Immunogenicity and safety study of a single dose of SpikoGen^® vaccine as a heterologous or homologous intramuscular booster following a primary course of mRNA, adenoviral vector or recombinant protein COVID-19 vaccine in ambulatory adults

Abstract

Background

SpikoGen^® is a subunit recombinant Wuhan spike protein produced in insect cells and formulated with Advax-CpG55.2^™ adjuvant. It is approved for adult and pediatric use in the Middle East. This study tested the safety and immunogenicity of SpikoGen^® as a 3^rd, 4^th or 5^th dose booster following a primary immunisation course of mRNA, adenovirus or SpikoGen^® vaccine.

Methods

The trial recruited participants who had received a previous doses of COVID-19 vaccine more than 3 months prior. Each received a single intramuscular booster dose of SpikoGen^® vaccine and spike and nuclear protein antibody levels measured at 1 and 3 months post-booster, together with collection of data on SARS-CoV-2 breakthrough infections and symptoms of long Covid.

Results

One-month post-booster, anti-spike IgG, sVNT, and pVNT levels were increased in all groups and there was ~4-fold neutralizing antibodies against the heterologous Omicron BA.2 and BA.4/5 strains. The SpikoGen^®-prime group had the highest levels of anti-spike IgG3, consistent with the Advax-CpG adjuvant driving IgG3 induction. There was no effect of age on the vaccine response. The booster dose was well tolerated with no vaccine-associated serious adverse events. Nine participants (9/74, 12.2%) had a breakthrough SARS-CoV-2 infection between 2 weeks and 3 months post-booster. No long COVID was observed after breakthrough infections. Breakthrough infection was negatively correlated with baseline anti-nuclear protein IgG seropositivity.

Conclusion

A single SpikoGen^® booster was well tolerated and stimulated cross-neutralising antibody responses against Omicron variants, regardless of the primary vaccine received. With SARS-CoV-2 variants continuing to evolve, ongoing research is needed into optimum booster strategies.

Clinicaltrials.gov registration. NCT05542862

Introduction

The pandemic phase of SARS-CoV-2 has officially ended [1]. Nevertheless, new COVID-19 outbreaks continue to occur due to SARS-CoV-2 immune escape variants and waning immunity. The emphasis has now shifted from identifying optimal primary vaccine courses, to identifying optimal booster strategies to restore waning immunity. In particular, there is a need to ensure adequate immunity against the many new Omicron variants including JN.1 and subsequent lineages [2].

High rates of nuclear protein (NP) antibody seroprevalence approaching 90% in many countries [3] confirms most individuals have now had one or more SARS-CoV-2 infections, irrespective of vaccination status. This shows vaccines have had limited impact on infection and transmission, with their major effects being to delay the timing and severity of SARS-CoV-2 infections [4]. Booster doses of mRNA, viral vector, inactivated and recombinant subunit vaccines have all been shown to increase serum neutralising antibody levels to various degrees [5–8]. However, many questions remain regarding optimal use of boosters, including when and how often they should be given, whether they should be monovalent or multivalent, and what variants should be included [9]. Withdrawal of the adenoviral vector vaccines due to low efficacy and safety issues including thrombosis with thrombocytopenia syndrome (TTS) in around 1 in 35,000 immunised individuals [10], has left mRNA and recombinant protein-based vaccines as the major remaining booster options. mRNA vaccines are associated with a low risk of myocarditis/pericarditis which is greatest in males aged 12–30 years within 1–14 days post the second dose of a primary series of mRNA vaccine, with incidence estimates ranging from 1:5000 to 1;20,000 in this younger population [11]. The exact mechanism and long-term significance remains unknown, and while it appears to be milder overall than viral induced myocarditis, rare deaths from mRNA-induced myocarditis have been reported [12]. Multiple doses of mRNA vaccines have also been associated with a IgG4 switch in the antiviral response and the significance of this is still unknown [13]. Traditional protein-based vaccines have a long history of safe and effective use and several protein-based COVID-19 vaccines are now licensed including SpikoGen^® [14], Nuvaxoid^®, Corbevax^® and VidPrevtyn Beta^® [15]., Nuvaxoid^® is the only one of the protein-based vaccine that has also been associated with a small increased risk of myocarditis [12].

SpikoGen^® is based on the extracellular domain of the spike protein expressed in insect cells as a recombinant protein. It is formulated with Advax-CpG55.2^™ adjuvant, which combines delta inulin [16] with CpG55.2 oligonucleotide, a toll-like receptor (TLR)-9 agonist which was the first licensed human pharmaceutical product to be designed by artificial intelligence. SpikoGen^® vaccine has been shown to provide robust protection against SARS-CoV-2 infection in hamster [17], ferret [18] and nonhuman primate models [19, 20]; including protection against infections caused by Beta, Delta and Omicron variants [21]. Although administered by intramuscular injection, it reduced nasal virus shedding and was shown to block virus transmission to naïve co-housed animals [17, 22].

