Fig 1.
The FILOcep1&2 vaccine design.
Conserved regions of the filovirus proteome (red) are the most similar parts of proteins common across the eight virus species of the filovirus family. These regions were identified by amino acid alignment of all known filovirus isolates in the database. An algorithm called Epigraph computed bi-valent amino acid sequences (epigraph 1 and epigraph 2), which complement each other and are used together in a vaccine to optimize match of potential T-cell epitopes between the vaccine and all input filovirus species [11]. For the FILOcep1 and FILOcep2 epigraphs, the four regions 1, 2, 3 and 4 are 280 (nucleoprotein 131–410), 123 (matrix 71–193), 315 (RNA polymerase 540–854) and 109 (RNA polymerase 952–1060) amino acid long, respectively, and were arranged into different orders to minimize potential induction of T cells recognizing irrelevant (non-viral) newly generated epitopes across the regional junctions. Synthetic ORF coding for these two proteins each 827 amino acid in length were inserted into engineered replication-deficient simian (chimpanzee) adenovirus ChAdOx1 and replication-deficient poxvirus MVA to generate four components of the vaccine abbreviated C1, C2, M1 and M2.
Fig 2.
Dose finding for rMVA and rChAdOx1 vaccines.
Groups of the BALB/c mice were administered intramuscularly increasing doses of 1x106, 1x107, 1x108 and 5x108 infectious units (IU) of individual ChAdOx1.FILOcep1 and ChAdOx1.FIOcep2 vaccines and their half-dose combinations A), or 1x105, 1x106, 5x106 and 1x107 plaque-forming units (PFU) of MVA.FILOcep1 and MVA.FILOEcep2 vaccines and their combined half-doses B), and the frequencies of the vaccine-elicited filovirus-specific T cells in the spleen were assessed 9 days after vaccination in an IFN-γ ELISPOT assay using 12 pools of FILOcep1&2-derived 15-mer peptides overlapping by 11 amino acids and spanning the full length of both immunogens. T-cell epitope variant peptide pairs were used together in pools to allow addition of pool-detected frequencies for overall magnitude of the anti-FILOcep1&2 responses. Data are shown as median (range), n = 3. Kruskal-Wallis test was used to determine the significance of variation among individual doses/vaccinations for immunodominant peptide pools P3 and P12 and the P values are shown above.
Fig 3.
Optimization of the regimen and characterization of FILOcep1&2 vaccine-elicited responses in the BALB/c mice.
Groups of mice were vaccinated A) with either the two adenovirus or two poxvirus components alone and compared to their C1C2 prime-M1M2 boost combination, B) exploring various vaccine distributions between two anatomical sites (hind legs) and compared to C1C2-M1M2, C) administering all 4 vaccine components on the same day and assaying at 4 weeks or 1 week later compared to C1C2-M1M2. The induced T cells were analyzed employing 12 FILOcep1&2 peptide pools in an IFN-γ ELISPOT assay. For A), B) and C), Kruskal-Wallis test was used to determine the significance of variation among regimens for immunodominant peptide pools P3 and P12 and the P values are shown above. D) Using the most efficient C1C2-M1M2 regimen, all 390 15-mer peptides were tested individually in an IFN-γ ELIPSPOT assay (S1 Fig.) and the strongest peptide pairs from that scan are listed, whereby SFU/M gives the frequencies of responding splenocytes per million. E) The 4 most immunodominant 15-mer peptide pairs used to characterize the functionality of vaccine-elicited CD8+ and CD4+ T cells, whereby the plurifunctionality of cells expressing 1 (black), 2 (light gray) and 3 (dark gray) cytokines/functions simultaneously are given as pie charts. Data in A), B), C) and E) are shown as median (range), n = 4. F) Two immunodominant CD8+ T-cell epitopes in 15-mers 105 and 336 were narrowed to their optimal length using IFN-γ ELISPOT assay with the frequencies of responding T cells on the right. G) In a separate immunization experiment, optimal-length variant epitopes derived from the 8 members of the filovirus family were assessed for recognition by C1C2-M1M2-induced T cells.
Fig 4.
Complete protection of the BALB/c mice against Ebola and Marburg virus challenges by FILOcep1&2 vaccination.
BALB/c mice were immunized with the candidate FILOcep1&2 vaccines or control vaccines expressing irrelevant eGFP, challenged by Ebola and Marburg viruses as shown in Table 1. A) Induction of filovirus-specific T cells was confirmed using Mabtech (left) and Cellular Technology Limited (CTL; right) IFN-γ ELSIPOT assay kits using two immunodominant peptide pools P3 and P12 on day 28. Frequencies are shown as median (range), n = 4. B) Eight mice in the FILOcep1&2 (blue) and 8 in the control eGFP (red) groups were challenged with 1000 LD50 of either mouse-adapted EBOV (Mayinga; left) or 1000 LD50 mouse-adapted MARV (Angola; right) virus on day 35 and the animals were daily measured for their body mass till day 14 post challenge (top) and survival till day 28 post challenge (bottom). The P values for survival used the Log-rank (Mantel-Cox) Test, n = 8. C) The EBOV and MARV viruses were quantified in various tissues on 3 and 5 days after the challenge (DPC). Data are shown as median (range), n = 4 per group.
Table 1.
Design of the experimental challenges.
Fig 5.
Complete protection of the C57BL/6J mice against Ebola and Marburg virus challenges by FILOcep1&2 vaccination.
A) A group of the C57BL/6J mice was vaccinated using the C1C2 and M1M2 3 weeks apart and the pattern of immunodominance for the 12 FILOcep1&2 peptide pools was determined in an IFN-y ELISPOT assay in the Oxford laboratory 1 week later (n = 4). B) In the Winnipeg laboratory, groups of the C57BL/6J mice were immunized with the candidate FILOcep1&2 vaccines or control vaccines expressing irrelevant eGFP and challenged by Ebola and Marburg viruses on day 35 (Table 1). Kruskal-Wallis test was used to determine the significance of variation among regimens for immunodominant peptide pool P4 and the P value is shown above. B) Four mice were killed and the induction of filovirus-specific T cells was confirmed in an IFN-γ ELSIPOT assay kits using the 4 immunodominant peptide pools P3, P4, P5 and P7. C) Eight mice in the FILOcep1&2 (blue) and 8 in the control eGFP (red) groups were challenged with 1000 LD50 of either mouse-adapted EBOV (Mayinga; left) or 1000 LD50 mouse-adapted MARV (Angola; right) virus and animals’ body mass was measured daily till day 14 post challenge (top) and survival was monitored until day 28 post challenge (bottom). The P values for survival used the Log-rank (Mantel-Cox) Test, n = 8.
Fig 6.
The FILOcep1&2 vaccine administration did not induce any filovirus-specific antibodies.
Proteins in lysates of 8 μg of purified irradiated EBOV (A) or MARV (B) preparations were separated by SDS-PAGE and used to assess the induction of filovirus protein-specific antibodies induced by administration of the FILOcep1&2 vaccines in a Western blot. Sera from 1 week after the MVA.FILOcep1&2 boost of the BALB/c or C57BL/6J mice were combined for each group, diluted 1:1000 and tested against the Western blotted virus lysates or PBS as indicated. MAbs specific for the EBOV and MARV NPs served as positive controls, and no serum or sera from the same time isolated from mice vaccinated with the control eGPF vaccines were used as negative controls. Relative molecular mass markers are shown.
Table 2.
Well defined human CD8+ T-cell epitopes in FILOcep1&2*.
Table 3.
Primers and probes for determination of virus load.