PRV-101 met the primary safety endpoint of the PROVENT trial, providing initial evidence of acceptable safety and tolerability of the product in healthy adult volunteers. No treatment-emergent SAEs, or adverse events leading to study drug discontinuation or study withdrawal were observed. All TEAEs in the PRV-101 groups were mild to moderate in severity. In addition, the vaccine-induced robust antibody responses to all five CVB serotypes included in the product. Altogether, these findings create a solid basis for the future development of the PRV-101 vaccine candidate.

The observation of robust neutralising antibody responses against the five CVB serotypes is highly important for the future development of PRV-101. Neutralising antibodies are specific for individual CVB serotypes and this finding therefore indicates that each inactivated virus type was immunogenic in the vaccine. Of note, previous studies with the poliovirus vaccine have shown that neutralising antibodies mediate protection against the virus [31, 32].

PRV-101 induced a clear increase in neutralising CVB antibody titres in both initially CVB seropositive and seronegative participants. In the initially seronegative participants, a clear dose–response relationship was observed, as the higher PRV-101 dose induced higher antibody levels than the lower dose. The results also showed the durability of the antibody response as all participants in the high-dose group maintained antibody titres of 8 or higher until the end of the study, with the exception of one participant whose CVB2 antibodies decreased to a titre of <4. The cut-off titre 8 corresponds to antibody levels that have been considered protective against another enterovirus, the poliovirus, measured as protection against paralysis as a consequence of blocking the spread of the virus to the central nervous system [33]. The high seroconversion rate among initially seronegative participants (100% across all CVB serotypes in both dosing groups) is another indicator of relevant immunogenicity. The finding that peak neutralising antibody titres occurred after the third vaccination suggests a booster effect of repeated vaccinations.

At the end of the study, 6 months after the third dose of the vaccine, >90% of participants had titres of 8 or higher against all five serotypes. The neutralising antibody levels were comparable with the neutralising poliovirus antibody levels previously seen in children who had received three or four doses of an inactivated poliovirus vaccine a few months before sampling, when using the same plaque reduction assay as in the current study [34]. The magnitude of the PRV-101-induced CVB antibody responses also compares well with those induced by prototype CVB vaccines in our preclinical studies in mice and rhesus macaques [24, 25]. These preclinical studies were carried out using a formalin-inactivated multivalent CVB vaccine similar to PRV-101, and antibodies were analysed using the same plaque reduction assay. The vaccinated animals were efficiently protected against experimental CVB infection.

In the current study, PRV-101 also induced high levels of IgG class CVB antibodies as measured with ELISA. Dose-dependent IgG responses developed rapidly, were already apparent 1 month after the first vaccine injection and were long-lasting, as the IgG levels remained elevated at the end of the study. PRV-101 also induced IgM and, somewhat more variably, IgA class antibody responses but these were less robust than the IgG responses. The IgA responses suggest that PRV-101 may also have the potential to induce mucosal immunity, even if such responses may be weak and vary from one individual to another.

The neutralising antibody response to CVB2 was somewhat weaker than the responses to the other CVB serotypes. This may be due to differences in the proportions of the individual CVB serotype components in the PRV-101 product, as both the protein concentration and the viral particle number of the CVB2 component were lower than those of the other components. In our preclinical animal studies with similarly inactivated CVB vaccines, CVB2 was equally immunogenic as the other CVBs, suggesting that the immunogenicity of CVB2 per se does not markedly differ from the other CVBs. Together with the clear dose–response pattern seen in the high-dose and low-dose PRV-101 groups, these findings suggest that the immunogenicity of each CVB component of PRV-101 may be optimised by adjusting their relative proportions in the vaccine.

PRV-101 was well tolerated in this trial and only relatively mild events, including headache, injection site pain or discomfort and itching were associated with dosing of PRV-101, being in line with the experience from the inactivated poliovirus vaccine. From this point of view, the results from this first-in-human trial support the feasibility of continuing this development programme. Biological proof-of-concept of preventing CVB-induced diabetes by a CVB vaccine has recently been obtained from preclinical mouse studies where the prototype CVB vaccine efficiently protected against experimental CVB infections and against beta cell damage and diabetes that might otherwise occur after a CVB infection [24, 25].