A human Phase 2 trial confirmed a primary course of two SpikoGen^® doses given 3 weeks apart to be safe and immunogenic [23] and able to induce broadly cross-neutralising antibodies [24]. A subsequent pivotal Phase 3 trial involving 16,876 participants (clinicaltrials.gov NCT05005559) confirmed SpikoGen^® significantly reduced SARS-CoV-2 infection as well as significantly reducing the risk of severe disease by 77.5% [25]. SpikoGen^® received emergency use authorisation for adult use from the Iranian FDA on 6 October 2021 and this authorisation was subsequently extended to children aged 5 years and above [26]. It was also shown to be effective as a booster dose in individuals who had received a primary course of inactivated, adenoviral vector or SpikoGen^® vaccine [27]. Until now, minimal data was available on SpikoGen’s ability to boost a primary course of mRNA vaccine. The current study was undertaken to assess the safety and immunogenicity of SpikoGen^® vaccine when used as a 3^rd, 4^th or 5^th dose booster in participants previously immunised with a primary course of mRNA, adenoviral vector vaccine or SpikoGen^® itself. The ability of a SpikoGen^® booster dose to boost spike antibody responses and induce cross-neutralisation of representative Omicron strains including BA.2, BA.4/5 and XBB.1 was assessed. Data was also collected on vaccine-breakthrough infections and long COVID symptoms following infection.

Methods:

Trial design and participants

The trial was initiated in June 2022 as an open-label study in adults who had previously received at least two, but up to four, doses of different COVID-19 vaccine. The most immediate prior dose had to have been received more than 3 months prior, with no upper limit on this duration. All eligible participants received a single booster dose of SpikoGen^® 25 μg formulated with Advax-CpG55.2^™ adjuvant. Exclusion criteria included: pregnant or lactating women; clinical signs or symptoms of SARS-CoV-2; history of severe adverse reactions (e.g., anaphylaxis) to any components of SpikoGen^® or other medications; receipt of any other research product within 30 days before screening (or intention to participate in another clinical study); previous vaccination with any other authorised vaccine within 28 days before screening (or the intention to receive such a vaccine within 14 days after the booster dose); or any serious medical, social or mental condition which, in the opinion of the investigator, would be detrimental to the subjects or the study. Blood samples for serology were obtained at baseline and at 1 month (+/− 1 week) and 3 months post the booster dose. The study was approved under the TGA Clinical Trial Notification Scheme (CTN) and registered at clinicaltrials.gov as NCT05542862.

ELISA and surrogate viral neutralisation (sVNT) assays

SARS-CoV-2 anti-spike IgG assay kits manufactured by Pishtaz Teb Company, Iran were used to measure serum Wuhan strain spike protein (spike^W) binding IgG, using the manufacturer’s instructions. The seropositive cut off was defined as > 8 RU/ml providing a sensitivity of 98.16%, and specificity of 99.1%. RU values were converted to BAU using the formula BAU=3xRU+20.

Spike^W binding Immunoglobulin isotype/subclasses antibodies were determined using biotinylated goat anti-human IgM, IgG1, IgG2, IgG3 and IgG4 (all from Sigma-Aldrich) mixed with HRP-conjugated Streptavidin (BD Biosciences).

SARS-CoV-2 neutralising ELISA kits, manufactured by Pishtaz Teb Company, Iran were used to measure serum antibodies that inhibit the binding of ACE2 to Wuhan receptor binding domain as a surrogate viral neutralisation (sVNT^w) assay. Based on analysis of healthy control sera, the mean plus 3 standard deviations was established as the seropositive cut-off for the sVNT^w assay, namely 2.5 μg/ml, which provided a sensitivity of 99.0% and a specificity of 100% when used against known samples.

ELISA kits (Xema, Finland) were used to measure SARS-CoV-2 nuclear protein (NP) binding IgG. Based on the results of healthy control sera, a seropositive cutoff of 0.2 optical density (OD) units was used.

Pseudotyped lentivirus neutralisation (pVNT) assay

An internally developed SARS-CoV-2 spike pseudotyped lentivirus neutralisation assay was used to measure the ability of sera to inhibit infection by a panel of pseudotyped lentiviruses expressing spike proteins of Wuhan and Omicron BA.2, BA.4/5 and XBB.1 variants, as previously described [24]. Briefly, serially diluted heat inactivated immune sera were incubated with pseudotyped virus particles for 1 h at 37°C then 100 μl of virus-serum mixture was added into 50 μl of 293T-hACE2 cells freshly plated at 125,000 cells per well in a 96-well white tissue culture plate. The cells were cultured at 37°C for 48h, followed by removal of the culture medium and addition of 30 μl of Phenol red-free DMEM medium (ThermoFisher). Then 30 μl of ONE-Glo EX Luciferase reagent (Promega) was added into each well and incubated at room temperature (RT) with shaking at 450rpm on a ThermoMixer (Eppendorf) before reading on FluoStar plate reader (BMG Labtech, Carbine Way, VIC, Australia). Neutralisation was calculated as the reduction in percent luciferase units relative to pseudotyped virus alone group without any serum treatment. The 50% inhibitory concentration (IC₅₀) was calculated using Sigmoidal 4PL robust fit regression method in GraphPad Prism V9.

