The study was performed on 12 adult male and four female rhesus macaques (Macaca mulatta). Males were divided into two vaccine groups and one non-vaccinated challenge control group (four males each), and two females each were added to the two vaccine groups ( Supplementary Table 1 ) to evaluate the effect of the vaccine in males and females. No obvious effect of sex on vaccine immunogenicity and efficacy was seen. The animals underwent a full physical examination prior to entering the study. They were negative for antibodies to simian T-cell leukemia virus (STLV) and simian retrovirus (SRV) and negative for binding antibodies to the RBD of the spike protein of SARS-CoV-2. The study was reviewed and approved by the Dutch “Centrale Commissie Dierproeven” (AVD5020020209404-2) according to Dutch law, article 10a of the “Wet op de Dierproeven,” and BPRC’s Animal Welfare Body (IvD).

At the time of vaccination, NHP were 4 to 8 years old and weighed 5.5 to 11.7 kg ( Supplementary Table 1 ).

ABNCoV2 (previously named RBDn-CLP) was generated as previously described (5). Briefly, a split-protein Tag/Catcher system (7–9) was used to conjugate and display the Wuhan RBD antigen aa319-591 (Sequence ID: QIA20044.1) on the protein surface of preassembled bacteriophage AP205 cVLPs (5). For vaccinations, 15 or 100 μg ABNCoV2 in 0.4 ml PBS with 200 mM sucrose was administered twice intramuscularly in the upper left leg, 14 weeks apart.

Six weeks after the second immunization, all 12 vaccinated animals and four non-vaccinated challenge control animals were exposed to 10⁶ TCID₅₀ SARS-CoV-2 via the combined intranasal (0.25 ml/nostril) and intratracheal (4.5 ml) routes using the CoV isolate BetaCoV/German/BavPat1/2020 (European Virus Archive, Germany) and monitored for 2 weeks.

Tracheal and nasal swabs, as well as bronchoalveolar lavage (BAL) samples, were analyzed for the presence of subgenomic mRNA (sgRNA) of coronavirus using a quantitative real-time PCR as previously described (10). Viral RNA was isolated using a QIAamp Viral RNA Mini kit (Qiagen Benelux BV, Venlo, The Netherlands) following the manufacturer’s instructions. The RT-PCR assay was carried out using the Brilliant II QRT-PCR Core Reagent Kit, 1-Step kit (Agilent Technologies BV, Amstelveen, The Netherlands), according to the instructions provided by the manufacturer in a 25-μl volume with final concentrations of 600 nM for both primers, 200 nM for the probe, and 5 nM MgCl₂, using 10 μl RNA, extracted from 140 μl sample volume. RNA was reverse transcribed for 30 min at 50°C. Then, after a 10-min incubation step at 95°C, the cDNA was amplified for 45 cycles, consisting of 30 s denaturation at 95°C, followed by a 1-min annealing extension step at 60°C. All the reactions were carried out with an iQ5 Multicolor Real-Time PCR Detection System (Bio-Rad Laboratories BV, Veenendaal, The Netherlands).

Serum samples and BAL fluid were tested for the presence of cytokines and chemokines CXCL9, CXCL10, CXCL11, CCL2, CCL3, CCL4, CCL5, CCL11, IL-1β, IL-6, IL-8, IFN-γ, and TNF-α using the LEGENDplex assay (Biolegend LEGENDplex™ NHP Chemokine/Cytokine Panel, 13-plex, art. nr. 740317) according to the manufacturer’s manual. Samples were measured on an Aurora machine (Cytek, Fremont, CA, USA) and analyzed by using company software.

Preimmunization serum samples, as well as serum samples obtained at weeks 2, 4, 12, 16, 18, 20, and 22 (2 weeks after SARS-CoV-2 challenge), were tested for the presence of anti-SARS-CoV-2 spike protein RBD IgG using an enzyme-linked immunosorbent assay (ELISA) with Wuhan-RBD coating antigen (ExpreS²ion Biotechnologies, Agern Allé 1, DK-2970 Horsholm, Denmark). For comparison, 10 human plasma samples positive for SARS-CoV-2 antibodies (SeraCare, Performance Panel 0820-0410) were included in the assay. The RBD-specific IgG levels were calculated against a standard curve made with a monoclonal antibody specifically directed against the RBD of the spike protein of SARS-CoV-2 that was used in a concentration range of 500–78,125 ng/ml. Samples were run in eight different dilutions. Only the OD values that fell within the linear range of the curve were used for calculating the antibody concentration. All samples and standards were run in duplicates and measurements were discarded if CV values of the duplicates were above 25%.

