Differential immunogenicity between HAdV-5 and chimpanzee adenovirus vector ChAdOx1 is independent of fiber and penton RGD loop sequences in mice

Abstract

Replication defective adenoviruses are promising vectors for the delivery of vaccine antigens. However, the potential of a vector to elicit transgene-specific adaptive immune responses is largely dependent on the viral serotype used. HAdV-5 (Human adenovirus C) vectors are more immunogenic than chimpanzee adenovirus vectors from species Human adenovirus E (ChAdOx1 and AdC68) in mice, though the mechanisms responsible for these differences in immunogenicity remain poorly understood. In this study, superior immunogenicity was associated with markedly higher levels of transgene expression in vivo, particularly within draining lymph nodes. To investigate the viral factors contributing to these phenotypes, we generated recombinant ChAdOx1 vectors by exchanging components of the viral capsid reported to be principally involved in cell entry with the corresponding sequences from HAdV-5. Remarkably, pseudotyping with the HAdV-5 fiber and/or penton RGD loop had little to no effect on in vivo transgene expression or transgene-specific adaptive immune responses despite considerable species-specific sequence heterogeneity in these components. Our results suggest that mechanisms governing vector transduction after intramuscular administration in mice may be different from those described in vitro.

Figure 1. HAdV-5 vaccine vectors elicit higher CD8⁺ T cell and antibody responses, and deliver superior transgene expression in vivo compared to ChAdOx1 vectors.

BALB/c mice (5/group) were vaccinated intramuscularly with 10⁸ infectious units (ifu) HAdV-5 or ChAdOx1 expressing Photinus luciferase. Two weeks post vaccination, (A) splenic CD8⁺ T cell responses against dominant Luc_160-168 epitope GFQSMYTFV were measured by IFN-γ ELISpot and (B) anti-luciferase IgG titers were assessed in serum by endpoint ELISA. Statistical analyses performed by student T-test. (C) Prior to termination of the experiment, luciferase expression in vivo was measured in the same animals by bioluminescence imaging on day, 1, 3, 7 and 14, post vaccination. Graph shows mean total flux units per second (p/s) ± range. (D) Overlay image indicating the location of the bioluminescence signal 24 h post vaccination in both HAdV-5 and ChAdOx1 vaccinated groups. Intensity of signal indicated, where red is most intense (10⁸ p/s) and violet is least intense (10⁷ p/s).

Figure 2. Transgene expression is higher within draining popliteal lymph nodes after HAdV-5 vaccination than ChAdOx1 vaccination. BALB/c mice, (3/group) were vaccinated intramuscularly with 10⁸ ifu HAdV-5 or ChAdOx1 vectors expressing Photinus luciferase.

24 h later, mice were sacrificed, and popliteal lymph nodes and muscle tissue removed for extraction of luciferase protein. Luciferase activity was assayed in vitro from (A) muscle and (B) lymph node extracts. Bars show mean and SEM. Statistical analysis by two-tail student t-test between HAdV-5 and ChAdOx1 vaccinated groups. Results are representative of three independent experiments.

Figure 3. HAdV-5 vectors transduce lymph node resident antigen presenting cells more efficiently in vivo than ChAdOx1.

BALB/c mice (3/group) were vaccinated intramuscularly with 10⁹ ifu of HAdV-5 eGFP or ChAdOx1 eGFP. A further two groups vaccinated with HAdV-5 and ChAdOx1 vectors expressing non-fluorescent Ag85A were included as a negative control. At 24 h post vaccination, proximal popliteal draining lymph nodes (dLN) were harvested and expression of GFP was assessed within live cells by flow cytometry. (A) Percentage of all live cells within the dLN expressing eGFP. Co-expression of surface dendritic cell (DC) marker CD11c is also plotted on the y-axis. Data shown from one representative individual in each group, and result is representative of three independent experiments. (B) Percentage of GFP⁺ cells (from A) expressing MHC Class II. (C) Percentage of GFP⁺ MHC II⁺ cells (from B) with the phenotype CD11c⁺B220⁻ (cDCs), CD11c⁺B220⁺ (pDCs) and CD11c⁻B220⁺ (B cells). (D) Total live cells from the same experiment were phenotyped and %GFP expression assessed within each subset. Graphs in B–D show mean and SEM of data from each individual per group.

Figure 4. Replacement of fiber and penton RGD sequences from ChAdOx1 with corresponding sequences from HAdV-5 modifies transduction efficiency in vitro.