Outcome measurements

The primary outcomes were anti-spike antibody seroconversion and geometric mean titre fold rise (GMFR) of neutralising antibodies as measured by sVNT and pVNT assays, 1 month after the booster dose. Data on PCR or RAT confirmed SARS-CoV-2 infections during the study period was collected via electronic participant diaries. Vaccine breakthrough infections were defined as any infections occurring from 2 weeks to 3 months post the booster dose as detected by PCR, RAT or NP antibody seroconversion. Seroconversion was defined as a change in the status of antibody levels from negative to positive based on the prespecified threshold or a four-fold or greater rise in levels over baseline values. Safety outcomes included the incidence of local and systemic solicited adverse events for seven days after the booster dose. Serious adverse events were evaluated up to the end of study follow up. Safety outcomes were reported based on the classifications provided in the Medical Dictionary for Regulatory Activities (MedDRA). Each patient’s severity score was assessed based on the FDA toxicity grading scale [28].

Statistical Analysis

A formal power calculation was not performed. All the participants who received the booster dose were included in the safety population. The immunogenicity objectives were reported based on the per-protocol set for participants that received a booster dose and had at least one post-immunisation blood sample taken. Participants who became infected with SARS-CoV-2 were excluded from the per-protocol population. Missing data were not imputed. Analyses were performed based on the type of vaccine used in the primary vaccination course. For the purposes of analysis, those who had received either Moderna and Pfizer mRNA vaccines or a combination of these as their primary vaccine course were analysed as a single mRNA-prime group. No adjustments for multiple comparisons were made. Continuous data were compared using t-test and categorical data were assessed using Fisher’s exact test. Hypothesis testing was two-sided, and p-values less than 0.05 were considered significant. GraphPad Prism was used for preparation of graphs and for statistical analyses.

RESULTS

Trial demographics

The trial was initiated in June 2022. Of a total of 81 participants who were consented and received the SpikoGen^® booster, 2 participants were later deemed ineligible, one because they had less than 3 months since their last vaccine dose and the other because they had not received a primary course of a specified COVID-19 vaccine. A further 5 participants failed to provide a post-vaccination blood sample which left 74 eligible participants who received the SpikoGen^® boost and had follow up blood results for analysis. The demographic data of eligible participants who completed the study are presented in Table 1. The minimum duration between the previous dose and having the booster was 3 months and there was no upper limit specified in the protocol. The range of time since the most immediate prior dose was 3 to 16.5 months with a median duration of 8.5 months. The eligible participants had a mean age of 58.7 years, ranging from 30 to 79 years with the mean age of the adenoviral-prime group (66.6 years) being slightly higher than the SpikoGen^®-prime (57.97 years) and mRNA-prime (54.13), groups. This slightly older age likely reflects that during the Australian COVID-19 vaccine rollout the AstraZeneca adenoviral vaccine was given to older individuals, whereas mRNA vaccines were reserved for younger individuals. There were approximately equal numbers of female (51.4%) and male (48.6%) participants. Those in the SpikoGen^®-prime group had participated in a previous vaccine trial where they had received two primary doses of SpikoGen^® vaccine (clinicaltrials.gov NCT05148871). Fifty percent of the adenoviral-prime and 31.8% of the mRNA-prime groups had had 3 or more previous COVID-19 vaccine doses whereas all those in the SpikoGen^®-prime group had only had two previous doses.

Table 1.

Participant demographics

Characteristics	All n=74	Adenoviral-prime (AstraZeneca) n=16	Protein-prime (SpikoGen®) n=36	mRNA-prime (Pfizer/Moderna) n=22
Sex — n (%)
Male	36 (48.6%)	9 (56.3%)	14 (38.9%)	13 (59.1%)
Female	38 (51.4%)	7 (43.8%)	22 (61.1%)	9 (40.9%)
Age (yrs) — mean ± SD	58.70 ± 12.36	66.62 ± 10.59	57.97 ± 11.47	54.13 + 12.7
Min. age	30	47	30	31
Max. age	79	79	79	71
Receipt of 3rd booster dose	13 (17.6%)	7* (43.8%)	0 (0%)	6# (27.2%)
Receipt of 4th booster dose	2 (2.7%)	1^ (6.3%)	0 (0%)	1$ (4.5%)

^*one participant received a 3^rd dose of AZ, three a 3^rd dose of Pfizer mRNA, two a Moderna mRNA, and one a 3^rd dose of Novavax

^{^}one participant received a 4^th dose of Moderna mRNA

^#three participants received a 3^rd dose of Novavax, two a 3^rd dose of Pfizer mRNA, and one a 3^rd dose of Moderna mRNA

^$one participant received a 4^th dose of an unspecified COVID-19 vaccine

Baseline serology

At baseline, anti-nuclear protein (NP) IgG seropositivity as a marker of previous infection was present in 7/16 (43.8%) of the adenoviral-prime, 15/36 (41.7%) of the SpikoGen^®-prime and 12/22 (54.5%) of the mRNA-prime, groups (Fig. 1A). In those who were NP seropositive at baseline, the levels generally showed some evidence of waning levels at 1 and 3 months post-booster (Fig. 1D). Six initially NP seropositive subjects became seronegative at 1-month post-booster (Fig. 1B), consistent with NP antibodies having a relatively short half-life. This indicated that approximately half the study participants already had hybrid immunity due to the combination of prior infection and vaccination.

