A Recombinant Subunit Vaccine induces a Potent and Broadly Neutralizing, Durable Antibody Response in Macaques against SARS-CoV-2 P.1 (Gamma) Variant

Abstract

FDA-approved and Emergency Use Authorized (EUA) vaccines using new mRNA and viral-vector technology are highly effective in preventing moderate to severe disease, however, information on their long-term efficacy and protective breadth against SARS-CoV-2 Variants of Concern (VOCs) is currently scarce. Here we describe the durability and broad-spectrum VOC immunity of a prefusion-stabilized spike (S) protein adjuvanted with liquid or lyophilized CoVaccine HT^™ in cynomolgus macaques. This recombinant subunit vaccine is highly immunogenic and induces robust spike-specific and broadly neutralizing antibody responses effective against circulating VOCs (B.1.351 [Beta], P.1 [Gamma], B.1.617 [Delta]) for at least 3 months after the final boost. Protective efficacy and post-exposure immunity were evaluated using a heterologous P.1 challenge nearly 3 months after the last immunization. Our results indicate that while immunization with both high and low S doses shorten and reduce viral loads in the upper and lower respiratory tract, a higher antigen dose is required to provide durable protection against disease as vaccine immunity wanes. Histologically, P.1 infection causes similar COVID-19-like lung pathology as seen with early pandemic isolates. Post-challenge IgG concentrations were restored to peak immunity levels and vaccine-matched and cross-variant neutralizing antibodies were significantly elevated in immunized macaques indicating an efficient anamnestic response. Only low levels of P.1-specific neutralizing antibodies with limited breadth were observed in control (non-vaccinated but challenged) macaques suggesting that natural infection may not prevent reinfection by other VOCs. Overall, these results demonstrate that a properly dosed and adjuvanted recombinant subunit vaccine can provide protective immunity against circulating VOCs for at least 3 months.

Short Description:

A recombinant subunit spike protein vaccine formulated with low or high antigen doses and liquid or lyophilized CoVaccine HT^™ reduces and shortens the duration of P.1 (Gamma) VOC viral replication in the upper and lower respiratory tracts of cynomolgus macaques 3 months after the final boost, although high antigen doses are required to completely protect against lung pathology. P.1 viral challenge restores peak spike-specific IgG levels and significantly elevates cross-variant neutralizing antibodies in all vaccinated animals, while control macaques (non-immunized, but challenged) only generate low antibody levels with limited breadth of neutralization.

COVID-19 is a respiratory disease caused by the Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) ^{1, 2} and is commonly characterized as an asymptomatic infection or a self-limiting, febrile illness co-presenting with cough, shortness of breath, and fatigue ^{3, 4}. Severe complications can result in approximately 6-10% of infected patients ⁵ developing pneumonia, acute respiratory distress syndrome, multi-organ dysfunction and arterial thromboembolic events that can result in hospitalization and/or death ^6–9. Transmission occurs through respiratory droplets during close contact with asymptomatic or presymptomatic infected individuals, as well as symptomatic patients ^10–12. The synergistic relationship between the infectivity and high transmissibility with low virulence has contributed to the ongoing global public health crisis, and the emergence of more transmissible and potentially more pathogenic variants of concern (VOCs). Longitudinal cross-sectional studies have indicated only a fraction of asymptomatic to mild and moderate SARS-CoV-2 natural infections of humans results in detectable neutralizing antibodies months after recovery ^13–15 and current evidence suggests that convalescent immunity may only provide transient protection ^16–18. Furthermore, the degree of immune evasion witnessed in vitro by spike (S) protein variants and the increasing occurrence of reinfection ^19–23 has raised the question whether previous infection or vaccine-derived immunity can provide heterologous protection against further transmission and limit additional mutation of circulating VOC, even while protection against severe disease and mortality has been sustained to date.

Some VOCs show increased transmissibility, cause more severe disease than original strains and demonstrate greater immune evasion to neutralizing antibodies as they harbor mutations at sites facilitating viral fusion and at critical epitopes in the N-terminal (NTD) and receptor binding domains (RBD) ^{24, 25}. The B.1.1.7 (Alpha), B.1.351 (Beta) and P.1 (Gamma) variants contain the same set of mutations, K417T, E484K, and N501Y, while the B.1.617 (Delta) carries L452R and T478K mutations, all of which are associated with enhanced infectivity and reduced serum neutralization ^26–28. These antigenic changes have abolished potent neutralizing epitopes targeted by monoclonal antibody therapeutics ²⁹ and antibodies elicited by mRNA vaccination ³⁰. Natural reinfection with the VOC strains after original strain infection, and vice versa, has been documented ^{20, 31–34}. Despite this, cross-neutralization of convalescent sera from patients infected with non-variants and mRNA immunized individuals suggests that the parental spike (S) protein in current vaccines affords some degree of protection ^35–37.

Phase 1-3 clinical trials of novel mRNA (Moderna mRNA-1273 and Pfizer-BioNTech BNT162b2) and viral-vector (Johnson & Johnson-Janssen Ad26.COV2.S) platforms, utilizing the original Wuhan-Hu-1 S protein ^38–40, have shown these vaccines to be highly protective with 94-95% and 67% efficacy against COVID-19 disease, respectively^41–43, and they all elicit uniformly high protection against severe disease/death (i.e., >85%). Post-hoc sequencing during clinical trials of these and other vaccine platforms in variant-dominant regions have revealed a slightly lower efficacy against B.1.1.7 and B.1.351 variants, albeit usually remaining above the 50% protection threshold ^{44, 45}. Attempts at predicting protection based on immune correlates, such as cross-neutralizing antibodies, have shown that vaccinated individuals had a 2 to 8-fold reduction in titers against current VOCs as IgG levels wane, which are nonetheless believed to confer a degree of resistance against infection ^46–49. However, these samples were taken during the peak period of the humoral response and may not be accurate predictors of protection months onwards when the likelihood of infection is greater. Thus, understanding the decay in antibody responses months after the final boost is an essential endpoint in vaccine development.

