Tetravalent Rabies-Vectored Filovirus and Lassa Fever Vaccine Induces Long-term Immunity in Nonhuman Primates

Abstract

Background

The objective of this study is to evaluate the immunogenicity of adjuvanted monovalent rabies virus (RABV)–based vaccine candidates against Ebola virus (FILORAB1), Sudan virus (FILORAB2), Marburg virus (FILORAB3), Lassa virus (LASSARAB1), and combined trivalent vaccine candidate (FILORAB1–3) and tetravalent vaccine candidate (FILORAB1–3 and LASSARAB) in nonhuman primates.

Methods

Twenty-four Macaca fascicularis were randomly assigned into 6 groups of 4 animals. Each group was vaccinated with either a single adjuvanted vaccine, the trivalent vaccine, or the tetravalent vaccine at days 0 and 28. We followed the humoral immune responses for 1 year by antigen-specific enzyme-linked immunosorbent assays and RABV neutralization assays.

Results

High titers of filovirus and/or Lassa virus glycoprotein-specific immunoglobulin G were induced in the vaccinated animals. There were no significant differences between immune responses in animals vaccinated with single vaccines vs trivalent or tetravalent vaccines. In addition, all vaccine groups elicited strong rabies neutralizing antibody titers. The antigen-specific immune responses were detectable for 1 year in all groups.

Conclusions

In summary, this study shows the longevity of the immune responses up to 365 days for a pentavalent vaccine—against Ebola virus, Sudan virus, Marburg virus, Lassa virus, and RABV—using a safe and effective vaccine platform.

rabies virus, filorviruses, Lassa fever virus, multivalent vaccine, long-lived immunity

Viral hemorrhagic fever (VHF) diseases—caused by Ebola virus (EBOV), Sudan virus (SUDV), Marburg virus (MARV), and Lassa virus (LASV)—occur in parts of sub-Saharan Africa [1]. These zoonotic diseases are transmitted among humans through contact with body fluids. Filoviruses (EBOV, SUDV, MARV) are highly pathogenic, with lethality reaching 80% in humans. Outbreaks have been documented since the 1960s, the largest of which being the 2014‒2015 EBOV outbreak in West Africa. This outbreak resulted in more than 28 600 cases and 11 325 deaths [2]. The second-largest EBOV outbreak began in August 2018 in the Democratic Republic of the Congo (DRC) and, as of 3 July 2020, has resulted in 3481 confirmed cases and 2299 confirmed deaths [3]. These outbreaks have accelerated the development of EBOV vaccines, and recently a vesicular stomatitis virus (VSV)–based vaccine has been approved by the United States (US) Food and Drug Administration (FDA). However, there are currently no approved vaccines against Sudan and Marburg filoviruses, which also pose serious health threats.

LASV is an emerging virus belonging to the family of arenaviruses. In contrast to filoviruses, LASV infection is often asymptomatic, but if clinical signs are observed, lethality reaches up to 40% [4]. LASV is a significant cause of abortion and deafness in humans in West Africa, where large outbreaks regularly occur [5]. There are currently no approved vaccines against LASV in the clinic.

Rabies virus (RABV) is endemic in Africa, where VHF viruses are prevalent. It is also a serious concern worldwide, accounting for up to 55 000 human deaths annually, mostly in Asia and Africa [6, 7]. The rabies vaccine is 100% effective when administered as a preexposure prophylactic as well as a postexposure prophylactic (when administered as soon as exposed). Despite the effectiveness of the rabies vaccine, it is not cost-effective for most countries.

Currently, only a combination of 4 live-attenuated VSV vaccines in a boosted regimen has been able to provide protective immune responses against all 3 filoviruses as well as LASV [8]. Such live virus vaccines are limited in use due to high storage costs and safety concerns, especially when administered to pregnant women, children, and immunocompromised individuals. Additionally, the introduction of the live VSV vaccine to nonendemic areas could have consequences for wildlife there, given that VSV causes disease in cattle, horses, and swine. The VSV-based vaccine against EBOV was found to provide long-term protection against EBOV in nonhuman primates (NHPs), but these findings were challenged in recent clinical studies [9, 10]. The other lead candidate for an EBOV vaccine requires a boost with another virus, modified vaccinia virus Ankara (MVA), to provide long-term immunity [9].