Figure 1. Anti-NP IgG levels.

Serum NP IgG levels measured by ELISA at baseline (A) and 1 (B) and 3 months (C) post-booster in adenovirus, Spikogen^® and mRNA-primed groups. Changes in NP IgG over time in baseline NP seropositive participants (D). Geometric mean and 95% confidence interval.

At baseline, spike^w-binding IgG, sVNT^w and pVNT^w levels were significantly higher in the mRNA-prime and adenoviral-prime groups than the SpikoGen^®-prime group (Fig. 2). Baseline seropositivity for spike^w-binding IgG was 21/22 (95.5%) for the mRNA-prime, 15/16 (93.8%) for the adenoviral-prime and 14/36 (38.9%) for the SpikoGen^®-prime group (Fig 2A). Similarly, baseline sVNT^w seropositivity which measures antibodies that inhibit binding of the Wuhan RBD to ACE2, was 17/22 (77.3%) in the mRNA-prime and 8/16 (50.0%) for the adenoviral-prime versus just 1/36 (2.8%) for the SpikoGen^®-prime group (Fig. 2B). Baseline seropositivity by the pVNT^w assay, using a seropositive cut-off of 1:40 or greater, was 22/22 (100%) for the mRNA-prime, 15/16 (93.8%) for the AZ-prime and 27/36 (75.0%) for the SpikoGen^®-prime group (Fig 2C).

Figure 2. Serum spike antibody responses to SpikoGen^® booster.

Sera were assayed for Wuhan spike binding IgG by ELISA, surrogate viral neutralisation (sVNT) titre and pseudotype viral neutralisation (pVNT) test at baseline (A-C), and 1 and 3 months post-booster. The fold change in antibody levels at 1 month post-booster over baseline levels (D-F). Levels of the antibodies at 3 months post-booster (G-I). Geometric mean and 95% confidence interval.

Response to SpikoGen^® booster dose

One month post receipt of the SpikoGen^® booster, all groups showed a significant rise in serum spike^w-binding IgG, sVNT^w and pVNT^w levels (Fig 2D-F). One month post-booster, the SpikoGen^®-prime group showed a ~4-fold increase in spike^w binding IgG (Fig 2D), ~16-fold increase in sVNT^w (Fig. 2E) and ~4-fold increase in pVNT^w (Fig. 2F). This group had the lowest baseline titers, which may have skewed in favour of this group having the greatest fold-increase of titers. Seropositivity by sVNT^w rose to 21/22 (95.5%) in the mRNA-prime, 9/16 (56.3%) in the adenoviral-prime, and 18/36 (50.0%) in the SpikoGen^®-prime, groups.

At 3 months post-booster, there were no significant differences between the groups in spike-binding IgG and pVNT^w levels, but sVNT^w levels were significantly higher in the mRNA-prime group compared to the adenoviral-prime and SpikoGen^®-prime groups (Fig. 2H).

Breakdown of spike-binding immunoglobulin into specific antibody isotpes, showed most participants at baseline had measurable levels of spike-binding IgM together with modest levels of IgG1, IgG2, IgG3 and IgG4 (Supplementary Figure S1). Elevated levels of spike-binding IgM were still apparent at 3 months, particularly in the SpikoGen^®-prime and mRNA-prime groups, suggesting the booster induced long-lived anti-spike IgM (Fig. 3). Most of the anti-spike IgG response at 3 months post-booster was made up by IgG1 and IgG3, with small rises in IgG2 and IgG4. It was notable that the SpikoGen^®-prime group had higher anti-spike IgG3 at baseline and 1 and 3 months post-booster than the adenovirus-prime or mRNA-prime groups (Fig. 3). This was consistent with what we have observed in influenza vaccine studies where Advax adjuvant similarly preferentially induced IgG3 plus IgG1 [16].

Figure 3. Anti-spike IgM and IgG subtype levels 3 months post SpikoGen^® booster dose.

Sera were assayed for spike-binding IgM and IgG subtypes by ELISA using Wuhan spike protein as the substrate. Shown are levels at 3 months post booster broken down by vaccine-prime received. Data on immunoglobulin subtypes at other time points is shown in Supplementary Figure S1. Geometric mean and 95% confidence interval. .

