MVA-based SARS-CoV-2 vaccine candidates encoding different spike protein conformations induce distinct early transcriptional responses which may impact subsequent adaptive immunity

Abstract

Introduction: Vaccine platforms such as viral vectors and mRNA can accelerate vaccine development in response to newly emerging pathogens, as demonstrated during the COVID-19 pandemic. However, the differential effects of platform and antigen insert on vaccine immunogenicity remain incompletely understood. Innate immune responses induced by viral vector vaccines are suggested to have an adjuvant effect for subsequent adaptive immunity. Integrating data on both innate and adaptive immunity, systems vaccinology approaches can improve the understanding of vaccine-induced immune mechanisms.

Methods: Two vaccine candidates against SARS-CoV-2, both based on the viral vector Modified Vaccinia virus Ankara (MVA) and encoding the native (MVA-SARS-2-S) or prefusion-stabilized spike protein (MVA-SARS-2-ST), were evaluated in phase 1 clinical trials (ClinicalTrials.gov: NCT04569383, NCT04895449). Longitudinal dynamics of innate and early adaptive immune responses induced by vaccination in SARS-CoV-2-naïve individuals were analyzed based on transcriptome and flow cytometry data, in comparison to the licensed ChAd and mRNA vaccines.

Results: Compared to MVA-SARS-2-S, MVA-SARS-2-ST (encoding the prefusion-stabilized spike protein) induced a stronger transcriptional activation early after vaccination, as well as higher virus neutralizing antibodies. Positive correlations were observed between innate and adaptive immune responses induced by a second MVA-SARS-2-ST vaccination. MVA-, ChAd- and mRNA-based vaccines induced distinct immune signatures, with the overall strongest transcriptional activation as well as monocyte and circulating T follicular helper (cTFH) cell responses induced by ChAd.

Discussion: Our findings suggest a potential impact of the spike protein conformation not only on adaptive but also on innate immune responses. As indicated by positive correlations between several immune parameters induced by MVA-SARS-2-ST, the distinct transcriptional activation early after vaccination may be linked to the induction of classical monocytes and activation of cTFH1 cells, which may in turn result in the superior adaptive immunogenicity of MVA-SARS-2-ST, compared to MVA-SARS-2-S. Overall, our data demonstrate that both the vaccine platform and antigen insert can affect innate immune responses and subsequent vaccine immunogenicity in humans.

Keywords: COVID-19; SARS-CoV-2; T follicular helper cells; innate immunity; modified vaccinia virus Ankara; spike protein; systems vaccinology; transcriptome.

Figure 1

Study overview. In four study cohorts, participants received two vaccinations with different SARS-CoV-2 vaccines. Participants of the first two cohorts received the recombinant MVA-based vaccine candidates MVA-S (MVA-SARS-2-S, yellow) and MVA-ST (MVA-SARS-2-ST, blue) in the scope of two phase 1 clinical trials. The control cohorts received different combinations of licensed vaccines, used as benchmarks for vaccine immunogenicity. One control cohort received two vaccinations with the licensed mRNA vaccine BNT162b2 (light grey). Participants of the fourth cohort (dark grey) received a dose of the licensed ChAd vaccine (ChAdOx1 nCov-19), followed by a dose of a licensed mRNA vaccine (BNT162b2 or mRNA-1273). The different vaccines encode either the native spike protein (yellow) or the prefusion-stabilized spike protein (blue). Blood samples were collected prior to vaccination (D0) and on D1, D3, D7 and D14 after the first and the second vaccination (V1 and V2). Transcriptional responses as well as innate and spike-specific circulating T follicular helper (cTFH) cell responses were analyzed and evaluated for correlations with adaptive immune outcomes (spike-specific neutralizing and binding antibodies, B and T cell responses).