Our laboratory utilizes the attenuated rabies vaccine strain SAD-B19 (BNSP) as the backbone to develop vaccines against infectious diseases. This vector was furthur attenuated to reduce neurotropism by introduction of a mutation at position 333 (arginine to glutamic acid mutation) within the rabies glycoprotein (GP) (BNSP333) [11]. To further increase the safety profile of our vaccines, we chemically inactivate our recombinant rabies-vectored viruses with β-propiolactone (BPL). We generated the RABV-vectored EBOV vaccine called FILORAB1 that induces immunity against EBOV as well as RABV and is one of the most extensively tested experimental anti-EBOV vaccines. Previously, FILORAB1 vaccine was protective against challenge in 3 large NHP studies by intramuscular (IM) injection and is immunogenic in chimpanzees by oral administration [12–15]. In addition to FILORAB1, our laboratory has also developed similar RABV-based vaccines against SUDV (FILORAB2), MARV (FILORAB3), and LASV (LASSARAB). We observed the FILORAB2 vaccine to be immunogenic in mice, FILORAB3 to be protective in mice, and LASSARAB protective in mice and guinea pigs [13, 14, 16–18].

We are currently developing an inactivated tetravalent vaccine combining these vaccines. It should protect against these 5 highly lethal viruses (all 4 VHF viruses plus RABV) and will likely be suited for the general population, including children and pregnant women. In this current study, we analyzed our adjuvanted inactivated RABV vaccine platform against EBOV, SUDV, MARV, or LASV GPs as single vaccines or as a blended trivalent or tetravalent vaccine for 1 year in NHPs. We observed high antibody titers of antigen-specific antibodies against the EBOV, SUDV, MARV, LASV, and RABV GPs that were unaffected when administered in a blended vaccine formulation. This study is the first to demonstrate the potential of an inactivated RABV vaccine platform against multiple antigens in a blended formulation.

MATERIALS AND METHODS

Antigens

All of the constructs (Figure 1) were previously engineered, characterized, and assessed in small animal and/or NHP models [13, 14, 17–19]. In brief, the codon optimized amino acid sequences of the following GPs were used in vaccine vector and antigen construction: EBOV (Mayinga strain, GenBank accession number AF499101.1), SUDV (Gulu strain, GenBank accession number AY729654.1), MARV (Angola strain, GenBank accession number KU978782.1), and LASV (Josiah strain, GenBank accession number MH778559).

Figure 1.

Schematic of vaccine constructs. We utilized the attenuated rabies vaccine strain SAD B19 as the backbone to generate our recombinant vaccines. The rabies genome encodes for 5 genes: N (nucleoprotein), P (phosphoprotein), M (matrix protein), G (glycoprotein), and L (RNA-dependent RNA polymerase). We previously introduced a trancription cassette with restriction sites BsiWi and NheI between N and P genes of the SAD B19 vaccine vector. The 333 (arginine to glutamic acid; shown as ★) mutation in the rabies glycoprotein (G) further reduces its neurotropism. In this study we used the chemically inactivated rabies-vectored Ebola (FILORAB1), Sudan (FILORAB2), Marburg (FILORAB3), and Lassa (LASSARAB) virus vaccines that express the viral hemorrhagic fever GPs inserted between the N and P genes of the rabies vaccine vector.

Vaccines

The recombinant rabies vaccines FILORAB 1, 2, 3 and LASSARAB were constructed, recovered, purified with sucrose, inactivated with BPL, and characterized as previously described [12–14, 17, 18, 20–23]. The vaccines were stored at –80°C.

