A live RSV vaccine with engineered thermostability is immunogenic in cotton rats despite high attenuation

Abstract

Respiratory syncytial virus (RSV) is a leading cause of infant hospitalization and there remains no pediatric vaccine. RSV live-attenuated vaccines (LAVs) have a history of safe testing in infants; however, achieving an effective balance of attenuation and immunogenicity has proven challenging. Here we seek to engineer an RSV LAV with enhanced immunogenicity. Genetic mapping identifies strain line 19 fusion (F) protein residues that correlate with pre-fusion antigen maintenance by ELISA and thermal stability of infectivity in live RSV. We generate a LAV candidate named OE4 which expresses line 19F and is attenuated by codon-deoptimization of non-structural (NS1 and NS2) genes, deletion of the small hydrophobic (SH) gene, codon-deoptimization of the attachment (G) gene and ablation of the secreted form of G. OE4 (RSV-A2-dNS1-dNS2-ΔSH-dGm-Gsnull-line19F) exhibits elevated pre-fusion antigen levels, thermal stability, immunogenicity, and efficacy despite heavy attenuation in the upper and lower airways of cotton rats.

Conflict of interest statement

M.L.M. co-founded Meissa Vaccines, Inc. and serves as Chief Scientific Officer for the Company. M.L.M., C.C.S., A.L.H., J.M. and C.A.R. are co-inventors of RSV vaccine technology subject to evaluation in this paper. The vaccine technology has been optioned to Meissa by Emory University. The remaining authors declare no competing financial interests.

Figures

Figure 1. MPE8 and D25 ELISAs.

(a) Ratio of direct ELISA using MPE8, a pre-F-specific mAb, to direct ELISA using motavizumab, a total F mAb. Values are normalized to strain A2. For A2-line19F mutants, the asterisks show significant differences compared with A2-line19F. (b) Ratio of direct ELISA using D25, another pre-F-specific mAb, to direct ELISA using motavizumab. All graphs represent the means+s.d.'s of at least two experimental replicates, and data were analysed by one-way ANOVA. When significant, P values are shown as a bracket between groups (P<0.0005) or by asterisk when compared with A2-line19F (*P<0.05; **P<0.005; ***P<0.0005).

Figure 2. Thermal stability assays.

Thermal inactivation was carried out either by incubation of virus at 4 °C (a,c,d) or 37 °C (b). The viruses in c labelled 79, 191, 357, 371, 557 and 357/371 represent A2-line19F containing substitutions at these indicated positions with A2 residues. The virus in c labelled DB1-357/371 represents DB1 with substitutions of line 19F residues at positions 357 and 371. Viruses were harvested at the indicated time points and titrated by FFU or PFU assays. All graphs represent the means+s.d.'s of at least two experimental replicates combined, and data were analysed by two-way ANOVA (*P<0.05; **P<0.005; ***P<0.0005; ****P<0.00005).

Figure 3. Design of live-attenuated vaccine OE4 and expression of viral proteins.

(a) Schematic of RSV LAV OE4 genome including codon deoptimization of the NS1, NS2 and G genes, deletion of the SH gene, and incorporation of the line 19F gene. (b) Western blotting of Vero cells infected with A2 (white), OE4 (green) or OE4 expressing wild-type G (OE4-wtG, grey) for NS1, NS2, N and G. An A2-Gnull mutant was included as a control. (c) Western densitometry analyses were normalized to A2 expression levels. (d) Western blotting of Vero cells infected with mock, A2-line19F (white), OE4-wtG (grey) or OE4 (green) for F, N and GAPDH. (e) Densitometry results were normalized to A2 expression levels. Densitometry results represent the means+s.d.'s of at least two experimental replicates and representative blots are shown. Statistical analyses were performed by one-way ANOVA (***P<0.0005; ****P<0.00005). d, codon-deoptimized; F, fusion protein; G, attachment glycoprotein; L, large polymerase; M, matrix; mK2, monomeric Katushka2; N, nucleoprotein; NS1/NS2, nonstructural proteins 1 and 2; P, phosphoprotein; SH, small hydrophobic protein.

Figure 4. Immunogold labelling of RSV surface glycoproteins F and G.

(a) Representative TEM images of BEAS-2B cells infected at an MOI of 10 with A2 (black) or OE4 (green) and labelled with MPE8 (pre-F mAb), 131-2A (post-F mAb), Motavizumab (total F mAb) or 131-2G (G mAb) and probed with gold-labelled secondary antibodies. (b) Quantification of the amount of immunogold particles per measured membrane length per virion. For each labeling condition, more than 100 virions (graph data points) were evaluated for each virus. The red lines represent the mean particle densities along the membrane for each condition. Significant differences are indicated by ***P<0.0005 determined by t-test with Welch's correction. Scale bars represent 200 nm.

