While the natural hosts of VSV are insects and livestock, a few incident cases have occurred in humans as a result of high-risk occupational exposure (i.e. laboratory workers, farmers, veterinarians) [22, 23]. Infected humans may be asymptomatic or may experience a mild febrile illness with symptoms lasting 2–5 days [23]. The low incidence of infection and disease results in an overall very low level of pre-existing immunity to the virus among the general human population. Areas of exception include rural communities of Central America where both predominant serotypes VSV-New Jersey (VSV-NJ) and VSV-Indiana (VSV-IN) are endemic [24, 25]. Other areas of note include the enzootic regions of coastal Georgia where seroprevalence of humans to VSV-NJ was approximated at 30% in the early 20^th century [26].

Viral vector vaccines should demonstrate stability of foreign gene expression to ensure high-level expression of the target antigen(s). VSV has a simple genome of 11KB encoding five major proteins. Transcriptional attenuation of approximately 30% occurs at each successive gene junction resulting in a pronounced 3’ to 5’ gradient of gene expression [27–30]. Therefore, the genomic site of foreign gene insertion strongly influences antigen expression levels. Minimal conserved nucleotide sequences (transcription start and stop signals) are required for normal gene expression [31] and foreign gene inserts must be flanked by these sequence elements.

Although there are no apparent structural limitations on the size of foreign gene insert for the VSV vector, larger inserts appear to reduce the rate of viral replication in animal models. For example, rVSVGagEnv encoding both the HIV envelope (Env) and group specific antigen protein (Gag) contributing approximately 4.4 kilobases (kb) of additional genomic sequence, modestly reduced viral titers by three-fold [32]. Since then, a larger insert of approximately 6 kb encoding Hepatitis C virus non-structural proteins (NS) has been expressed by a rVSV NJ vector, leading to a five-fold reduction in viral titer [33]. It is, however, also likely that some foreign gene products may further inhibit rVSV replication by other mechanisms such as biological activity, targeting and transport, or unforeseen toxicity.

The pathogenicity of VSV has been attributed in part to the glycoprotein (VSV G), as virulence is dependent on the ability of G protein to bind cellular receptors, and mediate entry and fusion with endocytic vesicles to initiate the replicative cycle [34]. Due to pivotal roles in receptor binding and membrane fusion, it has been a target for attenuation of rVSV vector vaccines. Replacement of the G gene with that of another foreign gene product acting as a viral receptor can generate rVSVΔG pseudotypes with altered cell tropism, which may also have attenuating effects. Foreign glycoproteins expressed by these pseudotypes are prime targets for cell-mediated and humoral immunity [35, 36]. Thus far, rVSV and rVSVΔG vectors expressing influenza hemagglutinin (HA) and Ebola/Marburg glycoproteins have demonstrated full protection against virus challenge and are non-pathogenic in mouse and non-human primate (NHP) disease models [37–44]. The strategy of using rVSV pseudo-typed with Ebola virus GP as a vaccine to combat Ebola virus induced disease has recently completed clinical testing and will be discussed in a separate vector analysis template due to the unique properties of the vector conferred by the Ebola virus GP protein as sole virus receptor. In vitro and in vivo attenuation of rVSV has also been demonstrated by truncation of the cytoplasmic tail (CT) of the G protein from 29 amino acids found in nature, to only 9 or 1 amino acids (CT9 and CT1 respectively) [17, 42, 45]. It is generally thought that this attenuation mechanism acts by impairing the interaction of the G CT with underlying viral core proteins, thereby reducing the efficiency of virus particle maturation and budding.

Another major approach to rVSV attenuation relies on down-regulation of expression of one or more key viral structural proteins. This attenuation strategy has been demonstrated for rVSV by translocation of the N gene further away from the 3’ transcription promoter to positions 2, 3 and 4 in the genome [28, 29]. The resulting step-wise reduction in N protein expression leads to corresponding incremental reduction of viral replication in vitro and reduced pathogenesis in a natural host [30].

Attenuation by either CT truncation or N gene translocation separately could not provide sufficient reduction in neuropathology in stringent murine and NHP neurovirulence (NV) models to support testing of rVSV as a vaccine vector in humans [46–48]. However, when both forms of attenuation were combined there was a dramatic and synergistic increase in vector attenuation, almost completely eliminating clinical and microscopic pathology following intra-cranial injection of mice and NHPs [47, 49, 50].