Animal Ethics Statement

A statistician randomly assigned animals to the vaccine groups. The animal study was blinded to the research staff. All animal studies were approved by the AlphaGenesis Animal Care and Use Committee and adhere to the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Study Design

The animal study was conducted at Alpha Genesis Inc (South Carolina). The objective of this study was to assess the difference in EBOV-, SUDV-, MARV-, and LASV-specific immunogenicity between the monovalent and multivalent vaccine groups. Twenty-four healthy, RABV-, filovirus-, and LASV-naive adult (~1.75 to 3.44 kg) cynomolgus macaques (Macaca fascicularis; 16 of them were imported from China, 8 born at AlphaGenesis) were randomized into 6 groups of 4 experimental animals (2 females and 2 males/group) and blood samples were collected for prescreening. The animals were presecreened for simian retrovirus, simian immunodeficiency virus, simian T-cell leukemia virus, herpes B virus, and tuberculosis at AlphaGenesis. Sera were then assessed for EBOV-, SUDV-, MARV-, LASV-, and RABV-specific background cross-reactivity (immunoglobulin G [IgG]) by enzyme-linked immunosorbent assay (ELISA) at Thomas Jefferson University. The absence of RABV cross-reactivity was further verified by the RABV neutralization assay by rapid fluorescent focus inhibition test (RFFIT).

The experimental animals were vaccinated by IM injection with a total volume of 1 mL (vaccine + adjuvant + 1× Dulbecco's phosphate-buffered saline). The 6 groups were vaccinated on day 0 and boosted on day 28 with 100 μg of either FILORAB1 (group 1), FILORAB2 (group 2), FILORAB3 (group 3), LASSARAB (group 4), FILORAB1–3 (group 5), or FILORAB1–3 and LASSARAB (group 6). All groups were adjuvanted with a Toll-like receptor 4 agonist, glucopyranosyl lipid A [13, 24], formulated in a stable oil-in-water nanoemulsion (SE) (15 μg in 4% SE). Blood was sampled from the vaccinated animals on days 0, 14, 28, 56, 84, 168, and 365 (Figure 2). All animals were given physical examinations and monitored daily.

Figure 2.

Study strategy. Cynomolgus macaques (2 male and 2 female) were vaccinated on day 0 and boosted on day 28 with 100 μg of either FILORAB1 (group 1), FILORAB2 (group 2), FILORAB3 (group 3), LASSARAB (group 4), FILORAB1–3 (group 5), or FILORAB1–3 and LASSARAB (group 6). All groups were adjuvanted with a Toll-like receptor 4 agonist, glucopyranosyl lipid A, formulated in a stable oil-in-water nanoemulsion (SE) (15 μg in 4% SE). Blood was sampled from the vaccinated animals on days –28, 0, 14, 28, 56, 84, 168, and 365.

Antigen Production for ELISA

The production of human influenza hemagglutinin (HA)–tagged filovirus GPs was described previously [13, 14, 17]. In brief, subconfluent T175 flasks of 293T cells (human kidney cell line) were transfected with a eukaryotic expression vector (pDisplay) that expresses HA-tagged filovirus GPs: human codon–optimized EBOV-GP encoding amino acids 33–632 of the ectodomain (EBOV-GP-ΔTM, Mayinga strain with an Igκ signal peptide, plasmid kindly provided by Erica O. Saphire of the Scripps Research Institute, La Jolla, California), the codon-optimized SUDV-GP (SUDV-GP-ΔTM, Gulu strain) encoding amino acids 1–650 of the ectodomain, or codon-optimized MARV-GP (MARV-GP-ΔTM, Angola strain) encoding amino acids 1–643 of the ectodomain.

Stripped RABV-GP (G) and LASV-GP complex (GPC) antigen were generated by infecting BEAS-2B cells with either recombinant VSV (rVSV)–ΔG-GPC (for LASV GPC antigen) or rVSV-ΔG-RABV-G-GFP (RABV G antigen) in Opti-Pro SFM (Gibco). Viral supernatants were concentrated 50× by KrossFlo tangential flow filtration using a 500-kDa hollow fiber cartridge. The concentrated virus was centrifuged at 110 000g through a 20% sucrose cushion. Viral pellets were then resuspended in TEN buffer (100 mM NaCl, 100 mM Tris, 10 mM ethylenediaminetetraacetic acid pH 7.6) containing 2% octyl β-D-glucopyranoside detergent and incubated for 30 minutes while shaking at room temperature. The mixture was then centrifuged at 3000g, and the supernatant was collected and further centrifuged at 250 000g for 90 minutes. Finally, the supernatant was collected and analyzed by sodium dodecyl sulphate–polyacrylamide gel electrophoresis and Western blot for LASV GP1 and GP2 or RABV G presence.