Next, we examined the effect of the booster on the pVNT response against vaccine-heterologous Omicron variants (Fig. 4). One month post-booster, the SpikoGen^®-prime group showed a ~4-fold increase in pVNT against Omicron BA.2 and BA.4/5 compared to minimal increases in the adenoviral-prime and mRNA-prime groups (Fig. 4A-B). All groups showed ~2-fold increases in XBB.1 pVNT (Fig. 4C). Three months post-booster there was no significant differences in pVNT against the BA.2, BA.4/5 and XBB.1 variants between the various groups (Fig. 4D-F).

Figure 4. Cross-neutralizing antibody responses after SpikoGen^® booster against the Omicron BA.2, BA.4/5 and XBB.1 variants.

Shown are fold-change increases from baseline to 1 (A-C) and 3 (D-F) months post-booster in serum pVNT for Omicron BA.2, BA.4/5 and XBB.1 variants. Geometric mean and 95% confidence interval.

To further assess for potential differences in the spike antibodies induced by the different vaccine types, the ratios of pVNT^w to total spike^w-binding IgG, sVNT^w to total spike^w-binding IgG and pVNT^w to sVNT^w were calculated. At baseline, the mRNA-prime group had a significantly higher sVNT^w:spike IgG^w ratio than the other two groups whereas the SpikoGen^® group had a significantly higher pVNT:sVNT ratio (Supplementary Fig. S2 A-C), suggesting that the mRNA vaccine induces a higher ratio of antibodies that inhibit RBD binding to ACE2 as a percentage of total spike-binding IgG. By contrast, the SpikoGen^®-prime group had good neutralising activity in the pVNT assays despite having low sVNT^w activity. This suggests the neutralising antibodies in the SpikoGen^®-prime group may bind to different neutralizing epitopes than antibodies in the mRNA-prime group. In response to the SpikoGen^® booster, sVNT levels increased in all groups including the SpikoGen^®-prime group. Three months post-booster, the mRNA-prime group had the highest sVNT^w:spike^w IgG ratio suggesting that the original effect of the mRNA-prime causing high sVNT activity remained despite the SpikoGen^® booster (Supplementary Fig. S2G-I).

Effects of age on booster vaccine response

Previous studies have suggested a negative impact of age on vaccine responses [29]. We therefore analysed whether there was any relationship between a participant’s age and their fold changes in pVNT from baseline against Wuhan and three major Omicron variants. There was no relationship between a participant’s age and their neutralizing antibody response to the Wuhan and the three Omicron variants, suggesting that older age did not impede responses to the SpikoGen^® booster (Supplementary Fig. S3).

Vaccine Breakthrough Infections

A total of nine SARS-CoV-2 infections occurred during the 3-month follow up period, with three being diagnosed by NP seroconversion and six by PCR or RAT after development of symptoms. One infection in a SpikoGen^®-prime participant occurred sooner than 14 days post-booster and was excluded from further analysis. The remaining eight vaccine-breakthrough infections were spread evenly across the vaccine-prime groups. Within the population of baseline NP seronegative participants, the rate of vaccine-breakthrough infections from 2 weeks after the booster dose was 3/9 (33.3%) in the adenoviral-prime group, 3/10 (30.0%) in the mRNA-prime group and 2/21 (9.5%) in the SpikoGen^®-prime group. Across all study participants (baseline NP seronegative or seropositive) the rate of breakthrough SARS-CoV-2 infections was 3/16 (18.8%) in the adenoviral-prime group, 3/22 (13.6%) in the mRNA-prime group and 2/36 (5.6%) in the SpikoGen^®-prime group. These infections occurred in late 2022 and early 2023 when the Omicron strains were widely circulating. Overall, the vaccine-breakthrough infections were relatively short lived and were rated by affected participants as mild to moderate in severity. None of the infections required medical assistance and none resulted in prolonged symptoms suggestive of long COVID. All in the adenoviral-prime and mRNA-prime groups with confirmed breakthrough infections had evidence of NP seroconversion, however 2/3 (66.6%) of the breakthrough infections in the SpikoGen^®-prime group failed to show NP seroconversion. Mean NP antibody levels 3 months post-booster were significantly lower (p=0.0083) in those with breakthrough infections in the SpikoGen^®-prime group versus those in the combined mRNA-prime and adenoviral-prime groups.

Antibody levels in those with breakthrough infections were compared to those who did not report infection to assess for potential correlates of susceptibility/protection. NP seropositivity at baseline showed a significant negative correlation with a subsequent vaccine breakthrough infection (Fig 5A). No participants who were NP seropositive at baseline (0/34, 0%) developed a breakthrough SARS-CoV-2 infection. Vaccine breakthrough infections were solely concentrated in those NP seronegative at baseline (9/43, 20.9%, p=0.004, Fishers exact test) including symptomatic breakthrough infections (6/43, 14.0%, p=0.031, Fishers exact test). This is consistent with hybrid immunity from prior infection plus vaccination providing robust protection against re-infection even by new variants. Serum levels of spike^W-binding IgG, sVNT^W, or pVNT against BA.2, BA.4/5 or XBB.1 at baseline did not show any relationship to breakthrough infection which occurred equally in those with high or low spike antibody levels (Fig 5. A-F).