SARS-CoV-2 neutralizing antibodies in serum samples obtained at weeks 2, 4, 12, 16, 20, and 22 (2 weeks after virus inoculation) were measured by live virus plaque reduction neutralization test (PRNT) as previously described (5). For comparison, 10 human plasma samples positive for SARS-CoV-2 antibodies (SeraCare, Performance Panel 0820-0410) were included in the assay. PRNT50 titers were calculated as IC50 using a 4PL non-linear fit model of the software GraphPad Prism 9.0.2. The Wuhan-like, early European B.1. isolate (Freiburg, FR-4286, GISAID accession no. EPI_ISL_852748) was kindly provided by Professor Georg Kochs, University of Freiburg. The B.1. lineage differs from Wuhan (WA1/2020) by the two amino-acid substitutions aa:S:D614G and aa:Orf1b:P314L only, which were acquired early after SARS-CoV-2 emergence in Europe and are present in subsequent B.1 lineage variants, including the alpha, beta, and delta variants.

SARS-CoV-2 variant neutralization was assessed using the same PRNT as described (5), using serum samples obtained 2 weeks after the second vaccination. For these assays, the SARS-CoV-2 variant of concern alpha (B.1.1.7, NCBI GenBank accession no. MZ314997) was kindly provided by Professor Arvind Patel, University of Glasgow, United Kingdom; beta (B.1.351, GISAID accession no. EPI_ISL_678615) was kindly provided by Professor Alex Sigal, African Health Research Institute, South Africa; and delta (B.1.617.2, SARS-CoV-2/DK/SSI-H11, NCBI GenBank accession no. OM444216) was kindly provided by Statens Serum Institute, Denmark. SARS-CoV2 B.1, B.1.1.7, and B.1.617.2 were propagated in VeroE6 cells expressing human TMPRSS2 (VeroE6-hTMPRSS2; kindly provided by Professor Stefan Pöhlmann, University of Göttingen) by infection with MOI 0.05. SARS-CoV2 B.1.351 was propagated in human A549 cells expressing human ACE-2. A supernatant containing new virus progeny was harvested 72 h post-infection. The supernatant was first centrifuged for 10 min at 3,000×g to pellet cellular debris. The virus-containing supernatant was subsequently filtered through a 45-µm filter and concentrated by centrifugation in 100 kDa Amicon tubes (Merck) for 30 min at 4,000×g. Virus titer was determined by TCID₅₀ assay and calculated by the Reed–Muench method.

Following virus propagation, all variants were verified by whole genome sequencing. Briefly, virus isolates were prepared for whole genome sequencing using Nimagen EasySeq and sequenced using Illumina sequencing. Adapter and primer sequences were removed, and reads were mapped to the NC_045512.2 reference using Minimap and iVar (11, 12). Variants and pangolin lineages (13) were determined based on iVar consensus sequences with a minimum 80% base frequency; 70.5% of the B.1.351 sequence contained the nine lineage-specific mutations (aa:E:P71L, aa:N:T205I, aa:orf1a:K1655N, aa:S:D80A, aa:S:D215G, aa:S:K417N, aa:S:A701V, aa:S:N501Y, aa:S:E484K), and 29.5% of the sequence carried eight of the nine lineage-specific mutations and the Wuhan K417.

Enumeration of specific IFN-γ and IL-4 secreting cells was performed by enzyme-linked immunosorbent spot (ELISpot) assay. For this assay, the monkey IFN-γ and IL-4 ELISpot kit from U-CyTech (Utrecht, The Netherlands) was used.

