Superior in vitro stimulation of human CD8+ T-cells by whole virus versus split virus influenza vaccines

Abstract

Pandemic and seasonal influenza viruses cause considerable morbidity and mortality in the general human population. Protection from severe disease may result from vaccines that activate antigen-presenting DC for effective stimulation of influenza-specific memory T cells. Special attention is paid to vaccine-induced CD8+ T-cell responses, because they are mainly directed against conserved internal influenza proteins thereby presumably mediating cross-protection against circulating seasonal as well as emerging pandemic virus strains. Our study showed that influenza whole virus vaccines of major seasonal A and B strains activated DC more efficiently than those of pandemic swine-origin H1N1 and pandemic-like avian H5N1 strains. In contrast, influenza split virus vaccines had a low ability to activate DC, regardless which strain was investigated. We also observed that whole virus vaccines stimulated virus-specific CD8+ memory T cells much stronger compared to split virus counterparts, whereas both vaccine formats activated CD4+ Th cell responses similarly. Moreover, our data showed that whole virus vaccine material is delivered into the cytosolic pathway of DC for effective activation of virus-specific CD8+ T cells. We conclude that vaccines against seasonal and pandemic (-like) influenza strains that aim to stimulate cross-reacting CD8+ T cells should include whole virus rather than split virus formulations.

Figure 1. Vaccine-induced DC maturation in healthy donor HD15.

Immature DC of donor HD15 were incubated for 48(-like) and seasonal influenza virus vaccines at 10 µg/mL (referring to HA content) and were subsequently analyzed by flow cytometry for expression of maturation markers on viable 7AAD-negative cells (grey histograms). Split virus formulations were unavailable from pandemic (-like) strains A/H1N1-California and A/H5N1-Indonesia. Unfilled histograms represent IgG isotype control stainings. MFI values were added to each histogram. For abbreviations of virus strains see Table 1. s, split virus; w, whole virus; w/o, without.

Figure 2. Split virus preparation substantially decreases the ability to activate and mature DC.

(A) Immature DC were incubated for 48 h with different pandemic (-like) and seasonal influenza virus vaccines at 10 µg/mL (referring to HA content) and were subsequently analyzed by flow cytometry for expression of maturation markers on viable 7AAD-negative cells. Representative data of a single donor are shown in Fig. 1. Relative fluorescence intensity was calculated for each individual marker from MFI value of marker-specific staining divided by MFI value of the related IgG isotype control staining and was measured in 6 randomly selected donors. Summarized data are shown here as box blot diagrams. Split virus formulations were unavailable from pandemic (-like) strains A/H1N1-California and A/H5N1-Indonesia. P-values comparing the maturation effect of whole virus vaccines and related split virus vaccines on DC were calculated using two-tailed paired Wilcoxon signed-rank test (*, p<0.10; **, p<0.05). (B) After incubation of immature DC for 48 h with different pandemic (-like) and seasonal influenza virus vaccines, culture supernatants of four randomly selected donors were measured for IFN-α by ELISA, as well as for IL-6 and TNF-α by cytometric bead array. Graphs show mean concentrations (± SD) derived from experiments in all four healthy donors. For abbreviations of vaccines see Table 1. w/o, without.

Figure 3. CD8⁺ T-cell reactivity to whole virus is superior compared to split virus preparations.

CD4⁺ and CD8⁺ T cells purified from PBMC of healthy individuals were screened for IFN-γ ELISpot reactivity to autologous DC pre-loaded with 10 µg/mL of influenza whole virus and related split virus vaccine formulations. DC also received maturation cytokines during vaccine pulsing. (A) Representative data obtained from donor HD20 with 1×10⁵ CD4⁺ T cells (grey columns) or 1×10⁵ CD8⁺ T cells (black columns) plated per well are shown. (B, C) Reactivity to influenza whole virus and related split virus vaccines were measured in 10 randomly selected healthy individuals as described in (A). Box plot diagrams include IFN-γ ELISpot data from purified CD4⁺ (B) and CD8⁺ (C) T cells. Effective SFC were determined by subtraction of background spot numbers (w/o vaccine) from spot numbers induced by each individual vaccine. P-values were calculated by two-tailed paired-sample Wilcoxon signed-rank test.

Figure 4. Whole virus vaccine material can be detected outside the endolysosomal compartment.

Confocal LSM analysis was performed on immature DC after 4/H3N2-Uruguay whole virus vaccine. DC were pre-treated with the irreversible proteasome inhibitor epoxomicin for 4 h before vaccine was added. (A) and (B) show two different sectional planes of the same DC that was selected as a representative example of the entire sample. A.1/B.1 show LSM pictures, including nucleus (Hoechst 33342) marked in blue color, endolysosomal compartment (Lamp1/Rab5) in red color, and nucleoprotein (NP) of H3N2 in green color, respectively. (A.2/B.2) Graphs represent the fluorescence intensity of staining dyes used in A.1/B.1. White arrows in A.1/B.1 mark the measuring points from basis to arrowhead. (A.3/B.3) To further analyze the localization of NP within the DC, A.1/B.1 pictures were post-processed. Blue color means Hoechst positive only, grey color means Lamp1/Rab5 positive only, orange color means Lamp1/Rab5 as well as NP positive, green color means NP positive only. Thresholds are indicated in A.4/B.4. (A.4/B.4) Diagrams show the intensities of red (Lamp1/Rab5) and green (NP) fluorescence within white rectangles of A.3/B.3. To quantify colocalization of NP and Lamp1/Rab5 within the white rectangle of A.3/B.3 Mander's overlap coefficient (MOC) was calculated. Value range 0–1 (0: no colocalization, 1: all pixels colocalize).

Figure 5. Endolysosomal escape of whole virus vaccine material.

Immature DC were incubated for 2(w/o) or with whole virus vaccines of pandemic-like avian strain A/H5N1-Vietnam (VNw) or seasonal strain B/Brisbane (BBw). Subsequently, samples were processed and analyzed by TEM using a Zeiss 912 Omega microscope operated at 120 kV. Pictures were taken at 5 increasing magnifications to allow better detection and tracking of virus material.

Figure 6. Proteasome inhibition strongly reduces CD8⁺ T-cell reactivity to whole virus vaccine.

CD8⁺ T cells purified from PBMC of healthy donor HD21 were screened for IFN-γ ELISpot reactivity to autologous immature DC loaded with seasonal A/H3N2-Uruguay whole virus or split virus vaccines, or an influenza A/H3N2 nucleoprotein oligopeptide mix. Targets also included DC infected with live influenza virus of the Puerto Rico 8 lab strain. During loading or infection, respectively, DC were treated with epoxomicin (black columns) or DMSO solvent (grey columns), which both did not reduce cell viability in trypan blue staining. Data are representative of 3 experiments performed in different donors.