Figure 5. Potential correlates of SARS-CoV-2 protection.

Serum NP and spike antibody levels at baseline (A-F) in participants that had or didn’t have a SARS-CoV-2 infection detected by PCR/RAT positivity or NP antibody seroconversion for 3 months follow up post booster dose. Age range of participants that either had or didn’t have a SARS-CoV-2 infection during follow up (G). Spike antibody levels (H-N) and pVNT:spike IgG ratio (M) at 1 month post-booster in participants that did or did not have a vaccine breakthrough SARS-CoV-2 infection. Geometric mean and 95% confidence interval.

Next, we looked to see whether there was any effect of age on the likelihood of SARS-CoV-2 infection. Infections were mainly clustered in those aged 50–80 years rather than younger subjects but there was no statistical difference when comparing ages of those infected and not infected (Fig 5. G).

If antibodies mediate SARS-CoV-2 protection, then the level of antibody levels 3 weeks post-booster would be expected to show an inverse correlation with infection risk. However, pVNT levels against the Omicron BA.2, BA.4/5 and XBB.1 variants did not differ significantly between subsequently infected and uninfected individuals (Fig. 5 J-L). Surprisingly, sVNT^W levels 3 weeks post-booster were significantly higher in those that subsequently got infected than those that didn’t (Fig. 5 H). The ratio of pVNT:spike binding IgG 1 month post-booster was not significantly different between those that got infected and those that didn’t.

Safety

Overall, the booster dose was well tolerated with only mild to moderate transient adverse events reported by trial participants and no vaccine-related serious adverse events reported. Repeated mRNA vaccine doses have been reported to induce switching to anti-spike IgG4 [13]. Despite the SpikoGen^® booster representing at least the 3^rd and up to the 5^th vaccine dose, minimal induction of anti-spike IgG4 was seen after the booster dose (Fig. 2).

DISCUSSION

Serum SARS-CoV-2 neutralising antibody levels fall rapidly from their peak, with RBD-binding IgG levels being reported to fall an average of ~90% within 90 days of infection [30]. The reason why SARS-CoV-2 antibody levels are so transient after either infection or vaccination remains poorly understood [31]. A rapid loss of serum neutralising antibodies is also a feature of other human coronavirus infections, with this drop being associated with recurrent susceptibility to re-infection on a seasonal basis [32]. Nevertheless, as we have found ourselves in non-human primate studies, SARS-CoV-2 protection may be maintained even in absence of neutralizing antibody [19]. One possibility is this ongoing protection is mediated by virus-specific T cells. In addition, the SARS-CoV-2 virus itself continues to evolve over time, resulting in regular development of antibody-escape variants [33].

The current study assessed the utility of SpikoGen^® vaccine as a 3^rd, 4^th or 5^th dose booster in participants who had previously received primary courses of either mRNA, adenovirus, or recombinant protein vaccines. It showed the SpikoGen^® booster dose was safe and boosted binding and neutralizing antibody levels including against the major Omicron variants, regardless of the primary vaccine type participants had received. Interestingly, the SpikoGen^® booster induced antibodies able to cross-neutralize the Omicron variants, but had only a modest effect on sVNT^W levels reflective of antibodies that inhibit Wuhan RBD binding to ACE2 Furthermore, few of the SpikoGen^®-prime group were positive for sVNT^W antibodies at baseline, compared to high rates of sVNT^W antibodies in the mRNA-prime and adenovirus-prime groups. This suggests SpikoGen^® induces antibodies that target SARS-CoV-2 other than by inhibition of RBD-ACE2 binding [34]. sVNT antibodies are induced by RBD domains in the “up” position needed to bind ACE2 [35]. The RBDs in SpikoGen^® spike protein may be predominantly in the “down” closed position, where the RBD is less likely to be seen by RBD-binding antibodies with sVNT activity [18]. This fits with our observation that the recombinant spike protein in SpikogGen^® has very low binding to ACE2 in ELISA assays, whereas denatured spike protein paradoxically has high ACE2 binding (unpublished data), Nevertheless, SpikoGen^® vaccine still induces serum antibody able to neutralize SARS-CoV-2 virus infectivity in pVNT assays.

At baseline, the mRNA-prime group had a significantly higher sVNT^W:spike^W IgG ratio than the SpikoGen^®-prime group, with this effect still evident 3 months post-booster. This suggests that the mRNA vaccines induce a higher proportion of antibodies that bind the RBD^W and inhibit binding to ACE2 as a proportion of total spike^W-binding IgG. This suggests different vaccine platforms may induce different types of neutralizing antibody with different effects. When comparing the different vaccine priming effects, the participants in the adenovector-prime group had both the lowest antibody responses to the SpikoGen^® vaccine booster as well as a trend to the greatest antibody decay afterwards, although there were no statistically significant differences between groups.