Freshly isolated peripheral blood mononuclear cells (PBMCs) were stimulated in triplicate with 1 µg/ml SARS-CoV-2 peptide pool covering the RBD of the spike protein aa 319–541 [PepMix SARS-CoV-2 (S-RBD), Cat no. PM-WCPV-S-RBD-2, JPT Innovative Peptide Solutions, Berlin, Germany]. The peptide pool contains 53 peptides of 15 aa, with an overlap of 11 aa. The negative control was medium alone, and the positive control was a PMA/ionomycin mixture (used at 50 ng/ml and 1 µg/ml, respectively). PBMCs (1.2 × 10⁶) were stimulated for 12–18 h in 300 µl RPMI-1640 medium supplemented with 10% fetal calf serum (FCS) and stimulus, in a 48-well tissue culture plate. After stimulation (37°C, 5% CO₂), the non-adherent cells (approximately 50% from the start) were collected by three gentle washes using prewarmed RPMI-1640 medium supplemented with 10% FCS. After centrifugation, the cells were resuspended in 150 µl RPMI-1640 medium supplemented with 10% FCS and stimulus. For the enumeration of antigen-specific cytokine production, the 150-µl cell suspension was divided into three wells (approximately 2 × 10⁵ cells/well) of a Millipore MultiScreen 96-well filter plate. The filter plate was prewet for 1 min with 15 µl/well of 35% EtOH and then coated with anti-IFN-γ or anti-IL-4 mAb (1 h at 37°C or O/N at 4°C) and blocked with blocking buffer (1 h at 37°C or O/N at 4°C). The wells contained 50 µl medium with the same supplements as present during stimulation (including stimulus). After 5 h (IFN-γ) or overnight (IL-4) incubation, the cells were washed away, and before adding detector antibodies, an inactivation step with 300 µl 1% Triton X-100 for 1 h at RT was included. Spots were visualized using streptavidin-HRP and an AEC (3-amino-9-ethylcarbazole) coloring system. Spots were automatically counted using the A.EL.VIS reader from Sanquin (The Netherlands).

RBD-specific responses were calculated by subtracting the mean and 2 times the standard deviation of the number of spots measured in the medium control wells from the mean number of spots measured in the RBD peptide-stimulated wells and were expressed as spot-forming units (SFU) per million PBMCs.

Statistical significance of differences between the groups in RBD-binding IgG levels, neutralizing antibody titers, and viral sgRNA copies was calculated by using the Mann–Whitney test. A two-sided α level of 0.05 was used to determine significance. P-values <0.05 were considered significant.

To evaluate the immunogenicity of ABNCoV2, six rhesus macaques per group were immunized intramuscularly with either 15 or 100 μg ABNCoV2 without adjuvant at week 0 and week 14 as outlined in Figure 1A . There were no local reactions observed in any animal at the vaccination site, i.e., the DRAIZE score was completely negative during 4 days post-immunizations (daily scoring). RBD-specific IgG by ELISA ( Figure 1B ), as well as SARS-CoV-2 neutralizing antibodies by live virus PRNT ( Figure 1C ), was already detected 2 weeks after the first immunization. Binding antibody levels increased with time in the high-dose group (100 μg ABNCoV2) from a geometric mean (GM) of 55 ng/ml at week 2 to GM 212 ng/ml by week 12; neutralizing antibodies increased from a PRNT50 geometric mean titer (GMT) of 30 to GMT 83 in the same period. In the low-dose group (15 μg ABNCoV2), antibodies induced after the first vaccination remained at a somewhat lower level (by week 12 reaching a significant difference of p < 0.05 compared with the high-dose group) with a GM of 25, 49, and 32 ng/ml in weeks 2, 4, and 12, respectively, and neutralizing antibodies at a GMT of 33, 37, and 23 at the same timepoints. Overall, binding as well as neutralizing antibodies induced by a single ABNCoV2 vaccination were durable within the monitored 3-month timeframe and were in the same range as those measured in 10 human convalescent samples (GM IgG 16 ng/ml and PRNT50 GMT 48), apart from higher levels (p = 0.0312) of RBD-binding antibodies in the 100-μg ABNCoV2 group. The second ABNCoV2 administration at week 14 significantly boosted both binding (p = 0.0087) ( Figure 1B ) and neutralizing (p = 0.0022) ( Figure 1C ) antibody responses in the two dose groups. At week 16, i.e., 2 weeks after the second ABNCoV2 administrations, IgG GM had increased to 1,835 ng/ml (high dose) and 284 ng/ml (low dose), i.e., 9-fold increase compared with 12 weeks after the first administration. Neutralizing antibodies had increased >50-fold to a GMT of 4,469 in the high-dose group and a GMT of 1,420 in the low-dose group. Although a decline in antibodies was detected in both vaccination groups by the time of challenge in week 20, the GM of binding antibodies was still 16 times higher in the high-dose group compared with human convalescent samples, and the GMT of neutralizing antibody titers was 38 times higher.