Evaluating a delayed vaccine response in non-human primate (NHP) models can provide valuable foresight into the relationship between waning immunity and protection, especially against VOCs. In this study, we describe the immunogenicity and protective efficacy of a recombinant pre-fusion spike subunit (based on the reference SARS-CoV-2 strain, Wuhan-Hu-1) adjuvanted with liquid or lyophilized (dry) CoVaccine HT^™, a squalane-in-water nanoemulsion adjuvant containing immunostimulatory sucrose fatty acid sulfate esters ⁵⁰, in cynomolgus macaques. Both zeta potential and emulsion droplet size of CoVaccine HT^™ are maintained when reconstituted with distilled water ⁵¹. Other investigators have shown that infection with SARS-CoV-2 (prototype strain) produced COVID-19-like disease in this species and reflects similar viral shedding kinetics and lung pathology as human infections ^{52, 53}. We have previously demonstrated that our vaccine candidate elicits a broad-spectrum IgG response including high neutralizing antibody (NtAb) titers against the prototypic SARS-CoV-2 and VOCs, specifically B.1.351 and P.1, and an in vitro antigen-specific IFN-γ secreting response from immune splenocytes taken from Swiss Webster mice ⁵⁴. To elaborate on our understanding of the elicited immunological response, we assessed the vaccine efficacy using a delayed challenge scheme with the P.1 VOC to delineate how the maturation and/or decay of the humoral response affects protection from a heterologous SARS-CoV-2 strain. Neutralizing antibody responses to the vaccine-matched WA1/2020 strain, and VOCs, B.1.351, P.1 and B. 1.617 were also determined just prior to challenge and 14 days post-challenge (to assess anamnestic responses). Furthermore, we characterize the viral kinetics and histopathological changes post-challenge in the lower and upper respiratory tracts of control (non-vaccinated, challenged), and protected NHPs. To our knowledge, this is the first published account of P.1 VOC challenge in NHPs.

RESULTS

Spike antigen with either liquid or lyophilized CoVaccine HT^™ elicits a durable humoral response

A stably transformed Drosophila S2 cell line, without subculturing, yielded approximately 30mg/L of the pre-fusion spike protein. Purification using CR3022 mAb immunoaffinity chromatography did not negatively affect the structural integrity of the purified protein (Fig. S1A) and size-exclusion analysis revealed that the spike antigen is primarily homogenous (Fig. S1B).

To evaluate immunogenicity, twelve Cynomolgus macaques were assigned into four groups (n=3 per group) and received two immunizations with 5 (Group A) or 25 (Group B and C) μg liquid prefusion spike trimer (S) antigen formulated with either lyophilized (Group A and C) or liquid (Group B) CoVaccine HT^™ adjuvant, or one dose of a co-lyophilized CoVaccine HT^™-adjuvanted control containing an unrelated viral glycoprotein antigen (Group D) (Fig. 1A). The two doses were administered within a three-week interval and all NHPs were challenged with the P.1 isolate 12 weeks post-boost (~ 3 months, study week 15). Wuhan-Hu-1 S- and RBD-specific IgG titers were measured by a multiplexed microsphere immunoassay (MIA) using insect cell expressed antigens coupled onto spectrally distinct, magnetic beads as described previously ⁵⁴. Serum S-specific IgG concentrations were interpolated using a standard curve generated from S-specific human IgG purified from vaccinated individuals (Fig. 1B). RBD-specific IgG titers were read out as median fluorescence intensity (MFI) (Fig. S2). All NHP immunized with the adjuvanted S at both antigen doses seroconverted after the prime (week 3) with S-specific antibodies in the range of 20 to 70 μg/mL, and peak serum IgG concentrations detected two weeks after the boost (week 5) in the range of 70 to 753 μg/mL. Macaques given a 25 μg dose of S demonstrated a greater IgG response to the antigen compared to those receiving 5 μg. RBD-specific IgG titers followed a similar trend. As expected, animals in group D receiving an unrelated antigen did not develop any detectable S-specific IgG during this phase of the study. S-specific IgG remained detectable 12 weeks after the boost (week 15) although IgG concentrations dropped 3.0 to 9.9-fold relative to the prior peak titer.

Figure 1. Vaccine Scheme, IgG and Neutralizing Antibody Kinetics.

(A) 12 cynomolgus macaques were separated into 4 groups and given either 5 or 25 μg S protein formulated with either liquid or reconstituted, lyophilized CoVaccine HT^™ adjuvant. Two doses were administered IM 3 weeks apart and serum was collected at indicated time points. Macaques were challenged IN and IT with a total of 1x10⁶ TCID₅₀ of the SARS-CoV-2 P.1 strain. (B) Serum Wuhan-Hu-1 S-specific IgG kinetics measured using a MIA with purified, human S-specific IgG standards to estimate serum concentration. Each sample was diluted 1:5,000 and plotted as the average of duplicate measurements. Duplicate values varied <10% of each other. Dashed line indicates the limit of quantification (LOQ, ⩽4 μg/mL). Individual values falling below the LOQ were set to ½ LOQ. Significance was calculated using a one-way ANOVA followed by a Dunnett’s Multiple Comparison to Group D at each time point. (C) WA1/2020 neutralizing antibody kinetics were measured using a WT SARS-CoV-2 PRNT assay. Curve-fitted PRNT₅₀ titers were calculated using a sigmoidal dose response curve. Dashed line indicates the limit of detection (LOD, ⩽1:20). Individual values falling below the LOD were set to ½ LOD. Significance was calculated using a Kruskal-Wallis Test followed by a Dunn”s Multiple Comparison. (*p⩽0.05, **p⩽0.01).

The titer of NtAb against the prototype WA1/2020 strain was determined using a standard PRNT with wild-type virus (Fig. 1C). NtAb from one animal in Group C was detectable after the prime (week 3). All S-immunized NHPs developed a potent neutralizing response, peaking two weeks after the boost (week 5) and generally remaining stable one week later (week 6). The group receiving 5 μg of the antigen showed the greatest variability. Nearly all immunized NHPs maintained neutralization capacity, with PRNT₅₀ greater than 1:150 dilution, 12 weeks after the final immunization (week 15). NtAbs from groups receiving 25 μg S either remained stable or decreased up to 3.0-fold, while the group receiving 5 μg S declined ~2.9-9.2-fold. A similar reduction in serum neutralization was also verified using a surrogate rVSV-SARS-CoV-2 S PRNT assay (Fig. S3).