All but one of the participants who subsequently developed a vaccine-breakthrough SARS-CoV-2 infection by heterologous Omicron strains from the timing had very high post-booster Wuhan sVNT^W levels (Fig. 7). This raises the possibility that high levels of heterologous sVNT^W antibodies may enhance the risk of a heterologous Omicron infection [36]. This possibility cannot be confirmed or excluded based on the small number of infections in our study participants Antibody-dependent enhancement (ADE) has been seen with SARS CoV and feline coronavirus, vaccines [37], and non-neutralizing antibodies have been shown to enhance SARS-CoV-2 infectivity in vitro [38]. Approximately half the study participants already had hybrid immunity due to the combination of prior infection and vaccination, and none of these individuals developed a breakthrough infection, attesting to hybrid immunity providing robust protection.

SpikoGen^® is manufactured in insect cells which glycosylate spike protein using paucimannose-type glycans whereas the spike protein induced by mRNA and adenoviral vaccines is produced in vivo thereby conferring mostly complex- and high-mannose-type glycans [39]. Such differences in glycosylation could result in different spike epitopes being accessible [40] and different antibodies being induced. Epitopes in the N-terminal domain (NTD), fusion peptide and stem-helix region of the S2 domain have been shown to have neutralizing activity [41]. Despite SpikoGen^® being based on the original Wuhan spike protein, the booster dose induced strong neutralizing responses against Omicron BA.2 and BA.4/5, albeit with lower responses to the XBB.1 variant. As most vaccine-escape virus mutants evade neutralizing antibodies via mutating their RBD, it is possible that the neutralising antibodies induced by SpikoGen^® vaccine may be more resistant to virus escape which may help explain how SpikoGen^® vaccine induces such broadly cross-neutralizing antibodies [24]. Another important difference we found between vaccines was that SpikoGen uniquely induced high levels of spike IgG3 as well as IgG1 whereas the other vaccine types primarily induced IgG1. IgG3 has the greatest antiviral activity of all the IgG isotypes, thanks to its unique features including an elongated hinge region, greater molecular flexibility, extensive polymorphisms, and additional glycosylation sites not present on other IgG subclasses, giving it potent effector functions including complement activation, antibody-mediated phagocytosis, and antibody-mediated cellular cytotoxicity [42]. The ability to induce high levels of anti-spike IgG3 may contribute to its ability to protect against heterologous SARS-CoV-2 variants, where in the absence of effective RBD-ACE2 inhibiting antibodies such anti-viral IgG3 effector functions may play a critical role in virus control.

The baseline antibody data indicated a trend to the highest NP seropositivity rate as evidence of past infection in the mRNA- and adenoviral-prime groups, with the lowest baseline NP seropositivity rate in the SpikoGen^®-prime group. This could suggest the SpikoGen^®-prime group has had a lower rate of past infection, a longer mean time since infection, or milder infections [43]. A higher rate of past infections in the mRNA-prime and adenoviral-prime groups may also explain why these groups had the highest baseline spike^W-binding IgG seropositivity. Interestingly, many individuals in the mRNA-prime and adenoviral-prime groups had already had 3^rd and 4^th booster doses, whereas everyone in the SpikoGen^®-prime group had only received 2 previous doses.

After 2 doses of BNT162b2 mRNA COVID-19 vaccine, the antibody half-life was 55 days in baseline seronegative subjects, but 80 days in baseline seropositive subjects [44]. In a previous study we found that ~50% of vaccinees no longer had detectable serum sVNT antibodies 6 months after their most recent vaccine dose, irrespective of whether they had received an inactivated virus, adenoviral or recombinant protein vaccine [27]. Just why SARS-CoV-2 antibody levels decay so rapidly is not known. One potential mechanism could be homology or mimicry between the SARS-CoV-2 RBD and a human self-protein such as transferrin or lactoferrin [45], resulting in B cells producing RBD antibody being rapidly deleted to avoid autoreactivity. Interestingly, this phenomenon of rapid SARS-CoV-2 antibody decay was apparent in both non-human primates [19] and humans [27], but not in mice where SARS-CoV-2 antibody levels remained high over time [17]. If the phenomenon of rapid SARS-CoV-2 antibody decay is primate-specific, this could support a self-protein mimicry hypothesis to explain rapid antibody decay. This also raises the question of whether excessively high and prolonged SARS-CoV-2 antibody levels might have adverse consequences such as causing autoimmune phenomena, given SARS-CoV-2 spike protein antibodies can cross-react with structurally similar host protein sequences [45]. Another explanation for rapid SARS-CoV-2 antibody decay after infection in primates could be that the virus is interfering with plasmablast maturation, thereby inducing production of short-lived plasma cells. As this phenomenon of rapidly waning antibodies is seen not just with SARS-CoV-2 infection [46] but also after vaccination with spike protein [47], this could suggest that spike protein in primates may mediate a block to B cell-T cell signalling. SARS-CoV-2 spike protein has a range of biological activities, including the ability to upregulate inflammatory cytokine production including IL-6, IL-1β, TNFα, CXCL1, CXCL2, and CCL2, while suppressing interferon production in human macrophages via a TLR2 dependent pathway [48], to bind and regulate nicotinic acetylcholine receptors [49–51] and promote inflammatory responses and cell apoptosis through upregulation of intracellular reactive oxygen species [52]. Generation and affinity maturation of spike-specific memory B cells has been shown to be T cell-dependent and correlates with IL-21⁺ CD4⁺ T cells in recovered individuals and CD40L⁺ CD4⁺ T cells in severely ill individuals [53]. This could fit with a model in which spike protein interferes in generation of long-lived plasma cells. Rapidly waning vaccine-induced antibody levels are particularly relevant in respect of the newer Omicron vaccine-escape variants where cross-neutralizing antibodies are much lower post-vaccination, thereby rapidly falling below protective levels [54].

