The whole-cell pertussis vaccine imposes a broad effector B cell response in mouse heterologous prime-boost settings

Abstract

ÍSince the introduction of new generation pertussis vaccines, resurgence of pertussis has been observed in many developed countries. Former whole-cell pertussis (wP) vaccines are able to protect against disease and transmission but have been replaced in several industrialized countries because of their reactogenicity and adverse effects. Current acellular pertussis (aP) vaccines, made of purified proteins of Bordetella pertussis, are efficient at preventing disease but fail to induce long-term protection from infection. While the systemic and mucosal T cell immunity induced by the 2 types of vaccines has been well described, much less is known concerning B cell responses. Taking advantage of an inducible activation-induced cytidine deaminase fate-mapping mouse model, we compared effector and memory B cells induced by the 2 classes of vaccines and showed that a stronger and broader memory B cell and plasma cell response was achieved by a wP prime. We also observed that homologous or heterologous vaccine combinations that include at least 1 wP administration, even as a booster dose, were sufficient to induce this broad effector response, thus highlighting its dominant imprint on the B cell profile. Finally, we describe the settlement of memory B cell populations in the lung following subcutaneous wP prime vaccination.

Keywords: Adaptive immunity; Immunoglobulins; Immunology; Memory; Vaccines.

Figure 1. A broader Ab response is induced by the wP prime vaccination.

(A) Blood was collected at day 0, day 14, and day 30 from AID-Cre-EYFP mice s.c. injected with aP or wP vaccines or with alum (ctr, control mice). IgM, IgG1, and IgG2b serum Ab titers against pooled proteins of the aP vaccine (PT, PRN, FHA, Fim2,3) (B) and IgM, IgG1, IgG2b, IgG2c, and IgG3 serum Ab titers against sonicated Bp (C) were detected by ELISA from serum of vaccinated and control mice. Ab titers are arbitrary values and each point in the graphs represents individual mouse data. At least 2 independent experiments were performed for each analysis. Means (±SEM) are shown. Kruskal-Wallis analysis with uncorrected Dunn’s test was performed to compare the different conditions at each time point and the different time points between the same conditions. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 2. The wP prime leads to stronger GC and memory B cell responses in dLNs.

(A) AID-Cre-EYFP mice primed either with aP or wP vaccines or controls injected with alum (ctr) received 2 doses of tamoxifen at days 7 and 10 after prime vaccination. Mice were analyzed at days 14 and 30. (B) B220⁺EYFP⁺ live cells from dLNs of mice primed with aP and wP vaccines were distinguished into GC (GL7⁺) and memory (GL7^–) B cells, by flow cytometry. (C) Total EYFP⁺ GC and EYFP⁺ memory B cell counts in the 2 dLNs are shown in the graphs. (D) A representative flow cytometry profile of heavy chain isotype distribution among EYFP⁺GL7⁺ B cells is shown for the aP and wP conditions at day 30 after prime. (E) IgM/IgD, IgG1, IgG2, and IgA distribution in the EYFP⁺ GC and memory subsets from mice analyzed at day 30 after prime are shown in the plots. Each point represents an individual mouse analyzed at day 14 or day 30 after prime vaccination. At least 2 independent experiments were performed for each analysis. Means (±SEM) are shown. Kruskal-Wallis analysis with uncorrected Dunn’s test was performed to compare the different conditions at each time point or each Ab isotype. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 3. The wP prime induces a broader PC population in BM.

