Advances in the Development of Vaccines Against Marburg and Ebola Viruses

‘The development of vaccines against MARV and EBOV … is a high priority because of their high virulence, ability to spread from person to person and the possibility for their use in bioterrorism.’

Marburg (MARV) and Ebola (EBOV) viruses are enveloped viruses (family Filoviridae) with a single negative-sense strand of genomic RNA of more than 19 kb. EBOV and, less frequently, MARV emerge sporadically in central Africa in localized zoonotic outbreaks in humans, an accidental host. The natural history and hosts of these viruses are not well understood. They are notorious for causing severe and highly lethal (up to 90% mortality in humans) hemorrhagic disease, and for the lack of any available immunoprophylaxis or treatment.

The development of vaccines against MARV and EBOV (which are sufficiently distinct to necessitate separate vaccines) is a high priority because of their high virulence, ability to spread from person to person and the possibility for their use in bioterrorism. Such vaccines probably would be used in outbreak control rather than mass vaccination.

The challenge

Early preclinical failures with filovirus vaccines based on inactivated virus, purified antigens or vaccinia virus vectors indicated that developing a successful vaccine would not be easily achieved [1]. These vaccines were either completely unsuccessful or demonstrated limited protective efficacy in rodent models of infection. Because nonhuman primates (NHPs) are anatomically and phylogenetically more closely related to humans, and because a very low dose (probably a single infectious unit of MARV or EBOV) is thought to be uniformly lethal, NHPs provide a more predictive model of filovirus infection in the human. Evaluation of these early vaccines in NHP models demonstrated reduced efficacy compared with rodent models. It is generally thought that the major filovirus neutralization antigen, the spike glycoprotein (GP), is poorly immunogenic, for reasons that are not well understood. Vaccination must also contend with the high level of virulence of the filoviruses, their multiple possible portals of entry, their widespread dissemination during infection, and their multiple mechanisms for immune dysregulation and evasion.

For viruses in general, live-attenuated vaccines often induce more robust and broadly reactive immune responses compared with inactivated vaccines. This might reflect a number of factors, including the ability of viruses to deliver foreign antigens into cells (including antigen-presenting cells), to amplify and express antigen over a longer time period, and to more efficiently stimulate innate and cellular immunity. Reverse genetic systems that produce complete recombinant MARV and EBOV have been developed [2,3], providing a method to develop ‘designer‘ live-attenuated vaccine candidates. Given the virulence of filoviruses, it is unclear whether live-attenuated versions would be sufficiently safe and free from reversion. However, partial or complete deletion of viral genes (as opposed to introducing point mutations) may be the way forward for the future development of attenuated vaccine candidates that are phenotypically stable.

‘Vaccination must … contend with the high level of virulence of the filoviruses, their multiple possible portals of entry, their widespread dissemination during infection, and their multiple mechanisms for immune dysregulation and evasion.’

Viral vectors provide an alternative live-vaccine strategy, one that is safer and easier to develop because it uses a well-characterized vector expressing only one or a few protective antigens of the pathogenic virus, and thus does not involve the complete pathogen. Although initial studies using recombinant vaccinia virus as a vector for filovirus antigen were unsuccessful, as noted above, more promising results have been obtained with several other viral vectors.

Preclinical successes

The first vaccine construct that showed substantial protective efficacy in NHPs was developed in 1998 by Hevey et al., and was based on a propagation-defective Venezuelan Equine Encephalitis virus (VEE) replicon, expressing the GP of MARV [4]. A single subcutaneous injection of 10⁷ plaque-forming units (PFU) of GP-expressing replicon resulted in complete protection against viremia, disease and death caused by a very high (8000 PFU) challenge dose of MARV. Recently, vaccine constructs based on VEE constructs expressing EBOV genes have been reported to protect monkeys against EBOV challenge [5], although the details of this study have not yet been published.

Subsequently, Sullivan and coworkers developed replication-defective adenovirus type 5 constructs expressing the EBOV GP or nucleoprotein (NP), the latter included as an antigen for cytotoxic T lymphocytes [6,7]. One or two consecutive immunizations with an equal mixture of vectors expressing GP and NP at a dose of 2 × 10¹² PFU completely protected the animals from viremia, disease and death caused by challenge with either a low (13 PFU) or high (1500 PFU) dose of EBOV [7]. A follow-up study demonstrated that a single immunization with 10¹⁰ PFU is minimally sufficient to protect against viremia, disease and death caused by 1000 PFU of EBOV [8].

‘While each of these constructs appears to be a promising candidate as a filovirus vaccine, none is free of shortcomings.’

Next, vaccine constructs based on a recombinant vesicular stomatitis virus (VSV) were developed in which the major envelope protein G was replaced with MARV or EBOV GP [9]. The resulting constructs were propagation-competent, but attenuated. A single intramuscular immunization of NHPs with 10⁷ PFU of vaccine construct resulted in complete protection against viremia, disease and death caused by 1000 PFU of the corresponding filovirus [10]. Interestingly, when naive NHP were infected with 1000 PFU of MARV or EBOV and innoculated with 2 × 10⁷ PFU of the corresponding vaccine construct 20–30 min later (as a postexposure treatment), 100 and 50% of the animals, respectively, were protected against severe disease and death, although some viremia was evident [11,12].