Filovirus Anti-GP, LASV Anti-GPC, and RABV-G IgG ELISA

Immulon 4 HBX 96-well flat-bottom Microtiter plates (ThermoScientific, catalog number 3855) were coated for 20 hours ± 4 hours at 4°C with 50 ng/well of recombinant filovirus GP or LASV-GPC or RABV-G, diluted in 0.1 mL of 1× coating buffer (15 mM Na₂CO₃, 35 mM NaHCO₃). The plates were washed 3 times with 250 μL of 1× phosphate-buffered saline (PBS) containing 0.05% Tween-20 and blocked for 2 hours ± 30 minutes at 21°C ± 3°C with 5% milk in 1× PBS containing 0.05% Tween-20. The plates were washed 3 times with 250 μL of 1× PBS containing 0.05% Tween-20 prior to the addition of NHP serum. The NHP serum samples, diluted with 1× PBS containing 0.5% bovine serum albumin, 0.05% Tween-20, and 0.05% sodium azide, were added to the ELISA plates at a starting dilution of 1:50 and further diluted 3-fold down the plates. Plates were kept at 4°C overnight 20 hours ± 4 hours and washed afterward 3 times with 250 μL of 1× PBS containing 0.05% Tween-20, followed by 2 hours ± 30 minutes at 21°C ± 3°C with 100 μL/well horseradish peroxidase–conjugated goat anti-human IgG antibody (Southern Biotech, 2040-05) diluted 1:20 000 (for EBOV-GP, SUDV-GP, LASV-GPC, and RABV-G) or 1:10 000 (for MARV-GP) in 1× PBS containing 0.05% Tween-20. The plates were then washed 3 times with 250 μL of 1× PBS containing 0.05% Tween 20 and developed by the addition of 200 μL/well of o-phenylenediamine dihydrochloride substrate (Sigma-Aldrich, P9187). The reaction was stopped after 13 minutes (for EBOV-GP, SUDV-GP, and RABV-G) and 15 minutes (for MARV-GP and LASV-GPC) by adding 50 μL/well of 3 M sulfuric acid (Fisher A300SI-212).

The plates were read at the absorbance wavelength of 630 nm (background) and 490 nm (experimental) on a BioTek EL×800 plate reader with GEN5 software. The 630 nm reading was subtracted from the 490 nm reading to calculate the delta value. The delta value was exported and analyzed in GraphPad Prism software. Endpoint titers were determined using cutoff values calculated from negative control sera. For each row/dilution, a separate cutoff was calculated using the formula:

$${\displaystyle \begin{array}{l}{\displaystyle Cutoff=X+SDf}\end{array}}$$

where X is the mean of negative control sera readings and SD is the standard deviation [25]. 𝑓 depends on the number of negative control samples used (n). Sample dilutions at each cutoff value were determined by interpolation from the 4-parameter curve fit. The endpoint titer was reported as the reciprocal sample dilution at the lowest cutoff value that gives 1 sample dilution higher than the dilution of the negative control from which the cutoff was derived. The 50% effective concentration (EC₅₀) values were calculated by interpolation, and antibody concentrations were expressed as reciprocal serum dilutions at the EC₅₀ value.

Rabies Virus Neutralization by RFFIT

Rapid fluorescent focus inhibition tests were performed as previously described [26]. In brief, mouse neuroblastoma cells (NA) cultured in serum-enriched RPMI media were seeded in 96-well plates and incubated for 48 hours. Separately, individual NHP sera from days 0, 84, and 365 were serially diluted 3-fold starting at a 1:20 dilution. A prediluted mixture of rabies virus (strain CVS-11, previously determined to achieve 90% infection in confluent NA cells) was added to each serum dilution. The US standard rabies immune globulin (World Health Organization standard [WHO STD]) was used at a starting dilution of 2 IU/mL. The sera/virus or WHO STD/virus mixture was incubated for 1 hour at 34°C. The medium was then removed from the NA cell plate, replaced with the sera/virus mix, and incubated for 2 hours at 34°C. Postinfection, the sera/virus mixture was aspirated and replaced with fresh medium. The plates were then incubated for 22 hours (24 hours’ total infection) at 34°C. After incubation, cells were fixed with 80% acetone, dried, and stained with fluorescein isothiocyanate Anti-Rabies N Monoclonal Globulin (Fujirebio) for >4 hours. Wells were assessed for percentage infection using a fluorescent microscope. Fifty percent endpoint titers (EPTs) were calculated using the Reed–Muench method and converted to international units per milliliter by comparing the sample 50% EPTs to that of the US standard rabies immune globulin.