A recombinant subunit vaccine induces stable neutralizing antibodies against VOCs

Circulating SARS-CoV-2 VOCs can evade vaccine-induced antibody responses and are associated with breakthrough infections in those fully immunized, especially as antibody titers wane ^55–61. To determine whether an adjuvanted, prototypic Wuhan-Hu-1 S subunit can generate durable cross-variant neutralizing antibodies, PRNTs using WT B. 1.351, P.1, and B. 1.617 isolates were determined with sera collected at the time of peak neutralization and 12 weeks after the final immunization, a timepoint at which antibody titers are expected to have waned. At week 6, potent neutralization of the B. 1.351, P.1 and B. 1.617 VOC was detected, although 10.7-, 10.7- and 5.7-fold lower, respectively, compared to WA1/2020 in all vaccine groups (Fig. 2A). The decrease in neutralization against the B. 1.351 and B. 1.617 is consistent with reports of titers in sera from NHPs and humans immunized with mRNA vaccines that show a ~8- and ~5-reduction, respectively ^62–67. As antibody titers waned 10 weeks later (study week 15) the gap between neutralization titers against the WA1/2020 strain to the VOC isolates narrowed to a 4.5-, 5.3- and 5.1-fold difference, respectively (Fig. 2B), driven by a proportionally larger decline in WA1/2020-specific neutralizing activity. At this later time point, all S-immunized macaques, except for a single macaque in Group A, maintained a detectable PRNT₅₀ titer greater than 1:40. A few macaques in Groups B and C demonstrated greater VOC neutralization at week 15 compared to week 6, suggesting a refinement or “maturation” of the humoral response towards neutralizing epitopes during this interval. Overall, 25 μg S protein generated a more durable and greater variant neutralizing response than the lower antigen dose.

Figure 2. Cross-Neutralizing Antibody Titers.

Neutralizing antibody titers generated by 5 or 25 μg of spike protein (S) formulated with either lyophilized (Lyo) or Liquid (Liq) CoVaccine HT (CoVac) at (A) week 6 and (B) week 15 were measured using a WT SARS-CoV-2 PRNT assay with the WA1/2020, B.1.351 (Beta), P.1 (Gamma), B.1.617 (Delta) VOC Curve-fitted PRNT₅₀ titers were calculated using a sigmoidal dose response curve. Dashed line indicates the LOD (≤1:20). Individual values falling below the LOD were set to ½ LOD. Significance was calculated using a Kruskal-Wallis Test followed by a Dunn’s Multiple Comparison to Group D at each time point. Horizontal bars within the symbols mark the GMT of each group. Horizontal bars above the groups of symbols indicate the group comparisons which showed statistically significant differences (*p≤0.05).

Immunization with recombinant subunits reduces viral burden from delayed P.1 (Gamma) challenge

The characterization of the prototype strain and B.1.351 variant infections in NHP models has been described previously ^{52, 53, 68–70}. Here we characterize the clinical signs and histological events of a P.1 VOC infection in cynomolgus macaques and determine whether a prefusion spike subunit formulated with CoVaccine HT^™ is effective at reducing viral load against a circulating VOC after a 12-week interval following the final immunization. All NHPs were challenged with a total of 1x10⁶ TCID₅₀ of the P.1 isolate using simultaneous intranasal and intratracheal inoculation routes. None of the NHPs challenged developed visible clinical signs of respiratory disease throughout the study, consistent with WA1/2020 strain infected cynomolgus macaques⁵². No changes in responsiveness, respiratory rate, respiratory effort, nasal or ocular discharge, nor cough was documented during the observation period. Core body temperature also did not change. A slight, and sometimes intermittent, decrease in weight was observed in some macaques across all experimental groups, however, most returned to near initial weight by day 14 (Fig. S4A–B). A single control macaque in Group D had sustained weight loss throughout and at the end of the observation period. Weight change did not correlate with a higher viral load nor degree of lung histopathology.

Bronchioalveolar lavages (BAL) from the lower respiratory tract, and nasal and oral swabs (NS and OS respectively) from the upper respiratory tract were collected at days 2, 4, 7, 10 and 14 after challenge to detect infectious virus, genomic RNA or signs of viral replication. High levels of infectious virus, 5.8 and 5.0 Log₁₀ TCID₅₀ titer, were recovered two days post-challenge from the NS and BAL, respectively, from Group D control macaques (Fig. 3A & 3D and S5A–B). Infectious virus continued to be detectable in both anatomic sites until at least day 7 post-challenge before becoming undetectable on day 10. All S-immunized NHPs presented TCID₅₀ titers at least 1-2 Log₁₀ lower than control macaques in both locations at all timepoints throughout the study, indicating a degree of protection conferred by both 5 and 25 μg S formulations of the subunit vaccine. Low levels of infectious virus in the BAL were recovered on day 2 in two out of the three macaques from both Groups A and B, while no culturable virus was recovered from all Group C macaques at any timepoint throughout the study. Except for a single macaque in Group B on Day 4 only, no virus could be cultured from the BAL of any S-immunized macaques from this timepoint onwards. Infectious virus from the NS was recoverable from immunized macaques throughout the study period and was variable between groups, however these macaques on average had TCID₅₀ at least 1-2 Log₁₀ below those of control macaques. Only low levels of virus could be recovered on day 7 from two macaques from Group C before dropping to undetectable levels at day 10. No virus could be seen in Groups A and B from day 7 onwards.

Figure 3. Viral Load Kinetics after P.1 Challenge.

Mean ± SEM of the (A & D) TCID₅₀ titers; (B, E, G) viral genomic; and (C, F, H) subgenomic N RNA copies from the (A, B, C) bronchoalveolar lavage (BAL); (D, E, F) nasal swab (NS); and (G, H) oral swab (OS) collected at the time points indicated. Dashed line indicates the assay-specific LOD.