Both MARV and EBOV can be transmitted by laboratory-generated aerosols [13,14], as well as by droplets and contact, including inoculation of mucosal surfaces of the eyes or the respiratory tract [15,16]. In order to develop a vaccine that would induce a strong local immune response in the respiratory tract, along with a systemic response, a vaccine construct was developed based on propagation-competent human parainfluenza virus type 3 (HPIV3) expressing EBOV GP [17]. A single intranasal and intratracheal inoculation of NHPs with 4 × 10⁶ TCID₅₀ (50% tissue culture infectious doses; 1 TCID₅₀ approximately equals 1 PFU) of the vaccine construct resulted in a high, but not absolute, protection of NHP against viremia, disease and death caused by challenge with 1000 TCID₅₀ of EBOV. Intranasal and intratracheal inoculation with the two doses of 2 × 10⁷ TCID₅₀ of the construct conferred complete protection against viremia, disease, and death caused by 1000 PFU of EBOV [18].

More development is needed

While each of these constructs appears to be a promising candidate as a filovirus vaccine, none is free of shortcomings. For VEE and other alphavirus replicons, concerns exist regarding the possibility of regeneration of an infectious virus due to recombination between the replicon and helper RNAs during preparation [19]. Vectors based on human adenoviruses or HPIV3, which are common human pathogens, may have significantly reduced immunogenicity in the adult human population due to substantial immunity to these viruses from natural exposure [20]. Replication-deficient adenoviruses are expensive to produce, and a very high dose of the vaccine construct is required to achieve a protective response, raising the possibility of reactogenicity in some recipients. VSV may potentially cause encephalitis [21], and its tropism and effects in humans are unclear. Although HPIV3 has a considerable safety record based on clinical studies with attenuated strains, two doses were required for complete, uniform protection.

Nonetheless, most of these limitations can probably be circumvented. Regeneration of infectious virus during production of alphavirus replicons can be avoided by using a bipartite or tripartite helper system [22]. Various approaches are currently being evaluated for the attenuation of VSV, and further studies will evaluate its safety in humans [23]. Possible reductions in the immunogenicity of constructs based on human adenovirus type 5 and HPIV3 due to pre-existing immunity might be less than initially feared, and might also be overcome by use of viral vectors based on rare adenovirus serotypes, or antigenically unrelated animal or avian counterparts [24], or by replacement of the envelope proteins – the major neutralization antigens – with homologous proteins of rare adenovirus serotypes. However, any modifications will need to be carefully evaluated for effects on attenuation and immunogenicity, since a strong immune response appears to be essential for protection against MARV or EBOV.

‘…immunogenicity and efficacy of any given viral vector may be influenced by characteristics of delivery, tropism, replication, gene-expression efficiency and host effects.’

Several explanations can be suggested for the success of these particular viral vectors in protection against MARV and EBOV challenge, as compared with vaccinia virus [1]. Vaccinia virus is a large, complex virus with various mechanisms of evasion of the host immune system by virally expressed proteins, which may antagonize antigen presentation and other steps of the immune response [25]. These mechanisms, as well as competition from the large number of expressed vector antigens, would antagonize the immunogenicity of any foreign antigen expressed by this recombinant virus. By contrast, the present vectors currently under evaluation are less complex (in most cases much less complex) than vaccinia virus and lack its large number of competing antigens and multiple means of immune antagonism. Other effects may also be implicated. For example, VEE replicon particles have intrinsic adjuvant activity [26] and may efficiently infect human dendritic cells [27], facilitating an increased efficiency of antigen presentation. The immunogenicity and efficacy of any given viral vector may be influenced by characteristics of delivery, tropism, replication, gene-expression efficiency and host effects, characteristics that may or may not be fully recognized or understood.

An alternative approach that lacks many limitations of vectored vaccines is the development of virus-like particle (VLP)-based vaccines. VLPs usually consist of at least one envelope protein and at least one internal protein of a virus, which self-assemble into particles that are morphologically similar to a virion but that lack the viral genome and thus do not express genes or propagate. EBOV VLPs, consisting of the viral GP, VP40 (matrix) and NP proteins [28] were recently demonstrated to protect NHP against 1000 PFU of EBOV [29]; however, details of this study have not yet been published. It should also be noted that this type of vaccine has its own limitations, including high production costs and potentially inefficient immune stimulation due to the lack of gene expression and intracellular antigen synthesis.

Conclusion

Dramatic successes in preclinical development of vaccines against MARV and EBOV viruses have been achieved over the past decade. To date, development of vaccines based on certain viral vectors and VLPs appear to be promising approaches. Remarkably, the successful vectored vaccine constructs are based on viruses that represent a number of distinct families and are very diverse biologically. Since the safety, immunogenicity and efficacy of each viral vector depends on multiple factors, it is not possible at this time to predict whether a particular vector will be safe and successful in inducing a protective MARV or EBOV-specific response in humans. It is therefore encouraging to have a diverse panel of candidates. It is also encouraging that the VSV-based constructs have substantial efficacy in NHP in postexposure treatment.

Future perspective

In the next 5 years, the current vaccine candidates will undergo detailed evaluations in NHPs and, as appropriate, in clinical studies. Modifications may be necessary, such as those that increase the level of attenuation or reduce sensitivity to pre-existing immunity. Additional vaccine candidates may also become available for evaluation. Although safety and immunogenicity can be evaluated in a clinical trial, evaluation of efficacy in humans will depend on outbreak situations. Viral vectors also have use in the development of vaccines against other highly pathogenic agents such as avian influenza virus as well as possible future emerging pathogens. Since any particular vaccine vector probably cannot be ‘reused‘ due to the host immunity that develops against the vector, it would be ultimately advantageous to develop a number of effective vectors.

Financial & competing interests disclosure

The authors‘ studies were funded as a part of the Intramural Research Program of NIAID, NIH.

The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

The authors are grateful to Brian Murphy and Alan Schmaljohn for careful reading of the manuscript.

Additional information

Funding

The authors‘ studies were funded as a part of the Intramural Research Program of NIAID, NIH.