(A) A549 cells were incubated with HAdV-5, ChAdOx1 or ChAdOx1 chimeric vectors with the HAdV-5 fiber (Ad5F) or HAdV-5 penton RGD sequences (Ad5RGD), or both HAdV-5 fiber and penton RGD sequences (Ad5F+RGD) all expressing TIPeGFP. Vectors were added at several different multiplicities of infection (MOI) as shown. 24 h post infection, the percentage of cells expressing GFP was measured by flow cytometry. (B) To investigate the efficiency of virus entry via surface receptor CD46, chinese hamster ovary (CHO) cells stably expressing human CD46, (CHO-BC1) and CHO cells lacking CD46 expression (CHO-K1) were incubated with HAdV-5 or ChAdOx1 vectors expressing TIPeGFP for 1 h at 4 °C. Virus was then removed and cells washed before incubation for a further 24 h and assessment of GFP expression by flow cytometry. (C) To investigate the efficiency of virus entry via the coxsackie and adenovirus receptor (CAR), CHO cells stably expressing human CAR (CHO-CAR) and CHO-K1 cells were incubated with HAdV-5, ChAdOx1 and ChAdOx1 chimeric vectors expressing TIPeGFP. Virus was incubated with cells at 4 °C, removed, and GFP expression measured exactly as described in (B). (D) To further investigate the contribution of CAR mediated entry during transduction of CHO-CAR and A549 cells, HAdV-5 or ChAdOx1 vectors expressing GFP were incubated with cells in the presence or absence of recombinant HAdV-5 fiber knob protein (Knob5) previously shown to inhibit CAR mediated viral entry Virus was removed, cells washed, and transduction measured as described in (B,C). All graphs show mean and SEM of triplicate wells.

Figure 5. Replacement of fiber and penton RGD sequences from ChAdOx1 with corresponding sequences from HAdV-5 fails to improve magnitudes of T cell and antibody responses.

BALB/c mice (5/group) were vaccinated intramuscularly with 10⁸ infectious units (ifu) HAdV-5, ChAdOx1, or ChAdOx1 chimeric vectors expressing Photinus luciferase. (A) Luciferase expression in vivo was measured by bioluminescence imaging 6 h, 24 h, 4 days, 8 days, and 14 days post vaccination. Graph shows mean total flux units per second (p/s) ± range. Statistical analysis performed by repeating measures two-way ANOVA with Bonferroni post test. Dagger (†) indicates statistical significance between HAdV-5 and ChAdOx1, ChAdOx1 Ad5F and ChAdOx1 Ad5RGD; hash (#) indicates significance between HAdV-5 and ChAdOx1 Ad5F+RGD; and asterisk (*) indicates significance between ChAdOx1 and ChAdOx1 Ad5F+RGD. (B,C) After 14 days post vaccination, (B) splenic CD8⁺ T cell responses against dominant Luc_160–168 epitope GFQSMYTFV were measured by IFN-γ ELISpot and (C) anti-luciferase IgG titers were assessed in serum by endpoint ELISA. Statistical analyses in (B,C) performed by one-way ANOVA.

Figure 6. Transgene-specific adaptive immune responses are comparable between ChAdOx1 and fiber and/or penton RGD chimeric vectors expressing the TIPeGFP antigen construct.

BALB/c mice were vaccinated intramuscularly with 10⁸ ifu HAdV-5, ChAdOx1 or ChAdOx1 chimeric vectors expressing TIPeGFP. After 14 days post vaccination, CD8⁺ T cell responses against immunodominant epitope Pb9 (in the TIP epitope string) were measured by ICS. (A) Pb9 specific IFN-γ⁺CD8⁺ T cells as a percentage of total CD8⁺ T cells. (B) Percentages of Pb9 specific CD8⁺ T cells secreting combinations of the three cytokines IFN-γ, IL-2 and TNF-α. (C) Anti-GFP IgG titers in serum measured by endpoint ELISA. Dashed lines in (C) indicate limit of detection of the assay. Statistical analyses in (A,C) performed by one-way ANOVA.

Figure 7. Distribution of transgene expression within lymph node and muscle tissue differs after vaccination with HAdV-5 and ChAdOx1.

BALB/c mice were vaccinated intramuscularly with 10⁹ ifu HAdV-5 TIPeGFP, ChAdOx1 TIPeGFP, ChAdOx1 Ad5F+RGD TIPeGFP, or PBS alone. After 24 h, popliteal dLNs and samples of muscle tissue from the site of injection were collected and mounted in OCT compound for cryo-sectioning and immunohistochemistry. LN tissue was stained for localization of GFP (green) and CD11c (red) expression. Co-localisation of GFP and CD11c is shown in yellow. Muscle tissue was stained with DAPI (blue) and for localization of GFP (green) expression. Images (all at magnification 5x) are from one representative individual from each group (n = 3).