Quantitative RT-PCR was used as an additional measurement for viral load. We detected the presence of viral RNA and viral replication throughout the 14-day study period in both the lower (Fig 3B–C, S6A–C and S7A–C) and upper (Fig 3E–F, S6D–F and S7D–F) respiratory airways and from the oral cavity (Fig 3G–H, S6G–I and S7G–I). Quantification of viral RNA showed that both 5 and 25 μg immunization had reduced the viral load by approximately 1-2 Log₁₀ RNA copies/mL (or swab) over Group D control macaques, corroborating the TCID₅₀ results. Viral replication in most immunized macaques was low or undetectable by day 7 and 10 in the BAL and NS, respectively, which is 3-4 days earlier compared to control macaques. Similarly, replication in the oral cavity was undetectable at least 3 days earlier in most S-immunized macaques than in the control group, suggesting a reduced potential for transmission. By contrast, high levels of viral RNA and replication, between 5-6 Log₁₀ RNA copies/mL, were observed in these control macaques before gradually resolving at day 10. Altogether, reduced viral loads and earlier clearance at all tested anatomic locations in Groups A, B and C macaques indicate a degree of durable protection from P.1 VOC infection provided by an adjuvanted protein subunit vaccine.

High dose spike antigen reduces lung histopathology caused by P.1 (Gamma) challenge

To evaluate vaccine efficacy against COVID-19-like pathology, a section slide from each lung lobe and the bronchi was prepared from each macaque and stained with hematoxylin and eosin to survey for histopathological changes indicative of acute lung injury. The histopathological scoring system based on the presence/severity of edema, intraalveolar and interstitial inflammation, perivascular lymphocytic cuffing, and increased bronchiolar-associated lymphoid tissue (BALT) is outlined in Table S1. A cumulative average score was determined for each macaque based on the evaluation of these five characteristics per section for a total of 30 scores (Fig 4A). Significant differences in the cumulative average scores for each vaccine formulation vs. controls were determined using a one-way ANOVA followed by a Dunnett’s multiple comparison test. Macaques in the control Group D appeared to have developed mild to moderate respiratory disease. The lower cumulative GMT scores of Groups B and C, immunized with 25 μg S, suggests that these animals, except for a single macaque, were protected from lung pathology while all Group A (immunized with 5 μg S) macaques appeared to have developed mild disease despite exhibiting lower viral loads. Week 15 WA1/2020 NtAb levels were determined to be inversely correlated, although weakly, with histopathology scores using a Spearman’s correlation test (r=−0.682, p=0.0178) (Fig. 4B). Week 15 pre-challenge neutralizing antibody levels were lower in animals with higher histopathology scores, a trend seen with breakthrough infections in fully immunized individuals ⁵⁶. The moderate disease seen in the control animals is demonstrated in Fig. 5A, where a representative, low-power magnification shows a distribution of lymphocytic perivascular cuffing and clusters of intraalveolar macrophages. In contrast, immunized animals (especially those receiving the higher antigen dose), appeared to have milder disease without these pathological changes (Fig. 5B). For reference, a lung section from an uninfected macaque is shown in Fig. S8. Macaques developing mild to moderate disease exhibited absent to moderate, focal and multifocal edema and perivascular cuffing with marked lymphocytic infiltration (Figs. 6A–D). Syncytia of intraalveolar multinucleated giant cells surrounded by acute and eosinophilic inflammatory infiltrate (Fig. 6A.) or increased intraalveolar macrophages with interstitial lymphocytic inflammation, in addition to alveolar septal thickening and complete perivascular cuffing can be seen in a moderately diseased, control macaque (Fig. 6B–C). Proteinaceous edema fluid filling alveolar spaces was also observed in unprotected macaques (Fig. 6D). While these histopathological changes were also noticed in protected macaques, the findings appeared less severe (Fig. 6E–F). A section from a protected macaque revealed a few intraalveolar multinucleated giant cells without an increase in intraalveolar macrophages, lymphocytic vascular cuffing, or interstitial inflammation (Fig. 6E). Similarly, only partial lymphocytic perivascular cuffing with adjacent alveolar epithelial hyperplasia was noticed without an increased intraepithelial macrophage infiltrate and normal alveolar septal thickness in immunized macaques receiving 25 μg of S (Fig. 6F).

Figure 4. Histopathology Score and Correlation with Neutralizing Antibodies.

(A) Cumulative average histopathology score was determined by averaging the scores of six sections cut from each lobe and the bronchi. The presence and severity of edema, intra-alveolar inflammation, perivascular cuffing, increased BALT and interstitial inflammation was determined and assigned scores ranging from 0-3. Significant differences in lung histopathology were calculated using a one-way ANOVA followed by a Dunnett’s multiple comparison to Group D with the scores from each section (30 scores per macaque) as replicates. Horizontal bars mark the GMT of each group. (**p≤0.01, ****p≤0.0001) (B) Correlation between histopathology scores and WA1/2020-specific PRNT₅₀ titers were determined using a Spearman’s correlation test.

Figure 5. Histopathology observed in Hematoxylin and Eosin-Stained Lung Sections of Control Cynomolgus Macaques Compared to Immunized Macaques after P.1 challenge.

(A) Representative low power view of a lung section showing distribution of lymphocytic perivascular cuffing (arrows) and clusters of intraalveolar macrophages (arrowhead) in an unprotected macaque. (B) Representative low power view of a lung section from an immunized macaque which did not exhibit pathologic changes. Scale Bar = 500 microns.

Figure 6. Histopathology Observed in Control Cynomolgus Macaques Compared to that Observed in Immunized Macaques after P.1 Challenge.

(A-D) unprotected macaques. (A) A syncytium of intraalveolar multinucleated giant cells is seen surrounded by an acute and eosinophilic inflammatory infiltrate. (B) Increased intraalveolar macrophages and an intraalveolar multinucleated giant cell accompanied by interstitial lymphocytic inflammation and thickening. (C) Complete lymphocytic perivascular cuffing with increased intra-alveolar macrophages. (D) Proteinaceous edema fluid in alveolar spaces. (E-F) Protected macaques. (E) Intraalveolar multinucleated giant cells (arrows) without increased intraalveolar macrophages, lymphocytic vascular cuffing or interstitial inflammation. (F) Partial lymphocytic perivascular cuff (arrow) with adjacent alveolar epithelial hyperplasia (arrowhead) without alveolar septal thickening. A-C, E&F : Scale bar = 10 microns. D Scale bar = 20 microns.