Statistical Analysis

For the ELISA, log-transformed EC₅₀ or endpoint titers were plotting against delta optical density value (490–630 nm). One-way analysis of variance with post hoc Tukey honest significance difference test was performed on log-transformed data for each time point.

RESULTS

Immunogenicity of Multivalent Vaccines in Macaques

Sera from vaccinated animals were assessed for EBOV, SUDV, MARV, LASV, and RABV GP–specific IgG responses over time. EBOV-GP, SUDV-GP, MARV-GP, LASV-GPC, and RABV-G responses (EC₅₀ titers) were detected as early as day 14 in groups 1–6, respectively (Figure 3). All of the animals in the 6 groups demonstrated the highest EBOV-GP, SUDV-GP, MARV-GP, LASV-GPC, and RABV-G IgG antibody EC₅₀ and endpoint titers on day 56, having been boosted on day 28 (Figures 3 and 4). Undetectable IgG responses were observed in 1 animal in group 6 (Figure 3: plum-colored inverted triangle) by day 365 for SUDV IgG and day 275 for MARV IgG, although this animal maintained detectable EBOV, LASV GPC, and RABV-G responses (Figure 3). The trivalent (group 5) and tetravalent (group 6) groups exhibited EBOV, SUDV, and MARV IgG EC₅₀ and endpoint titers that were not significantly different from the single vaccine groups (1, 2, and 3) at all time points (days 14, 28, 56, 84 168, and 365) (Figures 3 and 4). The tetravalent vaccine group (group 6) exhibited LASV-GPC IgG EC₅₀ and endpoint titers not significantly different from the single LASSARAB vaccine group 4 (Figures 3 and 4). EBOV and SUDV cross-reactive antibodies were observed in the FILORAB1- and FILORAB2-vaccinated animals, with higher cross-reactive titers seen in the FILORAB1 group (Figure 4). High LASV-GPC IgG background signal was seen on day 14 in the FILORAB2-vaccinated animals in groups 2 and 5, which waned after day 28 (Figure 4). Similar RABV-G IgG EC₅₀ and endpoint titers were elicited in groups 1–6 for all time points (Figures 3 and 5A). All groups exhibited a significant reduction in antibody titers after day 56, with an approximate 70%–80% reduction from peak EC₅₀ titers by day 365. No interference in the antibody responses was observed in any of the blended vaccine groups in comparison to their monovalent groups (Figures 3 and 4)

Figure 3.

Viral hemorrhagic fever and rabies G–specific immunoglobulin G (IgG) 50% effective concentration (EC₅₀) titers over time. Columns 1–5 depict glycoprotein-specific EC₅₀ of the antibody (IgG) responses to Ebola, Sudan, Marburg, Lassa, and rabies virus for each individual animal in every group at all time points up to 1 year. The log-transformed EC₅₀ titers were plotting against delta optical density value (490–630 nm). Mean EC₅₀ titers of triplicate values for each animal are depicted. Abbreviations: EBOV, Ebola virus; EC₅₀, 50% effective concentration; GP, glycoprotein; LASV, Lassa virus; MARV, Marburg virus; ND, none detected; RABV, rabies virus; SUDV, Sudan virus. Animal numbers depicted by - circle = 1, square = 2, triangle = 3, inverted triangle = 4.

Figure 4.