AID-Cre-EYFP mice primed with aP or wP vaccines or controls injected with alum (ctr) received 2 doses of tamoxifen at days 7 and 10 after prime vaccination. BM cells were collected 30 days after prime vaccination. (A) Representative flow cytometry profile of intracellular staining for EYFP⁺ PCs in BM. B220^–EYFP⁺ live cells were gated into CD138⁺ cells before determination of IgM, IgG1, and IgG2 distribution among PCs (a similar independent analysis was performed for IgA at the place of IgG2, not shown here). Total EYFP⁺ PC cell numbers (B) or EYFP⁺ PC isotype distribution (C) are represented in the plots. Numbers of IgG1⁺ ASCs against pooled proteins (PT, PRN, FHA, Fim2,3) (D) and numbers of IgG1⁺ and IgG2⁺ ASCs against sonicated Bp (E) were determined by ELISPOT from total BM cells. Representative spot images for each condition are shown at the left of each panel. Each point in the graphs represents an individual mouse. At least 2 independent experiments were performed for each analysis. Means (±SEM) are shown. Kruskal-Wallis analysis with uncorrected Dunn’s test was performed to compare the different conditions. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 4. wP:wP prime:boost vaccination is highly efficient at producing secondary B cell responses, including GCs, PCs, and systemic memory B cells, shortly after the boost.

(A) AID-Cre-EYFP mice were primed and boosted (day 30) with homologous and heterologous combinations of the aP and wP vaccines or injected with alum (ctr). Three doses of tamoxifen were administrated at days 7, 10, and 31. Mice were analyzed at day 35. Total cell numbers and isotype distribution of EYFP⁺ GC (B) or EYFP⁺ memory (C) B cells were determined by flow cytometry in dLNs. (E) Cell numbers (relative to 10⁵ splenocytes) and isotype distribution of EYFP⁺ memory B cells were determined by flow cytometry in the spleen. EYFP⁺ PCs were assessed by flow cytometry in dLNs (D) or spleen (F), and cell counts were reported similarly in the graphs. Each point in the plots represents an individual mouse. At least 2 independent experiments were performed for each analysis. Means (±SEM) are shown. Kruskal-Wallis analysis with uncorrected Dunn’s test was performed to compare the different conditions. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 5. Persistence of GC B cells and local and systemic memory B cells is favored by the wP vaccine.

(A) AID-Cre-EYFP mice were primed and boosted (day 30) with homologous and heterologous combinations of the aP and wP vaccines or injected with alum (ctr). Three doses of tamoxifen were administrated at days 7, 10, and 31. Mice were analyzed 50 days after boost injection. Cell numbers of EYFP⁺ GCs (B) and cell numbers and isotype distribution of total GC B cells (C) or EYFP⁺ memory B cells (D) were assessed by flow cytometry in dLNs and reported in the graphs. (E) Cell numbers and isotype distribution of EYFP⁺ memory B cells were assessed by flow cytometry in spleen. Numbers of IgG1⁺ ASCs against pooled proteins (PT, PRN, FHA, Fim2,3) (F) or numbers of IgG1⁺ and IgG2⁺ ASCs against sonicated Bp (G) were determined by a memory B cell ELISPOT assay performed 5 days after in vitro activation of splenocytes in the presence of IL-2 and R848. Each point in the graphs represents an individual mouse. At least 2 independent experiments were performed for each analysis. Means (±SEM) are shown. Kruskal-Wallis analysis with uncorrected Dunn’s test was performed to compare the different conditions. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 6. Three different signatures of memory B cells were identified by single-cell RNA-Seq analysis in vaccinated mice.