P.1 challenge recalls a broadly neutralizing anamnestic response in immunized macaques and a variant-specific primary neutralizing response in naïve macaques

To better understand the immunological basis of events happening after late challenge (12 weeks post booster immunization), and to determine if P.1 challenge triggered an anamnestic response in immunized NHPs, post-challenge S-specific IgG concentrations and PRNT₅₀ titers were measured from sera collected upon necropsy at day 14 after challenge. Late P.1 challenge significantly boosted serum IgG concentrations to levels attained at the post-immunization peak at Week 6, with less variance between animals, suggesting that a recall response was triggered by the heterologous infection (Fig. 7A). Vaccine-matched (WA1/2020) and VOC neutralizing (i.e., B.1.351, P.1 and B.1.617) antibody levels were also all significantly boosted to levels attaining or surpassing the peak levels generated through initial immunization by 5-fold and upwards of 20-fold respectively (Fig. 7B and S3). Unsurprisingly, the greatest increase from pre-challenge neutralization was against the P.1 challenge isolate. Group A showed slightly higher P.1 NtAbs compared to Group B and C macaques and was the only group to have significantly higher titers across different variants. The lower degree of histopathology seen in Groups B and C may indicate that 25 μg S with either form of CoVaccine HT^™ provides greater long-term protection. In Group D control macaques, P.1 challenge generated low levels of NtAbs compared to immunization alone, and it was only cross-reactive with the B.1.351 isolate and not to WA1/2020 nor B.1.617 isolates. This suggests that primary infection with one isolate of SARS-CoV-2 does not generate broad immunity against other variants in contrast to vaccine derived immunity, which is more broadly cross-reactive with VOCs, even months after peak immunity is observed.

Figure 7. Post-Challenge Anamnestic Response.

(A) Vaccine-matched S-specific IgG conc. from sera collected at week 6, week 15 and at 14-days post-challenge, measured by MIA. IgG concentrations were estimated using a human S-specific IgG standard as described previously. Horizonal bars mark the GMT for each group. Dashed line indicates the LOQ (≤4 μg/mL) Individual values falling below the LOQ were set to ½ LOQ (B) Neutralizing antibodies 14 days after P.1 challenge were measured using a WT SARS-CoV-2 PRNT assay with the WA1/2020, B.1.351 (Beta), P.1 (Gamma), B.1.617 (Delta) VOC. Curve-fitted PRNT₅₀ titers were calculated using a sigmoidal dose response curve. Dashed line indicates the LOD (≤1:20). Individual values falling below the LOD were set to ½ LOD. Significance was calculated using a Kruskal-Wallis Test followed by a Dunn’s Multiple Comparison to Group D at each time point. (*p≤0.05, **p≤0.01).

DISCUSSION

Understanding the durability and breadth of vaccine-generated immunity is critical for predicting long-term protection during the COVID-19 pandemic and for informing policy on strategic resource deployment to facilitate equitable vaccine access. Diversifying the type of vaccines currently used beyond mRNA and viral vectors to include other vaccine platforms, such as thermostable protein subunit vaccines ^71–73, can bolster global availability by mobilizing more vaccines to resource-poor areas, or overcoming anti-vector immunity and adverse effects, by using this platform as a booster to restore pre-existing natural or vaccine-induced immunity ^74–76.

We have demonstrated that two doses of a subunit vaccine consisting of a prefusion S trimer (Wuhan-Hu-1) formulated with liquid or lyophilized CoVaccine HT^™ reduces viral load and provides sufficient, durable, cross-variant protection from mild to moderate disease lasting at least 3 months after the final boost. This supports previous findings in rhesus macaques that S trimer-based subunit protein vaccines are highly immunogenic and generate long-lasting, robust antibody responses ⁷⁷. Furthermore, using a two-step purification method of immunoaffinity chromatography followed by a size-exclusion polishing step results in a homogenous antigen composition and reduces host-cell carryover plaguing conventional purification strategies for protein subunits ⁷⁸. Also promising is the prospect that the adjuvant, CoVaccine HT^™, can be lyophilized, and reconstituted simply with water for injection and still retain functionality. This technology has been previously used to develop mono- and multivalent filovirus vaccine formulations. Those studies demonstrated that single-vial, lyophilized formulations preserved the higher order antigen structure and the biophysical properties of the adjuvant ⁵¹. Robust levels of S-specific IgG and homologous, as well as B.1.351, P.1 and B.1.617 cross-neutralizing antibodies, were detected throughout the extended study period and inversely correlate with viral load and lung damage. Neutralization against the B.1.351 and P.1 VOC were equivalently reduced compared to the WA/2020 isolate immediately following the booster and suggests that E484K and N501Y may play a critical role in antibody evasion early on, however, this reduction equilibrates at the 3-month time point across all VOCs tested while still affording sufficient protection. This report supports growing evidence that the Wuhan-Hu-1 S peptide sequence encoded by highly effective mRNA, viral vectored, and subunit protein vaccines, generates immunity that affords protection against circulating VOCs ^{45, 46, 51, 79, 80}, even months after the final vaccine dose when antibody titers are waning. Furthermore, we noticed a few macaques receiving the higher antigen dose had developed increasingly potent cross-variant neutralizing titers immediately prior to viral challenge which were greater than titers observed during the peak humoral response. This suggests that a potentially remaining antigen depot or persistence of APC’s may foster continued accumulation of somatic mutations and affinity maturation in memory B cells ^{81, 82} beyond the initial vaccination phase.

This study is also the first to describe the course of infection and histopathology of a SARS-CoV-2 P.1 infection in a NHP model. Consistent with what was described in WA1/2020 infected cynomolgus macaques ^{52, 53}, macaques challenged with the P.1 isolate did not develop elevated temperature or decreased weight, nor did they show observable signs of respiratory disease. Viral load in the upper and lower respiratory tract peaked early during infection two days after challenge and gradually decreased to undetectable sgRNA levels by day 14 in unprotected animals, similar to viral kinetics described in another study using the same inoculating dose (1x10⁶ TCID₅₀), but with an early pandemic isolate ⁵³. Although P.1 is estimated to be upwards of 2.4-fold more transmissible ²⁶, it is unclear whether the higher viral load observed during peak P.1 replication in this study compared to time-matched reports for a non-variant isolate is a characteristic of P.1 replication or a discrepancy between readouts from different RT-PCR procedures. However, we show that immunization with both 5 and 25 μg doses of S effectively reduces viral load in both upper and lower respiratory tracts implying a lower likelihood of viral transmission even in mildly symptomatic macaques. Histopathological examination of lung and bronchial sections in unprotected macaques confirmed mild to moderate disease, which was abrogated in immunized individuals. COVID-19 like disease observed in this model consisted of increased intraalveolar and interstitial infiltration, as well as giant syncytial cells, similar to previous observations ^{52, 64, 83}. While both high and low doses of S antigen generate statistically negligible differences in humoral responses during peak immunity, it is clear from our findings that higher antigen doses, at least with CoVaccine HT^™ produces a more durable and protective response at later timepoints.