The mechanisms by which CVB infections can cause beta cell damage and clinical diabetes are not fully understood. The prevailing hypothesis is that CVBs infect insulin-producing beta cells, leading to cell damage, local inflammation in pancreatic islets, loss of immune tolerance in a genetically predisposed host and initiation of an autoimmune process. Since PRV-101 is an inactivated vaccine, it cannot cause type 1 diabetes by such a mechanism. However, molecular mimicry between CVBs and host proteins might induce cross-reactive immune responses that could lead to cell damage. In fact, one mimicry epitope has been discovered in the non-structural viral protein 2C and the GAD65 autoantigen expressed in beta cells, and this epitope has been shown to be recognised by the immune system [15, 16]. Importantly, PRV-101 does not contain this epitope since it only consists of structural virus proteins. Nevertheless, some other cross-reactive epitopes may potentially still exist in CVBs, so the trial participants were carefully monitored for signs of type 1 diabetes. None of them developed type 1 diabetes, signs of subclinical beta cell dysfunction or clinically relevant levels of diabetes-associated autoantibodies. One trial participant turned weakly positive for IAA during the study but all follow-up samples from this participant were found to be IAA negative in another laboratory using an RBA similar to that employed in the trial. High-affinity IAAs are predictive of progression to type 1 diabetes. An additional, ECL-based assay for detecting high-affinity IAAs [35, 36] was also negative for this trial participant. The IAA-positive participant also remained negative for all other islet autoantibodies. Altogether, we conclude that this participant had no signs of clinically relevant islet autoimmunity. The low-titre and low-affinity IAAs detected by one laboratory but not confirmed by another laboratory may be considered to lie within the background variation and to have no predictive value for the development of type 1 diabetes. These findings are in line with the safety results from toxicological studies with PRV-101, including a 4 week repeat-dose study in mice (data summary available on request via the corresponding author). In addition, other preclinical studies with similar CVB vaccines have indicated no induction of diabetes or IAA in different mouse models or in rhesus macaques. It is also important to note that PRV-101 did not induce autoantibodies against tissue transglutaminase, which would be predictive for coeliac disease, another autoimmune disease that has been linked to CVB infections. Overall, there was no evidence to suggest that PRV-101 would promote autoimmunity, even in genetically predisposed individuals (almost half of the study participants carried HLA markers predisposing to type 1 diabetes and/or coeliac disease).

This study has some limitations. Evaluation of long-term safety and immunogenicity of PRV-101 is needed, as well as studies of T cell-mediated immune responses to CVB that may help to maintain long-term immunity. Inactivated poliovirus vaccination schedules include later booster vaccinations after the primary vaccination regimen and it is possible that this also holds for PRV-101. Studies evaluating PRV-101-induced T cell responses as well as later-stage CVB antibody levels are currently in progress. Another limitation is that the current study only included adults. Therefore, we cannot exclude the possibility that some of the CVB seronegative individuals had been exposed to CVB in the past and that their antibodies had decreased to undetectable levels by the initiation of the current study. In such cases, the pre-existing memory T cells could have facilitated the immune responses induced by PRV-101. Therefore, subsequent studies in young, exposure-naive children are needed. A critical next step would be to initiate a Phase Ib study of PRV-101 in children, particularly infants, who are the future target population of the vaccine.

In conclusion, this first-in-human, proof-of-mechanism study has demonstrated, for the first time, that a multivalent formalin-inactivated CVB vaccine was both well tolerated and immunogenic in humans. The vaccine induced robust and dose-dependent immune responses, in both male and female participants, towards all five CVB serotypes included in the vaccine. Future studies are planned to progressively increase age and ethnic and regional diversity of the programme. The results of this randomised, placebo-controlled trial support further development of this first-in-class vaccine to prevent CVB infections and several CVB-associated diseases, and potentially also ultimately decrease the global incidence and disease burden of type 1 diabetes and coeliac disease.