RBD IgG: Against the ancestral (D), Delta (E), and Omicron (F) variant RBDs. (G) IgG antibody titers to Ancestral, Delta, and BA.1 RBD in hamster serum at week 5 and week 9. Abbreviations: LLOD, lower limits of detection; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; PBS, phosphate buffered saline; Wk, week; RBD, receptor-binding domain.

In the S-2P:Al group, challenge significantly boosted neutralization against the ancestral virus (5.1-fold, P < 0.05) and Delta (3.7-fold, not significant), but had limited effect on BA.1, only two of five sera showed low neutralization. No BA.1 neutralization was detected before challenge.

The S-2P:ISA720 group exhibited high neutralizing titers against the ancestral and Delta strains at week 5. Post-challenge, neutralization against both strains was only marginally increased (1.2-fold, P > 0.05), whereas neutralization against BA.1 was significantly enhanced (4.5-fold, P < 0.05). Notably, pre-challenge sera from all animals in this group neutralized BA.1, with a GMT NT₅₀ of 480.2.

Since the RBD is a major target of neutralizing antibodies [29], we also quantified RBD-specific IgG antibody levels. Overall, RBD-binding antibody profiles correlated with neutralization titers.

In S-2P:Al-immunized animals, challenge moderately increased RBD antibodies against ancestral and Delta viruses (not statistically significant) (Fig. 4D, E, G), but markedly increased anti-BA.1 RBD IgG by 10.3-fold (P < 0.05) (Fig. 4F, G). Despite this increase in binding antibodies, no corresponding enhancement in BA.1 neutralization was observed (Fig. 4C).

In contrast, the S-2P:ISA720 group showed a slight reduction in ancestral- and Delta-specific RBD IgG after challenge (Fig. 4D, E, G). Although anti-BA.1 RBD IgG did not increase significantly (Fig. 4F, G), a substantial rise in BA.1 neutralizing titers was detected (P < 0.05) (Fig. 4C).

These results demonstrate that homologous breakthrough infection enhanced neutralizing activity against the highly divergent BA.1 variant in S-2P:ISA720-immunized hamsters, but did not boost neutralization against the homologous or less divergent Delta strain. Conversely, the opposite pattern was observed in S-2P:Al-vaccinated animals.

3.5. Divergent antibody affinity maturation profiles following homologous virus challenge

Antibodies elicited by SARS-CoV-2 breakthrough infection have been reported to confer cross-variant neutralization through affinity maturation of IgG targeting conserved epitopes. To investigate whether similar mechanisms operate in vaccinated animals, we assessed antibody affinity maturation against the RBD and spike protein in serum samples collected from hamsters immunized with S-2P:Al or S-2P:ISA720, both before and after challenge with ancestral SARS-CoV-2. Avidity index, which reflects the overall strength of antigen–antibody binding, was quantified using a urea dissociation ELISA as previously described [27]. Briefly, immune complexes were treated with 8 M urea prior to incubation with an HRP-conjugated secondary antibody. As expected, urea treatment substantially reduced antibody titers compared to PBS-treated controls, consistent with the dissociation of low-affinity antibodies from the coated antigen, while high-affinity antibodies remained bound (Fig. 5A−C).

Fig. 5.

Analysis of antibody avidity pre- and post-challenge. Serum IgG antibodies against RBD and spike proteins at weeks 5 and 9 were measured under both PBS-treated and 8 M urea-treated conditions (to disrupt antigen–antibody complexes), and the avidity index was calculated. A–C) Levels of serum IgG antibodies targeting ancestral RBD, Omicron BA.1 RBD, and ancestral spike (The LLOD is 200-fold dilution). D–F) Avidity indices of serum IgG for ancestral RBD, Omicron BA.1 RBD, and ancestral spike. Abbreviations: SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; PBS, phosphate buffered saline; Wk, week; RBD, receptor-binding domain; LLOD, lower limits of detection.