Late P.1 challenge induced an anamnestic response in all immunized macaques that restored S-specific antibody titers to peak serum concentrations seen shortly after the booster immunization, and significantly enhanced vaccine-matched and VOC-neutralizing antibody titers. In most macaques receiving 25 μg S, the immune recall occurred with little to no disease. The durability and cross-variant neutralizing nature of immune responses generated by mRNA and viral vectors has been documented up to six months or later ^{63, 84–86}. However, breakthrough infections, particularly with variant strains, have been reported even in the presence of detectable and high-level NtAbs ^{55, 57, 60, 63, 87} Anamnestic responses in fully immunized NHPs have been characterized previously with homologous and B.1.351 variant challenge and agree with our observations that viral challenge boosts functional antibody levels. Our analysis shows that viral challenge with a heterologous strain not only rapidly boosts homologous and challenge strain specific NtAbs, but also NtAbs against the unencountered variants, B.1.351 and B.1.617, underpinning the broad-spectrum potential of subunit protein vaccines. In contrast, control macaques, receiving no S antigen but infected with P.1 strain, developed moderate levels of P.1 NtAbs that are only cross-neutralizing with B.1.351, and barely neutralize WA1/2020 and B. 1.617, suggesting that natural infection with one strain of SARS-CoV-2 may confer only limited protection against other VOCs. Of course, further affinity maturation could not be observed here as the study ended 14 days after challenge.

The small size of our treatment groups (n=3) is a limitation of our study and may therefore not provide enough statistical power to strongly correlate antibody concentrations to viral load or histopathology score. Furthermore, we cannot directly compare the outcome of our study to other studies that evaluate vaccine efficacy soon after the final immunization when immunity is greatest, as vaccine efficacy is known to decline over time. Likewise, we could not benchmark the efficacy of our vaccine formulation when the immunity is greatest at week 5, however, our late challenge scheme accurately reflects the current urgent need for decisions about timing and candidates for possible booster vaccinations during the pandemic situation as it provides useful information regarding the real-life durability and breadth of vaccine protection. Finally, while our results and evidence from other COVID-19 vaccine serological studies demonstrate that the prototypic spike sequence confers sufficient protection against Beta, Gamma and Delta VOC^{30, 46, 47, 65, 67, 79} protection against highly divergent variants such as the Omicron variant may require a second booster for greater coverage^88–92. Further investigation is required to determine the full utility of protein vaccines to serve as prime or boost in combination with other vaccines for this purpose.

In conclusion, we show that a two-dose regimen of a prefusion-stabilized trimeric S subunit protein vaccine formulated with lyophilizable CoVaccine HT^™ adjuvant reduces viral burden and high antigen doses can confer durable cross-variant immunity. Future efforts will therefore focus on developing a thermostabilized vaccine formulation in a single-vial format, potentially enabling facile worldwide distribution.

METHODS

Ethical Statement

The investigators adhered fully to the “Guide for the Care and Use of Laboratory Animals” by the Committee on Care of Laboratory Animal Resources Commission on Life Sciences, National Research Council. Cynomolgus macaques (Macaca fascicularis) were housed at BIOQUAL Inc. (Rockville, MD). All macaque experiments were reviewed and approved by BIOQUAL’s Animal Care and Use Committee. BIOQUAL Inc. is accredited by the American Association for Accreditation of Laboratory Animal Care (AAALAC).

Study Design

Cynomolgus macaque studies were performed using three adjuvanted vaccine formulations and an adjuvanted control using unrelated antigens: A) 5 μg of SARS-CoV-2 S protein with 5 mg lyophilized CoVaccine HT^™, B) 25 μg S protein with 5 mg liquid CoVaccine HT^™, C) 25 μg S protein with 5 mg lyophilized CoVaccine HT^™, and D) 25 μg of Ebola Virus (EBOV) glycoprotein with 5 mg CoVaccine HT^™, co-lyophilized in one vial. CoVaccine HT^™ (BTG International Ltd, London, United Kingdom) was lyophilized as previously described ⁵¹ and reconstituted with sterile water before mixing with the antigen. Each group consisted of both male and female cynomolgus macaques (n = 3 for each formulation) weighing between 2.8 and 8.2 kg. Cynomolgus macaques of groups A-C were immunized intramuscularly (IM) in both deltoids (split dose) at weeks 0 and 3, control animals were immunized only once at week 0. Pre-challenge sera were collected at weeks 0, 3, 5, 6, 15. Challenge at week 15 with SARS-CoV-2 P.1 variant, hCoV-19/Japan/TY7-501/2021, TY7-501 (BIOQUAL-generated stock [lot no. 031921-1215] in Calu-3 cells from seed stock no.TY7-501 was performed with an inoculum dose of 5x10⁵ TCID₅₀/mL administered to each animal in volumes of 1 mL by intratracheal and intranasal injection at each site. Nasal and oral swabs (NS and OS respectively) were collected on days 2, 4, 7, 10 and 14 after challenge. Bronchoalveolar lavage (BAL) samples were collected on days 2, 4, 7 and 10 after challenge. Necropsies and lung tissue collection were performed at the endpoint of the study at day 14 post-challenge. Samples were immediately processed and subsequently stored at −80°C prior to analysis. S-specific serum IgG concentrations were determined using a microsphere immunoassay (MIA; see below). NtAbs were measured using wild-type and rVSV-SARS-CoV-2 S PRNTs (see below). Viral load in the NS, OS and BAL was measured using plaque assays and quantification of genomic and subgenomic N transcript RNA (see below). Lung and bronchi tissues were processed, and H&E stained to delineate histopathological signs of disease in each animal.