Viral hemorrhagic fever glycoprotein-specific immunoglobulin G (IgG) endpoint titers over time. Columns 1–4 depict glycoprotein-specific endpoint titers (EPTs) of the antibody (IgG) responses to Ebola, Sudan, Marburg, and Lassa virus for each individual animal in every group at all time points up to 1 year. The log-transformed EPTs for each animal were plotting against delta optical density value (490–630 nm). One-way analysis of variance with post hoc Tukey honest significant difference test was performed on log-transformed data for each time point. Mean EPTs for each group were compared for each time point., *P < .0332; **P < .0021; ***P < .0002; ****P < .0001; P > .1234 (not significant [ns]). Abbreviations: EBOV, Ebola virus; GP, glycoprotein; LASV, Lassa virus; MARV, Marburg virus; SUDV, Sudan virus. Animal numbers depicted by - circle = 1, square = 2, triangle = 3, inverted triangle = 4.

Figure 5.

Rabies glycoprotein–specific immune responses. A, Rabies glycoprotein–specific endpoint titers (EPTs) of the antibody (immunoblogulin G) responses for each individual animal in every group at all time points up to 1 year. The log-transformed EPTs for each animal were plotting against delta optical density value (490–630 nm). One-way analysis of variance with post hoc Tukey honest significant difference test was performed on log-transformed data for each time point. Mean EPTs for each group were compared for each time point. *P < .0332; **P < .0021; ***P < .0002; ****P < .0001; P > .1234 (not significant [ns]). B, Rabies neutralizing titers on days 84 and 365 by rapid fluorescent focus inhibition test. Individual animals are depicted for each group. C, Correlation between rabies G 50% effective concentration and the neutralizing titers (IU/mL). r² depicts the Pearson correlation coefficient. Abbreviations: EC₅₀, 50% effective concentration; ELISA, enzyme-linked immunosorbent assay; RFFIT, rapid fluorescent focus inhibition test. Animal numbers depicted by - circle = 1, square = 2, triangle = 3, inverted triangle = 4.

Rabies Neutralization by RFFIT

RABV neutralization assays were performed on sera collected on days 0, 84, and 365. All groups elicited high RABV neutralizing titers above the WHO-suggested protective threshold of 0.5 IU/mL on both days tested. Most animals in groups 1–6 showed a decrease in RABV neutralizing titers from days 84 to days 365, with the exception of 3 animals in group 4 (Figure 5B). A strong correlation (r² = 0.732 for day 84 and r² = 0.826 for day 365) was observed between the RABV IgG EC₅₀ titers and the RFFIT neutralizing titers.

DISCUSSION

Global efforts have focused on controlling VHF diseases caused by WHO-defined priority pathogens: EBOV, SUDV, MARV, and LASV. The FDA has approved 2 EBOV vaccines, both vectored in live-attenuated viruses: rVSV-ZEBOV (Ervebo) and a vaccine combining 2 vectors: the first component is the Ad26.ZEBOV derived from human adenovirus serotype 26 (Ad26) expressing the Ebola virus Mayinga variant GP, while the second component MVA-BN is the MVA–Bavarian Nordic filovector [27]. The absence of any approved active or passive intervention strategies against other filoviruses—SUDV and MARV—and Lassa fever, despite the recurrence of outbreaks in Africa, continues to result in deaths. In June 2020, the DRC declared a new Ebola outbreak, while in early January 2020, a Lassa fever outbreak in Nigeria led to 472 confirmed cases and 70 deaths [28].

EBOV, SUDV, MARV, and LASV have overlapping endemic ranges, so the ideal prophylactic vaccine for use in Africa would confer protection against all 4 pathogens. So far, 2 prior studies have reported using such multiantigen vaccines in NHPs using a VSV vaccine vector and assessing short-term immune responses. The Marzi et al study primed with a VSV-based LASV vaccine and boosted with a VSV-based EBOV vaccine [18]. The Cross et al study evaluated the quadrivalent VesiculoVax vaccine against EBOV, SUDV, MARV, and LASV [29, 30]. The day 28 endpoint titer responses against filovirus GPs seen in the VesiculoVax study are comparable to the responses seen in our study. Compared to our study, the boosted responses for the filovirus GPs on day 66 were much higher in the VesiculoVax study. However, it should be noted that the VesiculoVax study had a higher limit of quantitation (1:17 714 700) in comparison to our study (1:109 350). Interestingly, the VesiculoVax study induced low LASV GPC titers until the boost. The boosted LASV GPC responses on day 66 are comparable to the titers induced in our study on day 365. The EBOV GP and LASV GPC endpoint titers observed in the Marzi et al study were much lower than those generated in our study. Additionally, Geovax, in collaboration with the University of Texas Medical Branch and Battelle Memorial Institute, has received funding from the National Institute of Allergy and Infectious Diseases to test their MVA platform against EBOV, SUDV, and MARV. While both the MVA and the VSV platforms have elicited robust protective T-cell and antibody responses in animals and humans, they are live viral vaccine platforms that can pose significant safety concerns within the human immunodeficiency virus–endemic populations of Africa [31].

