Restrained expansion of the recall germinal center response as biomarker of protection for influenza vaccination in mice

Abstract

Correlates of protection (CoP) are invaluable for iterative vaccine design studies, especially in pursuit of complex vaccines such as a universal influenza vaccine (UFV) where a single antigen is optimized to elicit broad protection against many viral antigenic variants. Since broadly protective antibodies against influenza virus often exhibit mutational evidence of prolonged diversification, we studied germinal center (GC) kinetics in hemagglutinin (HA) immunized mice. Here we report that as early as 4 days after secondary immunization, the expansion of HA-specific GC B cells inversely correlated to protection against influenza virus challenge, induced by the antigen. In contrast, follicular T helper (TFH) cells did not expand differently after boost vaccination, suggestive of a B-cell intrinsic difference in activation and differentiation inferred by protective antigen properties. Importantly, differences in antigen dose only affected GC B-cell frequencies after primary immunization. The absence of accompanying differences in total anti-HA or epitope-specific antibody levels induced by vaccines of different efficacy suggests that the GC B-cell response upon revaccination represents an early and unique marker of protection that may significantly accelerate the pre-clinical phase of vaccine development.

Fig 1. The applied flowcytometric B-cell phenotyping panel is able to identify HA specificity within B-cell subsets.

Mouse iliac lymph node cells were obtained 19 days post-prime immunization with alum adjuvanted FL H1#2316 (A) or PBS (B). Cells were subsequently stained for CD4, CD19, CD138, GL7, CD95, and a biotinylated HA probe (FL H1#5070) conjugated to APC (HA-APC) and PE (HA-PE) fluorochromes was used to assess HA binding among B cell subsets. Gate frequencies are indicated as frequency of parent or, if followed by a “*”, as frequency of CD19+ B-cells. Arrows drawn from gates to plots show the applied sequential gating steps to identify the cell populations shown in plot titles in (B) all gating steps prior to the identification of B cells are not shown. Data are representative for n = 8 immunized mice.

Fig 2. Germinal center B cell kinetics post-prime immunization do not correlate to vaccine protective efficacy.

Following the immunization schedule (A) a large cohort of mice were immunized three times at three week intervals (day 0, 21 and 42) with 30 μg alum-adjuvanted FL H1#2316, 0.3 μg alum-adjuvanted FL H1#2316, 30 μg alum-adjuvanted UFV#4157, or PBS alum-adjuvanted. At the indicated days (4, 7, 12 and 19 post-prime immunization), 8 mice per treatment group were sacrificed and iliac lymph nodes used for FACS analysis. The remaining mice (n = 10 per immunized cohorts, n = 8 for positive control CR6261 cohort) went trough a lethal influenza virus challenge on day 70 (25xLD50 of H1N1 A/Netherlands/602/2009) and were monitored for 21 days. A positive control group (n = 8) for influenza virus protection was intravenously injected with 300 μg of bnAb CR6261 at day 69. The challenge outcome is displayed as Kaplan-Meier survival (B). Relative counts of GC B cell subsets in iliac lymph nodes (quantified using the flow cytometric analysis outlined in Fig 1A) were measured at the indicated time-points post-prime immunization in mice (n = 8 per time-point per cohort) vaccinated with 30 μg alum-adjuvanted FL H1#2316, 0.3 μg alum-adjuvanted FL H1#2316, 30 μg alum-adjuvanted UFV#4157, or PBS alum-adjuvanted. Graphs in (C) show the total counts of GL7+CD95+ B cells while in (D) GL7+CD95+ B cells binding to both FL H1#5070-PE and FL H1#5070-APC conjugates are included. Each symbol represents one animal while group means are indicated by a horizontal bar.

Fig 3. Increased GC B cell expansion after boost-immunization with a non-protective UFV#4157 immunization.

Frequencies of cells binding to both FL H1#5070-PE and FL H1#5070-APC conjugates (HA+) among GC (GL7+CD95+) B cells (A) and non-GC (GL7-CD95-) B cells (B) in iliac lymph nodes were measured 4 days post-boost (day 25) in mice (n = 8 per time-point per cohort) vaccinated with 30 μg alum-adjuvanted FL H1#2316, 0.3 μg alum-adjuvanted FL H1#2316, 30 μg alum-adjuvanted UFV#4157 or alum-adjuvanted PBS. Each symbol represents one animal while group means are indicated by a horizontal bar. Statistical comparisons are made by comparing group means of the immunized groups in a one-way ANOVA, corrected for multiple comparisons using Tukey’s statistical hypothesis testing.

Fig 4. Early post-boost immunization antibodies do not correlate with protection conferred by mini-HA immunizations.

(A) ELISA titers against full-length HA derived from A/Brisbane/59/07 (same antigen as FL H1#2316) were determined in serum obtained at day 25 (4 days after the first boost immunization) from mice (n = 8 per group) immunized with high doses (30 μg or 60 μg) of the indicated antigens or PBS, all adjuvanted with alum. (B) Inhibition of CR9114 binding to A/Brisbane/59/07 determined by ELISA after pre-incubation of A/Brisbane/59/07 with day 25 post-boost serum antibodies from the same mice. Every dot represents data from a single animal, horizontal bars specify group means. In panel A, the grey area between dotted lines represents the highest and lowest LOD of the assay, which is calculated per each plate. Statistical comparisons are made by comparing group means of the miniHA vaccinated animals in a one-way ANOVA, corrected for multiple comparisons using Tukey’s statistical hypothesis testing.

Fig 5. Protective immunization regimens induce lower HA-specific B cells responses post-boost immunizations.

Frequencies of HA-binding cells among GC B cell (A) and non-GC B cell (B) populations in iliac lymph nodes as measured at day 25 (4 days post-boost immunization) in mice (n = 8 per time-point per cohort) vaccinated with the alum-adjuvanted Mini-HA UFV#4900, UFV#4650, UFV#4157 antigens (30 μg), FL HA FL H1#2316, at an equimolar dose of 60 μg, or PBS. Each symbol represents one animal while group means are indicated by a horizontal bar. Statistical comparisons are made by comparing group means of the miniHA vaccinated animals in a one-way ANOVA, corrected for multiple comparisons using Tukey’s statistical hypothesis testing.