Recombinant protein expression and purification

The plasmid was generated to express the pre-fusion, protease-resistant, trimeric transmembrane (TM)-deleted spike (S) glycoprotein from SARS-CoV-2 as described previously⁵⁴. Modifications to the gene include the removal of the furin and S2’ cleavage site, the addition of 2 prolines between the heptad repeat 1 and central helix region, and a foldon trimerization domain. Briefly, Drosophila S2 cells were transfected with the S gene using Lipofectamine LTX with PLUS reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions using ExCell420 media (Sigma-Aldrich, St. Louis, MO). Serial passage of the cell culture into fresh media spiked with 300 μg/mL hygromycin B generated a stably transformed cell line. To verify S protein expression, transformants were induced at a density of 1x10⁶ cells/mL with culture medium containing 200 μM CuSO₄ and the supernatant was analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and western blotting one week later. To scale up protein expression, the cell culture was expanded to 1L in a 2L Cellbag^™ bioreactor on a WAVE bioreactor system (Cytiva, Marlborough, MA) and induced with 200 μM CuSO₄ the following day. The culture supernatant was harvested one week later and Sodium azide (Sigma-Aldrich, St. Louis, MO) was added to 0.01% (w/v) before cold storage or purification.

Recombinant S protein was purified from clarified cell culture supernatants by immunoaffinity chromatography (IAC) using the SARS-CoV-2 cross-reactive mAb CR3022 (provided by Mapp Biopharmaceutical, RRID:AB_2848080) coupled to NHS-activated Sepharose at a concentration of 10 mg/mL. The antigen was eluted with a glycine buffer (pH 2) in tandem into a HiPrep 26/10 desalting column (Cytiva, Marlborough, MA) equilibrated with PBS. The oligomeric content was evaluated by size-exclusion chromatography using a HiLoad 16/600 column (GE Healthcare, Piscataway, NJ) equilibrated with PBS. The S protein eluted as a single peak and the final product migrated as two bands, corresponding to the monomer and trimer on SDS-PAGE, under denaturing conditions, and was reactive to CR3022 mAb on a western blot. Antigens were filter sterilized with a 0.22 μm syringe filter (Cytiva, Marlborough, MA) and stored at −80°C until use.

Analysis of antibodies by multiplex microsphere immunoassay (MIA)

The IgG antibody titers in sera were measured using a multiplex microsphere-based immunoassay as described previously ^{54, 93–95}. Spectrally distinct, magnetic MagPlex^® microspheres (Luminex Corporation, Austin, TX) were coupled to purified S, RBD or bovine serum albumin (BSA). A mixture of the antigen-coupled beads was incubated with sera diluted with PBS+ 1% BSA and 0.02% Tween 20 (PBS-BT) at 1:5,000 or 1:10,000 in black-sided 96-well plates for 3 hours at 37°C with agitation. Bound IgG was detected using 1 μg/mL red phycoerythrin (R-PE)-conjugated goat anti-human IgG antibodies (Jackson ImmunoResearch, Inc., West Grove, PA) and resuspended in MAGPIX^® drive fluid before being analyzed on a MAGPIX^® Instrument (Luminex Corporation, Austin, TX).

To determine S-specific IgG concentrations in the sera, the median fluorescence intensity (MFI) readouts of each sample was interpolated against a standard curve generated using purified human IgG at concentrations in the range of 7.44 to 1000 ng/mL. To produce the antibody standard, IgG was purified from pooled sera of COVID-19 vaccinated human volunteers using protein A affinity chromatography, followed by immunoaffinity chromatography (IAC) using NHS-Sepharose (Cytiva, Marlborough, MA) coupled with recombinant S to select for S-specific IgG. Purity was assessed using SDS-PAGE and antibody concentration was quantified using UV₂₈₀ absorbance. The resulting MFI values were plotted against the Log₁₀-transformed concentrations and fitted using a sigmoidal dose-response, variable slope model (GraphPad Prism, San Diego, CA). The resulting curves yielded r² values > 0.99 with a well-defined top and bottom and the linear range of the curve. The experimental S-specific IgG concentrations in experimental samples were determined by interpolation on the standard curves, multiplied by the dilution factors and plotted as antibody concentrations (ng/mL).

Recombinant vesicular stomatitis virus (rVSV) neutralization assay

Replication-competent rVSV expressing SARS-CoV-2 S protein (Wuhan-Hu-1) was generated as described previously ⁹⁶ and the virus stocks were amplified in Vero E6 cells. For the plaque reduction neutralization test (PRNT), individual NHP serum samples were heat-inactivated at 56°C for 30 minutes. Six 3-fold serial dilutions of serum samples, starting at 1:40 dilution, were prepared and incubated with 100 plaque-forming units (PFU) of rVSV-SARS-CoV-2-S at 37°C for 1 hour. Antibody-virus complexes were added to Vero cell monolayers in 6-well plates and incubated at 37°C for another hour followed by addition of overlay media mixed with 1% agarose. 72 hours later, cells were fixed and stained with a solution containing 1% formaldehyde, 1% methanol, and 0.05% crystal violet overnight for plaque enumeration. Neutralization titers (PRNT₅₀) were generated using a variable slope, nonlinear regression, with upper and lower constraints (100% and 0% neutralization, respectively), using Prism 9 (GraphPad Software, San Diego, CA).

TCID₅₀ and Wild-type SARS-CoV-2 virus PRNT₅₀ assay

TCID₅₀ and PRNT₅₀ assays were performed in a biosafety level 3 facility at BIOQUAL, Inc. (Rockville, MD). The TCID₅₀ assay was conducted by addition of 10-fold graded dilutions of samples to Vero TMPRSS2 cell monolayers. Serial dilutions were performed in the cell culture wells in quadruplicates. Positive (virus stock of known infectious titer in the assay) and negative (medium only) control wells were included in each assay set-up. The plates were incubated at 37°C, 5.0% CO2 for 4 days. The cell monolayers were visually inspected for CPE, i.e., complete destruction of the monolayer with cellular agglutination. The TCID₅₀ value was calculated using the Read-Muench formula ⁹⁷. For samples which had less than 3 CPE positive wells on the first dilution, the TCID₅₀ could not be calculated using the Reed-Muench formula, and these samples were assigned a titer of below the limit of detection (i.e., <2.7 log10 TCID₅₀/mL). The TCID50 value of the positive control should test within 2-fold of the expected value.