Figure 5. Cryo-electron tomography of RSV virions and subvolume averaging of the F glycoprotein.

(a–c) Tomographic slices (6.14 nm) of A2, OE4 and A2-heat (55 °C for 30 min) virions showing overall virus structure and the organization of surface glycoproteins (insets). Inset in OE4 is rotated 180°. Scale bars are 200 nm for A2 and OE4, and 100 nm for A2-heat. (d–l) Subvolume averages and modelling of RSV F structures in pre- and post-fusion conformations. Central slices (6.14 Å in thickness) of the averaged structures lowpass filtered to 40 Å for A2 (d), OE4 (e) and A2-heat (f). Quasi-atomic models generated by fitting the RSV pre-fusion F (PDB ID 4JHW) and RSV post-fusion F (PDB ID 3RRT) crystal structures into the subvolume averages, with side views (g–i) and top views (j–l) for A2 (g,j), OE4 (h,k) and A2-heat (i,l). Note the height difference between the ectodomain of A2/OE4 and A2-heat. The measurements were made from the top of the membrane to the top of the head domain. Scale bars, 10 nm (d–f); 5 nm (g–l).

Figure 6. OE4 replication in immortalized and primary cell cultures.

Vero (a), BEAS-2B (b), primary normal human bronchial epithelial cells differentiated at air–liquid interface (NHBE) (c), primary human tracheobronchial airway cells differentiated at air–liquid interface (HAE) (d) were infected with A2-line19F (black), OE4 (green) and in HAE, OE4+wtG (green dash) at MOI=0.01 (Vero and BEAS-2B), MOI=2.6 (NHBE), or MOI=6.7 (HAE). Samples were titrated by fluorescent focus unit (FFU) assays on Vero cells. (e) Representative images of infected HAE cultures. Scale bar represents 200 μm. Graphs depict the means±s.e.s of the means combined from three experiments (Vero and BEAS-2B), from two donors in duplicate (NHBE), or from six cultures from a single donor per virus (HAE). When significant, P values are shown relative to A2-line19F (*P<0.05; by two-way ANOVA).

Figure 7. Attenuation and efficacy of OE4 in mice and cotton rats.

(a) Lung viral loads were determined in mice inoculated i.n. with 106 FFU of A2, A2-line19F or OE4 at the indicated time points. (b) Serum nAb titres were measured in mice inoculated with 106 FFU of A2, A2-line19F or OE4. (c) Mice were inoculated with 106 FFU of A2, A2-line19F or OE4 then challenged with 105 PFU of A2-line19F on day 102. Lung viral loads were determined day 4 post challenge. (a–c) Graphs represent combined data from two experiments of 5–10 mice per group. (d) Mice (five per group) were inoculated with mock, A2-line19F, OE4 or A2-del-M2-2, and lungs were harvested 8 days post inoculation for histological quantification of airway mucin expression. Each dot represents an airway, and graph shows >300 airways per group in one of two experiments with similar results. (e,f) Viral load on day 4 in cotton rat lung homogenates (n=3) (e) and nasal washes (n=3) (f) following i.n. inoculation with 105 FFU of A2, OE4 or A2-del-M2-2. (g) Cotton rats (six per group) were inoculated with mock, RSV A(Tracy), OE4 or A2-del-M2-2, and serum nAb titres against representative RSV strains were determined on day 42 post inoculation using pooled sera. EC50 was calculated by non-linear regression, and data represent EC50+upper limit of the 95% confidence interval. (h,i) Cotton rats (five per group) were inoculated with mock, A2, or OE4, challenged on day 42 with 106 FFU of RSV A2-line19F, and viral loads on day 46 were measured in nasal washes (i) and lung lavages (h). *P<0.05; **P<0.005; ***P<0.0005; ****P<0.00005 by one-way (c–f,h,i) or two-way (a,b) ANOVA.

Figure 8. Histopathology following RSV challenge in cotton rats.

To evaluate for vaccine attributable enhanced disease post challenge, groups of five cotton rats were inoculated intramuscularly with FI-RSV or intranasally with either mock or OE4. Animals vaccinated with FI-RSV also received a boost on day 21 p.i. All animals were challenged with A2-line19F on day 42 p.i., lungs were harvested 4 days later and histopathology scores were performed (a). Representative haematoxylin and eosin stains for FI-RSV (b), mock (c) and OE4 (d) vaccinated rats are shown. Scale bars represent 200 μm. Data are represented as mean+s.d. *P<0.05, **P<0.005 by two-way ANOVA.