At week 5, the S-2P:ISA720 group elicited significantly higher antibody titers against the ancestral RBD or spike protein than the S-2P:Al group, both with and without urea treatment. This difference was further accentuated after urea dissociation (6.1-fold vs. > 21.8-fold for RBD; 5.2-fold vs. 10.1-fold for spike) (Fig. 5A, C). These findings are consistent with the avidity index measurements (Fig. 5D, F), which indicate that the antibodies induced by S-2P:Al had significantly lower avidity than those induced by S-2P:ISA720, in addition to lower overall titers.

Following virus challenge, the S-2P:Al group showed moderate increases in antibody titers against RBD and spike (3.0-fold for RBD, not significant); 2.8-fold for spike, not significant) (Fig. 5A, C), along with a significant boost in antibody avidity for both antigens (Fig. 5D, F).

In contrast, the S-2P:ISA720 group displayed a distinct antibody response profile: despite showing no increase in anti-RBD or anti-spike IgG titers (with a slight decrease for RBD), it exhibited a statistically significant increase in avidity against spike (P < 0.01), with a non-significant increasing trend for RBD (Fig. 5D, F).

When assessed against BA.1 RBD, the S-2P:Al group exhibited minimal antibody reactivity at week 5, with three animals below the lower limit of detection (LLOD = 200). In comparison, all S-2P:ISA720-vaccinated animals remained seropositive, despite a 33.3-fold reduction in antibody levels relative to ancestral RBD. Following viral challenge, the S-2P:Al group exhibited a significant increase in total IgG against BA.1 RBD (P < 0.05), accompanied by a non-significant increase in avidity (Fig. 5B, E). Conversely, the S-2P:ISA720 group maintained stable total IgG antibody levels but demonstrated a significant increase in avidity.

These results suggest that homologous breakthrough infection in S-2P:ISA720-vaccinated animals enhances antibody avidity primarily toward conserved, cross-reactive epitopes within the RBD. In contrast, S-2P:Al-vaccinated animals show avidity maturation largely directed to strain-specific epitopes that lack cross-neutralizing activity against Omicron BA.1.

4. Discussion

The continuous evolution of SARS-CoV-2, compounded by immune imprinting, remains a continuing challenge to the development of broadly protective COVID-19 vaccines. Despite knowing that boosters and breakthrough infections induce broad cross-reactive immunity, and that longer intervals and increased antigenic distance lessen original antigenic sin [[30], [31], [32], [33], [34], [35]], we still poorly understand the extent to which the level of pre-existing immunity governs the magnitude of the recalled antibody response to subsequent challenges. Our study addresses this question using a controlled hamster model to examine how adjuvants, as pivotal modulators of primed immunity, influence not only protective efficacy but also the evolution of the antibody response following homologous viral challenge.

A key finding underscores the decisive role of adjuvant selection in shaping outcomes following breakthrough infection. Priming with S-2P:ISA720 not only elicited high initial neutralizing titers and conferred complete protection against disease, but also established an immune state predisposed to broad protection. Notably, only animals in this group demonstrated substantial cross-neutralizing activity against the antigenically distant Omicron BA.1 variant prior to challenge, indicating that the adjuvant used during priming critically influences the development of immune breadth. Certainly, the superior protective efficacy of S-2P:ISA720 immunization should also be attributed to its ability to induce a potent cellular immune response, which was significantly stronger than that elicited by S-2P:Al, as demonstrated in a mouse model [25].

In contrast, although S-2P:Al induced neutralizing antibodies against ancestral and Delta viruses, the titers were significantly lower and cross-neutralization against BA.1 was absent. Following challenge, the S-2P:Al group exhibited a strong quantitative boost in antibodies that were largely strain-specific, with minimal improvement in BA.1 neutralization, a pattern consistent with classic immune imprinting. This recall response was likely dominated by memory B cells targeting immunodominant, strain-shared epitopes, reflecting a pre-committed and narrow trajectory that allowed little flexibility for broadening. S-2P:Al-immunized hamsters showed higher challenge viral loads in the respiratory tract (suggesting more antigen exposure) than the S-2P:ISA20 group, but nevertheless failed to generate a stronger cross-reactive antibody response to BA.1, emphasizing the powerful role of pre-existing antibodies in directing immune recall.