(A) The expression of CD73, CD80, and PD-L2 was determined on dLN EYFP⁺GL7^– memory B cells by flow cytometry 50 days after boost from AID-Cre-EYFP mice that were primed and boosted (day 30) with homologous and heterologous combinations of the aP and wP vaccines or injected with alum (ctr). Three doses of tamoxifen were administrated at days 7, 10, and 31. Representative flow cytometry plots are shown for aP:aP and wP:wP conditions. CD73⁺CD80⁺ and PD-L2⁺CD80⁺ memory EYFP⁺ populations in dLNs are shown in the graphs for all vaccine combinations. (B) Uniform manifold approximation and projection (UMAP) and clustering of 707 EYFP⁺ GC and 1,962 EYFP⁺ memory sorted B cells analyzed by scRNA-Seq from dLNs of AID-Cre-EYFP mice 50 days after boost injection, and tamoxifen administration on days 7, 10, and 31, from 2 aP:aP-, 3 aP:wP-, 3 wP:aP-, and 2 wP:wP-immunized mice. (C) Selected gene expression for the 6 different clusters (numbered 0–5) is presented in dot plot scale on normalized unique molecular identifier (UMI) counts. Frequencies (D) and absolute numbers (E) of EYFP⁺ memory B cells within clusters 0, 3, and 4 are shown for the 4 vaccine combinations. (F) Isotype profile determined by flow cytometry during cell sorting is shown for the EYFP⁺ memory B cells from clusters 0, 3, and 4. Means (±SEM) are shown in panels A, D, and E.

Figure 7. Isotype-switched resident-like memory B cells localize in the lungs of wP-primed mice.

(A) AID-Cre-Tomato mice, primed and boosted (day 30) with homologous and heterologous aP and wP vaccine combinations, received tamoxifen at days 7, 10, and 31. Analysis was performed 50 days after boost. (B) Representative flow cytometry analysis of d-Tomato⁺ cells selected from B220⁺CD45.2^– resident lung B cells. d-Tomato⁺GL7^– cells were further analyzed for the expression of IgM and IgG1 membrane Ab isotypes. (C) Numbers of total d-Tomato⁺ resident memory cells (graph on the left) and their isotype subclasses (graph on the right) are shown. (D) dLN, splenic, and lung total d-Tomato⁺ memory B cells belonging to the wP:wP condition were analyzed for the expression of CD73, CD80, and PD-L2 markers. A representative flow cytometry plot is shown. Geometric MFI relative to each membrane marker is indicated in the graphs for all organs. Each point in the graphs (C and D) depicts an individual mouse. At least 2 independent experiments were performed for each analysis. Means (±SEM) are shown. Kruskal-Wallis analysis with uncorrected Dunn’s test was performed to compare the different conditions. *P < 0.05, **P < 0.01, ****P < 0.0001.

Figure 8. Formation of large numbers of Bp-specific long-lived PCs in the BM requires 1 wP vaccination.

AID-Cre-EYFP mice were primed and boosted (day 30) with homologous and heterologous combinations of the aP and wP vaccines or injected with alum (ctr). Three doses of tamoxifen were administrated at days 7, 10, and 31. Mice were analyzed 50 days after boost injection. (A) Cell numbers and isotype distribution of EYFP⁺ PCs assessed by flow cytometry in BM are shown in the graphs. (B) Numbers of IgG1⁺ or IgG2⁺ ASCs against sonicated Bp were determined by ELISPOT performed on total BM cells. Each point in the graphs represents an individual mouse. At least 2 independent experiments were performed for each analysis. Means (±SEM) are shown. Kruskal-Wallis analysis with uncorrected Dunn’s test was performed to compare the different conditions. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 9. Production of Bp-specific Abs with broader antigen and isotypic diversity in recall responses is favored by combining wP and aP vaccines.

AID-Cre-EYFP mice were primed and boosted (day 30) with homologous and heterologous combinations of the aP and wP vaccines or injected with alum (ctr). Blood was collected at 5, 25, 40, and 50 days after boost injection. IgG1 and IgG2b Ab titers against pooled proteins (PT, PRN, FHA, Fim2,3) (A) or IgG1, IgG2b, IgG2c, and IgG3 Ab titers against sonicated Bp (B) were detected by ELISA from serum of vaccinated and control mice. Ab titers are arbitrary values and each point in the graphs represents data from an individual mouse. At least 2 independent experiments were performed for each analysis. Means (±SEM) are shown. Kruskal-Wallis analysis with uncorrected Dunn’s test was performed to compare the different conditions at each time point and the different time points between the same condition. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.