Serum antibody levels are just one aspect of vaccine protection. The data shows no correlation between spike antibody levels and protection against vaccine breakthrough infection. Notably, this mirrors findings in by SpikoGen^®-immunized monkeys which remained protected against COVID-19 infection despite decay of theirserum neutralizing antibody (pVNT) against the infecting strain to negligible levels [19, 20]. The lack of relationship between serum spike antibody levels and protection may reflect a more dominant role of either mucosal [55] and/or T cell immunity [56] in long-term SARS-CoV-2 protection.

Severe illnesses are often associated with higher antibody titres [57], whereas absence of antibody seroconversion is seen in those with mild disease [46]. It was interesting that 2/3 participants (66.7%) in the SpikoGen^®-prime group with vaccine-breakthrough infection failed to show NP seroconversion afterwards consistent with their reports these were mild infections.

The current data again highlight the power of hybrid immunity to prevent recurrent infection, with none of the baseline NP seropositive subjects developing a breakthrough infection. This is consistent with results of a recently published meta-analysis which showed that the effectiveness of hybrid immunity against hospital admission or severe disease was 97·4% at 12 months [58].

Despite its small size it was reassuring that no safety issues were encountered in response to administration of the SpikoGen^® booster. The study included participants up to 80 years of age with no detrimental effects of age seen on the antibody responses to the SpikoGen^® booster. Reactions were mild and short-lived with full recovery. The SpikoGen^® booster did not induce high levels of anti-spike IgG4. No cases of myocarditis or pericarditis were observed, including in the boosted mRNA-prime group, although the study was not powered to detect such rare adverse events. None of the participants with a vaccine breakthrough infection developed symptoms of long COVID. Long COVID is said to affect at least 10% of COVID-19 infections with the majority of cases occurring in non-hospitalised patients with a mild acute illness [59]. Samples from patients with long COVID have shown presence of SARS-CoV-2 viral RNA and circulating spike protein months after infection, which suggests long COVID may represent a form of chronic infection [60]. Other potential causes include immune dysregulation, reactivation of chronic infections such as with Epstein-Barr virus or human herpesvirus, microbiota effects, autoimmunity, microvascular blood clotting, endothelial dysfunction or dysfunctional brainstem signalling [59].

While providing valuable data, this study has several limitations. It was not practicable to run this booster trial as a blinded or randomised study and hence various confounding factors need to be considered. Those in the adenoviral-prime group tended to be older and many had had additional mRNA booster doses following their primary adenoviral vaccine course. In the early stages of the pandemic older individuals were directed to receive the adenovirus vaccine with mRNA vaccines being reserved for younger adults, to avoid risk of life-threatening adenovirus-associated thrombotic events [61]. Based on baseline NP seropositivity, about half the study participants had already had a prior SARS-CoV-2 infection. Neither T cell nor mucosal responses were able to be assessed as part of this study. However, SpikoGen^® induced robust spike protein memory CD4 and CD8 T cell responses in previous murine [22] and human [23] studies and Advax-CpG adjuvant has been shown to be a potent inducer of CD8 T cell responses in NHP [62]. The small number of study participants meant the study was not powered to assess for potential rare vaccine side effects such as myocarditis/pericarditis. Reassuringly, no cases of vaccine-associated myocarditis were observed in the large phase 3 trial of SpikoGen^® that involved 16,876 participants, nor in pharmacovigilance studies after delivery of 8 million doses. The low number of breakthrough SARS-CoV-2 infections in this study also meant it was not possible to assess whether the SpikoGen^® booster prevented long Covid, even although no cases were reported.

Overal, this study showed that it convenient and safe to administer SpikoGen^® vaccine as a booster irrespective of the previous primary course of COVID-19 vaccine an individual had received. The SpikoGen^® booster despite still being based on the original Wuhan spike protein induced broadly cross-neutralising antibodies against a diversity of Omicron variants. Studies are underway to assess whether updating the SpikoGen^® spike protein to a currently circulating variant such as LB.1 will further enhance its efficacy against future strains. Another study is testing whether a SpikoGen^® booster administered via the oral route can induce protective mucosal immunity.