To assess whether ABNCoV2 affords protection against infection, all 12 vaccinated animals and 4 non-vaccinated controls were challenged with 10⁶ TCID₅₀ SARS-CoV-2 via the combined intranasal and intratracheal route at week 20, i.e., 6 weeks after the second vaccine administration. In the 2-week observation period post-challenge, only mild clinical symptoms were observed and there were no signs of enhanced respiratory disease ( Supplementary Table 2 ). Indeed, cytokine profiling revealed a vaccine dose-dependent reduction of infection-induced inflammatory C-X-C motif chemokine ligand 10 (CXCL10) and 11 (CXCL11) in both serum ( Figure 2A ) and BAL ( Figure 2B ) compared with non-vaccinated controls. CXCL10, a chemokine described to contribute to exacerbating lung inflammation and progression to acute respiratory disease syndrome with importance as a prognostic and predictive marker for SARS-CoV-2 outcome (14), was transiently elevated in all four control NHP, with peak levels ranging from 191 to 599 pg/ml in serum and 4,423 to 10,473 pg/ml in BAL 2 days post-challenge. In the group vaccinated with 15 μg ABNCoV2, only three of the six animals showed transiently elevated levels of maximal 151–494 pg/ml CXCL10 in serum and 1,085–4,277 pg/ml in BAL. The high-vaccine dose group (100 μg ABNCoV2) showed no elevation of CXCL10 at any time post-challenge in serum, i.e., values were below 100 pg/ml, and only in a single animal 1,234 pg/ml was detected on post-challenge day 2 in BAL, a value well below those measured in control animals. Similarly, CXCL11, another chemokine responsible for the recruitment of inflammatory cells toward infected lungs (14), was transiently elevated in all control animals with peak mean amounts of 188 pg/ml in serum ( Figure 2A ) and 4,178 pg/mL in BAL ( Figure 2B ), compared with reduced peak mean amounts of 52 and 60 pg/ml in the low-vaccine dose group and 11 and 13 pg/ml in the high-vaccine dose group in serum and BAL, respectively. IL-1β and TNF-α (serum and BAL) as well as CCL5, CCL11, and IFN-γ (BAL) were not detected in any group, and changes of other cytokines measured were modest and/or more variable or restricted to the same control animal ( Supplementary Figures 1, 2 showing serum and BAL, respectively).

The efficacy of ABNCoV2 in terms of prevention of SARS-CoV-2 replication was monitored at regular intervals post-challenge. The level of sgRNA, attributed to replicating virus (10), was assessed by quantitative real-time PCR in BAL and throat and nose swabs. As depicted in Figure 3A , all four non-vaccinated control animals harbored a high number of sgRNA copies in BAL 2 days post-challenge (GM 8.5 × 10⁶, range 3 × 10⁶ to 1 × 10⁸ copies/ml), which decreased only slowly with time to a GM of 1 × 10⁵ copies/ml on post-challenge day 4 and 5 × 10³ copies/ml on post-challenge day 6, at which timepoint three of the four non-vaccinated controls were still positive with 6 × 10³ to 4 × 10⁴ sgRNA copies/ml. In contrast, no viral load in BAL was detected at any timepoint in the majority of animals (four of six) vaccinated with 100 μg ABNCoV2 ( Figure 3A ). Compared with controls, the only two positive animals in this group showed reduced levels of 6 × 10³ and 3 × 10⁴ sgRNA copies/ml on post-challenge day 2 (group GM of 511) and had completely cleared the virus 2 days later. The group vaccinated with the lower dose of 15 μg ABNCoV2 measured reduced GM of 9 × 10³ and 433 sgRNA copies/ml on post-challenge days 2 and 4, respectively, and total viral clearance on day 6. When calculating the area under the curve (AUC) to consider both level and kinetics of viral load ( Figure 3B ), a significant (p = 0.0095) difference of both vaccine dose groups could be shown compared with the control group. Similarly, viral load measured in throat swabs was significantly reduced in vaccinated animals (median AUC 1 × 10³ in the high-dose group and 2.25 × 10⁴ in the low-dose group), compared with controls (AUC 5.4 × 10⁵) in a vaccine dose-dependent manner ( Supplementary Figure 3A ). In nose swabs ( Supplementary Figure 3B ), though measuring a vaccine dose response, we failed to detect sgRNA in most control animals for unknown reasons and thereby also failed to show a significant difference such as those seen in BAL and throat.