Our study demonstrates that 2 doses of our inactivated RABV-vectored VHF vaccines induced stable filovirus and LASV antigen-specific immunity for up to 1 year. The trivalent filovirus and the tetravalent VHF vaccines induced high EBOV, SUDV, and RABV antibody responses. The monovalent or tetravalent vaccines induced lower antibody titers against MARV and LASV compared to titers against EBOV and SUDV. The reduction in antigen-specific EC₅₀ titers after the boost until the study’s duration mimics the responses in the Callendret et al study, which reported higher EBOV-GP and SUDV-GP responses than MARV-GP responses [32]. While our study has not assessed the presence of neutralizing titers, prior studies have indicated that rabies-vectored VHF vaccines induce nonneutralizing antibodies, as seen for FILORAB1 in NHPs, FILORAB3 in mice, and LASSARAB in mice and guinea pigs [13, 17, 18]. Although immune thresholds for protection have not yet been established for hemorrhagic fever viruses, antibodies induced by FILORAB1 have proven protective in several NHP challenge studies.

We have previously demonstrated the protective efficacy of FILORAB1, the inactivated RABV-vectored EBOV vaccine used in this study, in several NHP studies [13, 14]. Protection similar to EBOV is expected in FILORAB2-vaccinated animals due to high SUDV IgG responses generated [14, 33]. The RABV-vectored MARV and LASV vaccines have been extensively tested for efficacy in mouse and guinea pig models, respectively [17, 18]. Despite seeing protection in LASSARAB-vaccinated animals against the guinea pig–adapted Josiah strain of LASV, the development of a multivalent vaccine is complicated as LASV has 4–6 distinct lineages. It is unknown whether antibodies against 1 LASV lineage can protect against lethal challenge with an isolate from a heterologous lineage of LASV.

In addition to preventing the VHF diseases, all of these RABV-vectored vaccines also protect against RABV, another deadly disease that also poses a serious health threat in Africa. RABV kills an estimated 25 000 people annually in Africa [34]. Children are the most affected by the disease, with 4 of every 10 deaths occurring in children <15 years of age [35]. Therefore utilization of our rabies-vectored VHF vaccine would target 5 deadly pathogens—EBOV, SUDV, MARV, and LASV, as well as RABV.

Additionally, an ideal vaccine would be stable at room temperature and elicit a protective immune response with fever booster doses, given the remote locations where VHF agents often circulate, the shortage of healthcare professionals and clinics, and the limited mobility of human populations. FILORAB1 was observed to be stable at temperatures of 4°C and 37°C for 6 months and 50°C for 2 weeks [19]. Since FILORAB2, FILORAB3, and LASSARAB are prepared similarly, they will also likely be stable at these temperatures.

While this study assessed immunogenicity, further studies will be performed to determine the efficacy of the tetravalent vaccines in challenge studies. The long-term immunity detected in this study supports the further testing of our tetravalent VHF vaccine formulation in phase 1 clinical trials. The next step is to identify and advance a vaccine that is effective but safe for pregnant women and immunocompromised persons.

Notes

Acknowledgments. The authors thank Jennifer Wilson for carefully editing the manuscript.

Financial support. This work was supported by the National Institute of Allergy and Infectious Diseases, National Institutes of Health (contract number HHSN272201700082C).

Potential conflicts of interest. M. J. S. is an inventor of the patent “Multivalent Vaccines for Rabies Virus and Filoviruses” (International Publication Number WO 2012/106490 A1). All other authors report no potential conflicts of interest.

The authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

Author notes

G. S. and C. Y. contributed equally to this work.