Conversely, the S-2P:ISA720 group mounted a qualitatively distinct recall response. Instead of a substantial increase in homologous neutralizing titers, we observed marked affinity maturation toward conserved epitopes, as evidenced by significantly enhanced BA.1 neutralization and elevated RBD avidity. In contrast, neutralizing activity against the homologous strain showed only limited enhancement. These findings suggest that high-quality priming can redirect recall responses away from immunodominant epitopes and instead promote the refinement of cross-reactive B cell clones targeting subdominant but conserved regions, an outcome that appears to contradict classical antigenic sin.

We propose that this phenomenon may be explained by a mechanism of antibody-mediated epitope masking. Growing evidence suggests that pre-existing antibodies, particularly those targeting immunodominant epitopes, can suppress the booster response upon re-exposure, while dose not interfere with the response to subdominant epitopes [[36], [37], [38]]. High-affinity antibodies induced by S-2P:ISA720 against immunodominant epitopes may rapidly bind and mask these sites upon challenge, thereby preventing the reactivation of strain-specific B cell clones and their re-entry into germinal center reactions. This masking effect effectively shifts immune attention toward conserved but previously under-represented epitopes. If a pool of cross-reactive memory B cells has been established during priming, even those with low initial affinity, these clones may undergo affinity maturation upon encountering unmasked conserved epitopes, resulting in enhanced breadth without a substantial increase in total antibody titer.

The increased affinity for the prototype strain spike likely stems from combined enhancements in affinity for conserved epitopes within the RBD as well as those outside the RBD (such as in the NTD and S2 subunits), resulting in an overall improvement that reaches statistical significance. This suggests that antibodies targeting subdominant epitopes outside the RBD exhibit limited neutralizing activity, which is consistent with the observation that neutralizing antibodies are primarily concentrated in the RBD [29].

In the S-2P:Al group, however, the moderate antibody levels and affinity against strain-specific epitopes are likely insufficient to mediate effective epitope masking. Consequently, upon challenge, naïve B cells specific to subdominant conserved epitopes cannot effectively compete with memory B cells targeting immunodominant epitopes for antigen binding. This results in a narrow, recall-focused response amplified through antigenic sin.

The underlying mechanism likely stems from the adjuvant’s capacity to modulate germinal center reactions during priming. ISA720 promotes a Th1-biased response and robust naïve B cell activation, potentially fostering a more diverse and cross-reactive memory B cell pool [39]. Subsequent viral challenge can then recruit these clones for further affinity-driven refinement, consistent with our avidity data and the early neutralization activity observed after primary immunization.

These findings reshape the paradigm of immune recall: the outcome of the recall response is not predetermined but can be guided by both the quality (epitope specificity) and quantity (antibody titer) of the priming response. Effective adjuvants can thus promote immune plasticity, enabling the recall response to broaden—not merely boost—the immune repertoire.

This study has several limitations. First, the use of a homologous viral challenge, while enabling a clearer dissection of immune imprinting mechanisms, fails to replicate the heterologous immune pressures typically encountered in human populations. Second, the initial analysis overemphasized the impact of viral challenge on pre-existing immunity, consequently underestimating the contribution of time-dependent maturation to the humoral immune profile.

In conclusion, we have demonstrated that adjuvant-modulated priming fundamentally shapes the dynamics of antibody recall. Qualitative and qualitative priming is essential to establish a cross-reactive B cell reservoir, enabling viral re-exposure to serve as an opportunity for affinity maturation toward breadth. This model offers a new framework for developing broadly protective vaccines against SARS-CoV-2 and other antigenically variable pathogens.

Ethics statement

All animal experiments were conducted in strict accordance with the guidelines of the Chinese Regulations of Laboratory Animals (Ministry of Science and Technology of the People’s Republic of China) and Laboratory Animal Requirements of Environment and Housing Facilities (GB 14925–2010, National Standard of the People's Republic of China). All procedures were approved by the Animal Experiment Committee of Naval Medical University, China (20SWAQX23-004).