Next, immune responses 2 weeks following challenge were measured to evaluate whether the significantly reduced viral load in the lungs and throat of ABNCoV2-vaccinated animals provided sufficient antigen for a booster response upon infection. All non-vaccinated control animals developed SARS-CoV-2 neutralizing antibodies ( Figure 4A ) and, apart from one animal, also RBD-specific IFN-γ T-cell responses with 32 to 88 SFU per million PBMCs ( Figure 4B ). Neutralizing antibodies in controls raised to a GMT of 119 2 weeks post-challenge (week 22) and were comparable to those measured in human convalescent samples (GMT 48). In contrast, high-dose ABNCoV2-vaccinated animals showed no increased neutralizing antibody titers 2 weeks post-challenge (week 22, GMT 1,764) compared with the time of challenge (week 20, GMT 1,854). Interestingly, T-cell responses were undetectable post-vaccination in this group and remained undetectable post-challenge in the majority (four of six) of these high-dose animals, while the remaining two animals recorded low T-cell responses (SFU/10⁶ PBMCs of 23 and 14) post-challenge. These findings are consistent with the largely undetectable viral load in this group. In line with a somewhat higher viral load post-challenge in the group vaccinated with 15 μg ABNCoV2, two animals with PRNT50 titers of 107 and 198, which were at the lower end of the group titer range at the time of challenge (PRNT50 of 107–3,539, GMT 390), experienced an increased PRNT50 titer of 1,462 and 1,006, respectively. One of these animals also recorded the highest RBD-specific IFN-γ T-cell response post-challenge (399 SFU/10⁶ PBMCs). No significant increase in neutralizing antibody titers was detected in the remaining four animals in that group. IL-4 T-cell responses were not detected in any animal in the study at any timepoint.

More and more SARS-CoV-2 variants of concern are emerging with increased resistance to antibody neutralization and reduced vaccine efficacy (2–4, 15, 16). We tested the capacity of NHP sera obtained 2 weeks after the second ABNCoV2 administration for their capacity to neutralize the Wuhan-like isolate (FR-4286) in comparison to the variants alpha (B.1.1.7, first identified in the UK), beta (B.1.351, first identified in South Africa), and delta (B1.617.2, first identified in India). As depicted in Figure 5A , no significant difference in neutralizing efficacy was measured between Wuhan-like FR-4286 (PRNT50 of 3,818–32,180, GMT 10,827), B.1.1.7 (PRNT50 of 3,282–14,713, GMT 6,899), B.1.351 (PRNT50 of 6,845–22,602, GMT 10,334), and B1.617.2 (PRNT50 of 10,092–20,000, GMT 14,544) in the 100-μg vaccine group. No significant effect on neutralization of variants was seen in the 15-μg vaccine group either (PRNT50 GMTs of 1,721, 2,207, 2,227, and 4,423 in the Wuhan-like FR-4286, B.1.1.7, B.1.351, and B1.617.2 specific assay, respectively), with the caveat of wide titer ranges in this group. In contrast, 10 human convalescent samples showed a significant drop in neutralizing titers against B.1.351 (GMT 12) compared with the Wuhan-like isolate (GMT 55), as reported by others (15) ( Figure 5B ).