Figure 2

MVA-S and MVA-ST vaccination induce differential transcriptional responses. (A) Number of upregulated or downregulated differentially expressed genes (DEGs) on D1, D3 and D7 after each vaccination with MVA-S, MVA-ST, mRNA or ChAd/mRNA, defined as genes with a false discovery rate (FDR) <0.1 and an absolute log fold change ≥1 compared to V1D0. (B) Canonical pathways involved in the immune system, which are up- or downregulated with -log (FDR) >5 and absolute z-score >2 for at least one time point in one cohort, as analyzed by Ingenuity Pathway Analysis (IPA, Qiagen). The number of DEGs in each pathway is shown in Supplementary Tables 31 - 34 . (C) Circos plots connecting canonical pathways, which are significantly up- (green) or downregulated (brown) at more than one time point or in more than one group. (D) Ingenuity upstream regulator cytokine analysis using DEGs with FDR <0.1 and an absolute log fold change ≥1 compared to V1D0. The heatmap shows the top upstream cytokines with a -log (FDR) >5 and absolute z-score >2 for at least one time point in one cohort. (E) Venn diagrams showing shared DEGs on V1D1 and V2D1 following MVA-S and MVA-ST vaccination. All DEGs following MVA-S and MVA-ST vaccination are listed in Supplementary Tables 35 , 36 , respectively. (F) Longitudinal expression of selected genes involved in canonical signaling pathways, which are differentially up- or downregulated on at least one time point following MVA-S or MVA-ST vaccination. Depicted are the normalized counts calculated by variance stabilizing transformation (VST). Boxplots depict median and interquartile range, dots resemble individual data points. Longitudinal expression levels of all DEGs of the MVA-S/ST cohorts, according to the canonical signaling pathways shown in (C), are shown in Supplementary Figures 3 - 18 .

Figure 3

ChAd- and mRNA-based vaccines induce a strong response of classical monocytes. (A) Gating schemes used for differentiation of monocyte subsets within HLA-DR+CD11b+ cells, and dendritic cell (DC) subsets within HLA-DR+CD11b- cells, respectively. The complete gating strategy for identification of monocytes and DCs within whole PBMCs, is shown in Supplementary Figure 1 . (B) Longitudinal dynamics of the frequency and activation (CD40 expression) of monocyte and DC subsets at baseline (D0) and on D1, D3 and D7 following each vaccination. Displayed is the median frequency of each subpopulation, and the median fold change of the CD40 expression compared to V1D0. MFI, median fluorescence intensity. Descriptive statistics for all parameters are reported in Supplementary Tables 8 - 21 .

Figure 4

ChAd-based primary vaccination induces a stronger cTFH cell response compared to mRNA- and MVA-based vaccination. (A) Exemplary contour plots used for identification of cTFH cell subsets (cTFH1, cTFH2, cTFH17) and their activation status (based on CD38 and ICOS expression). The complete gating strategy for identification of cTFH cells within PBMCs is shown in Supplementary Figure 2 . (B) Longitudinal dynamics of cTFH subpopulation frequencies, as well as ICOS+ and CD38+ cTFH cells, represented as median values of the subpopulations cTFH1, cTFH2 and cTFH17 on D0, D7 and D14 after V1 and V2. (C) Fold induction of cTFH1 cell frequency (upper panel), ICOS+ (middle panel) and CD38+ cTFH1 cells (lower panel) on V1D7 and V2D7 compared to baseline. Descriptive statistics for all parameters are reported in Supplementary Tables 22 - 30 .

Figure 5

Increased immunogenicity of MVA-ST correlates with transcriptional, monocyte and cTFH cell responses. (A) Neutralizing antibodies (nAB), as measured by virus neutralization test (VNT₁₀₀) against wild type SARS-CoV- 2 at baseline and 14 days after second vaccination with MVA-S and MVA-ST (B) Spearman correlation analysis between early immune responses following first (left panel) and second vaccination (right panel) with MVA-ST, and adaptive immune outcomes after completion of the vaccination schedule. The early immune parameters include the peak time point of previously selected DEGs, as well as the frequency and activation of classical monocytes (cl MO) and circulating T follicular helper 1 (cTFH1) cells. The adaptive immune parameters include nAB, S1- and S2-specific IgG, B cell and T cell responses, measuered at V2D14. All parameters were normalized to their baseline levels (V1D0), using fold changes for early immune responses and baseline subtraction for adaptive immune parameters. Circle size and color indicate Spearman correlation coefficient (r).