To measure neutralization, sera from each NHP were diluted to 1:10 followed by a 3 fold- serial dilution. Diluted samples were then incubated with 30 plaque-forming units of wild-type SARS-CoV-2 USA-WA1/2020 (BEI–NR-52281), B.1.351 (BEI NR-55282), or P.1. (BEI NR-54982) variants, in an equal volume of culture medium for 1 hour at 37°C. The serum-virus mixtures were added to a monolayer of confluent Vero E6 cells and incubated for one hour at 37°C in 5% CO₂. Each well was then overlaid with culture medium containing 0.5% methylcellulose and incubated for 3 days at 37°C in 5% CO₂. The plates were then fixed with methanol at −20°C for 30 minutes and stained with 0.2% crystal violet for 30 min at room temperature. PRNT₅₀ titers were calculated using a variable slope, nonlinear regression, with upper and lower constraints (100% and 0% neutralization, respectively), on Prism 9 (Graphpad Software, San Diego, CA).

Histopathology

NHP Lung tissue specimens from each lung lobe and bronchi were harvested at time of necropsy and preserved in 10% formalin before processing and parafilm embedded, fixed and stained with hematoxylin & eosin. For each NHP, one section from the bronchi and one section from each lobe on the right and left lung were selected for scoring, for a total of six sections. Pathologic findings on each slide were scored on a scale of 0 – 2 for intraalveolar edema, the amount of BALT (bronchiolar-associated lymphoid tissue), and the presence of Interstitial inflammation; and 0 – 3 for perivascular inflammatory infiltrates (cuffing), and intraalveolar inflammation. (See Table S1.) Scores for the six slides from each macaque were tabulated, and a cumulative average score was calculated for each NHP using a total of 30 scores as replicates.

SARS-CoV-2 Viral Genomic and subgenomic RNA quantitative RT-PCR

The presence of viral RNA and viral replication in the BAL, NS, and OS after SARS-CoV-2 P.1 strain challenge was determined by quantitative RT-PCR. RNA was isolated from 200 μL sample using the QIAamp MinElute Virus spin kit (Qiagen, Frederick, MD). To generate a control for the amplification reaction, RNA was isolated from the applicable virus stock using the same procedure. The number of copies for the control were calculated using known RNA weights per mol. A master mix was prepared containing Taq-polymerase, obtained from the TaqMan RT-PCR kit (Bioline cat# BIO-78005), RT, RNAse inhibitor, a primer pair at 2 μM concentration (2019-nCoV_N1-F: 5’-GAC CCC AAA ATC AGC GAA AT-3’ and 2019-nCoV_N1-R: 5’-TCT GGT TAC TGC CAG TTG AAT CTG-3’) and probe (2019-nCoV_N1-P: 5’-FAM-ACC CCG CAT TAC GTT TGG TGG ACC-BHQ1-3’) at a concentration of 2 μM. For the reactions, 45 μL of the master mix and 5 μL of the sample RNA were added to the wells of a 96-well plate. All samples were tested in triplicate. Control RNA was prepared to contain 10⁶ to 10⁷ copies per 3 μL. Eight (8) 10-fold serial dilutions of control RNA were prepared and produced a standard curve with a range of 1 to 10⁷ copies/reaction. For amplification, the plate was placed in an Applied Biosystems 7500 Sequence detector and amplified using the following program: 48°C for 30 minutes, 95°C for 10 minutes followed by 40 cycles of 95°C for 15 seconds, and 1 minute at 55°C. Duplicate samples of each dilution were prepared as described above. If the copy number exceeded the upper detection limit, the sample was diluted as needed. The number of copies of RNA per mL was calculated by interpolation from the standard curve and multiplying by the reciprocal of 0.2 mL extraction volume. This gave a practical range of 50 to 5 x 10⁸ RNA copies per mL for BAL samples and for nasal and oral swabs the viral loads were given per swab.

The RT-PCR assay for the sgRNA utilizes primers and a probe specifically designed to amplify and bind to a region of the N gene mRNA from the Coronavirus, which is not packaged into the virion. The signal was compared to a known standard curve of plasmid containing the sequence of part of the messenger RNA and calculated to give copies per ml. The control RNA was prepared to contain 10⁷ copies. Seven 10-fold serial dilutions of control RNA were prepared using Buffer AVE and generated a standard curve with a range of 1 to 10⁶ copies/reaction. Duplicate samples of each dilution were prepared as described above with the primer pair (SG-N-F: CGATCTCTTGTAGATCTGTTCTC and SG-N-R: GGTGAACCAAGACG CAGTAT) and probe (FAM- TAACCAGAATGGAGAACGCAGTGGG -BHQ). If the copy number exceeded the upper detection limit, the sample was diluted as needed. For amplification, the plate was placed in an Applied Biosystems 7500 Sequence detector and amplified using the following program: 48°C for 30 minutes, 95°C for 10 minutes followed by 40 cycles of 95°C for 15 seconds, and 1 minute at 55°C. The number of copies of RNA per ml was calculated by interpolation from the standard curve and multiplying by the reciprocal of 0.2 ml extraction volume. This gave a practical range of 50 to 5 x 10⁷ RNA copies per mL for harvested samples.

Statistical analysis

Statistically significant differences between the geometric mean of IgG concentrations or MFI in groups given different vaccine formulations at each timepoint were determined using a two-way ANOVA followed by Dunnett’s Multiple Comparison. Comparisons of PRNT₅₀ values between vaccine formulations was done using a Kruskal-Wallis Test followed by a Dunn’s multiple comparisons test. Correlations between IgG or PRNT titers to viral load in the BAL and NS was examined using the non-parametric Spearman’s correlation test. Differences between histopathological scores in different treatment groups was calculated using a one-way ANOVA followed by Dunnett’s multiple comparison test with the cumulative scores of all slides per macaque as replicates. All statistical analysis was completed using Graphpad Prism 9 software (San Diego, CA).

Data and material availability:

All data are available in the main text or the supplementary materials.