One additional attenuation mechanism relies on either mutation or deletion of amino-acid 51 of the VSV M protein These VSV M mutants grow quite robustly in cell culture but demonstrate a marked reduction of virulence in vivo. It is thought that the attenuating mutation(s) reduce the ability of virus to shut down host innate immune responses which normally restrict virus growth in vivo [51–53].

Studies using rVSVΔG vectors expressing Ebola and Marburg virus glycoproteins achieved post-exposure prophylaxis in both rodent and NHP models [20]. If administered in one dose within 24 hours of virus challenge, 50–100% of both guinea pigs and mice were protected. Similarly, there was 50% protection of NHPs if treatment was administered within 30 minutes of challenge.

A live viral vaccine safety standard for all licensed vaccines requires assessment of viral NV by intracranial inoculation of NHPs with the vaccine [54, 55]. Vaccines for measles, mumps, yellow fever, polio and others have all been assessed for NV by this method [56–59]. A pilot NV study in NHPs demonstrated that prototypic rVSV vectors expressing HIV gag and env were not adequately attenuated for clinical evaluation [48]. However, extensive testing in mouse NV studies and two additional, sequential NHP NV studies led to the identification of rVSV vectors that were safe for clinical testing [49, 50]); one of these highly attenuated vectors known as rVSVN4CT1gag1 was selected for a first in man clinical trial. The rVSVN4CT1gag1vector was attenuated by translocation of the N gene to the 4^th position in the genome (N4), truncation of the G protein CT to a single amino acid (CT1) and the gag gene was located in the 1^st position of the genome (gag1) to maximize gag protein expression. The rVSVN4CT1gag1 vector has now demonstrated safety and immunogenicity in phase 1 clinical trials [18] and the rVSVN4CT1 expressing Ebola virus GP is on a clinical development pathway as a candidate Ebola virus vaccine [41].

To provide clinical trial materials (CTM) for Phase 1 studies, an HIV-1 vaccine production process was developed in a 10L bioreactor under good manufacturing practices (GMP). An approved Vero cell line was used as substrate for vaccine vector amplification. Following infection, culture medium from infected cells was harvested once cell cytopathology was extensive (80 – 100%), and centrifuged to remove cellular debris. This unprocessed harvest material (UHM) was then conditioned with a virus stabilizer at a final concentration of 7.5% sucrose, 3.8mM KH₂PO₄, 7.2mM K₂HPO₄ and 5mM L-Glutamate (SPG) and passed through an anion exchange membrane which binds rVSV particles. The membrane was then rinsed to remove cellular proteins, and DNA and virus particles were eluted in a high salt buffer. The high salt eluate was exchanged with a low salt phosphate buffer suitable for injection by a process of tangential flow ultra-filtration. The resulting virus preparation was then formulated with SPG and 0.2% hydrolyzed gelatin as additional virus stabilizer, sterile-filtered, and dispensed in vials as drug product (also known as CTM). CTM was stored frozen at −70°C to −80°C until ready for injection. CTM material generated by this process (or equivalent material generated by the same process) underwent toxicology testing in rabbits under GMP. Data from the toxicology study, the results of compendial safety tests performed at all key stages of vaccine manufacturing, and all data from pre-clinical development and safety testing of the rVSVN4CT1gag1 vector, were submitted to the FDA as part of an investigational new drug (IND) application in 2011. The FDA approved the rVSVN4CT1gag1 vector for clinical evaluation, and enrollment for HVTN 090, a Phase 1, double blinded, placebo controlled clinical trial began in October 2011, marking the first time an rVSV vaccine vector was administered to healthy trial participants. Data from this first in human trial have now been published [18]. The rVSVN4CT1gag1 vector has also demonstrated safety and immunogenicity in a second HIV-1 Phase 1 clinical trial as part of a pDNA prime, rVSV boost, vaccination regimen (HVTN 087: http://clinicaltrials.gov/).

The safety and immunogenicity of the rVSVN4CT1gag1 vector in animal models and in clinical trials has demonstrated the potential of rVSV vectors targeting other infectious diseases. Robust and stable gene expression, a safe, attenuated phenotype, and induction of foreign antigen-specific immune responses, support further development of rVSV and other vesiculoviruses as platforms for vaccine development.