香京julia种子在线播放

Adult CB6F1/OlaHsd females were purchased from Harlan (Horst, The Netherlands) together with BALB/c OlaHsd females and C57BL/6 OlaHsd males. The latter were crossed to produce F1 CB6F1 mice. All mice were bred, kept in pathogen-free animal facilities in accordance with local guidelines and used at 1 week (neonates) or 6–8 weeks (adults) of age. All animal experiments were approved by the Geneva veterinary office and conducted under relevant Swiss and European guidelines.

We used an experimental monovalent purified subunit influenza vaccine composed of HA from the influenza strain H1N1 A/California/7/2009 [Novartis Vaccines (a GSK company), Siena, Italy]. Groups of 5–8 CB6F1 neonatal (1-week-old) and adult mice were immunized subcutaneously (s.c.) with 100 µl of the plain HA (1 µg) or in combination with either CAF01 (250 µg DDA/50 μg TDB, Statens Serum Institut, Copenhagen, Denmark), IC31^® (KLK/ODN1a = 100 nmol/4 nmol, Valneva Austria GmbH), GLA-SE (5 µg GLA and 2% v/v squalene, Infectious Diseases Research Institute, Seattle, WA, USA), or DDA-Curdlan (250 µg DDA/50 μg Curdlan, Statens Serum Institut, Copenhagen, Denmark) produced according to the protocol previously described for DDA-TDB (25). Curdlan was purchased from Sigma-Aldrich.

Mice were immunized at the base of the tail and inguinal draining lymph nodes (LNs) were harvested, except for the experiments with GLA-SE in which mice from both age groups were injected s.c. (100 µl) at the scruff of the neck and brachial draining LNs harvested. This use of the base of the tail as injection site was required to comply with the new local animal welfare guidance to limit procedures requiring anesthesia. We carefully checked that this change did not affect our results (Figure S1 in Supplementary Material).

Viral challenge was performed as recently described (6) using a mouse-adapted H1N1 Influenza strain (A/Netherlands/602/2009, passage 2 in mice). Virus was grown on MDCK cells (ATCC). Fifty-six days post-immunization, 2 × 10³ PFU of virus in 20 µl sterile PBS were instilled intranasally under anesthesia induced by Ketasol^® (Graeub) and Rompun^® (Bayer). Mice were observed daily to monitor body weight and survival. Mice showing more than 20% of body weight loss were humanely euthanized.

Mice were bled from the tail vein at the indicated time points except for neonatal mice at day 0 that were bled by decapitation. Titration of HA-specific total IgG, IgG1, and IgG2a titers was performed by ELISA on individual serum samples as previously described (6).

Avidity was measured by ELISA as the overall strength of binding between antibody and antigen, using plates incubated for 15 min with increasing concentration of ammonium thiocyanate (NH₄SCN) from 0 to 1.5 M. Antibody avidity was defined as the amount of antibody eluted for each increment of NH₄SCN concentration. Arbitrarily, the percentage of antibody eluted between 0 and 0.25 M of NH₄SCN was assigned as low-avidity antibody fraction, between 0.25 and 0.75 M as intermediate and >0.75 M as high-avidity antibody fraction.

Hemagglutinin-specific plasma cells were quantified by direct ex vivo ELISpot assay as previously described (6).

Cells from the two draining LNs of each individual mouse were pooled and stained with fluorescently labeled antibodies to GL7, CD8, B220, TCR-β, CD95 (Fas) (all from BD Biosciences), PD-1, Ter119, GR1, CD11c (all from eBioscience), and CD4 (all from BioLegend). CXCR5 staining was performed using purified anti-CXCR5 (BD Biosciences), followed by FITC anti-rat IgG (Southern Biotech), and normal rat serum (eBioscience). The stained cells were analyzed using a Gallios cytometer (Beckman Coulter) and the generated data analyzed using FlowJo Software (Tree Star).

Germinal centers in the draining LNs of immunized mice were stained and quantified as previously described (6). Sections were visualized and photographed with a Zeiss LSM700 confocal microscope (objective: 20×) or a Mirax scan microscope (Zeiss). Images were acquired with Zeiss LSM image browser software (Zeiss) or the Pannoramic Viewer software (3DHistec).

Data were analyzed using Prism 6.0 (GraphPad Software) and presented as mean ± SEM of at least three independent experiments. Difference between groups was analyzed as described in figure legends. P-values less than 0.05 were considered statistically significant.

We first compared the antibody titers elicited by a subunit monovalent influenza vaccine containing HA administered alone or formulated with GLA-SE, IC31^®, or CAF01. Neonates (referred to as “1 week”) and adult mice were immunized twice at days 0 and 21. In adults, a single dose of unadjuvanted HA was sufficient to elicit HA-specific antibody responses (Figure 1). These responses were strongly enhanced by each tested adjuvant (Figure 1), as recently reported (12). The highest primary IgG responses were induced by HA/GLA-SE (Figure 1A), followed by HA/IC31^® (Figure 1B), and HA/CAF01 (Figure 1C). HA/GLA-SE induced antibodies with a more pronounced IgG2a profile, while HA/IC31^® and HA/CAF01 primarily induced IgG1 antibodies.

Consistently with our previous findings (6), neonates did not raise detectable primary IgG responses to unadjuvanted HA. They also poorly responded to a single dose of HA/GLA-SE or HA/IC31^®. In contrast, HA/CAF01 strongly increased primary HA-specific IgG and particularly IgG1 responses (Figure 1C). Neonatal responses to HA/CAF01 were significantly higher than to HA/IC31^® (P ≤ 0.004) and HA/GLA-SE (P ≤ 0.0004). Neonatal responses to the second dose of HA/CAF01 were also significantly higher than those elicited by unadjuvanted HA, although neither primary nor secondary responses reached the titers elicited by adjuvanted vaccines in adults (Figure 1).

Thus, adult and neonatal requirements for B cell adjuvanticity differ: HA/GLA-SE is the strongest antibody inducer in adults, whereas only CAF01 adjuvantation succeeds in eliciting potent primary responses in neonates.

T_FH cells are critical for GC formation, controlling the number of GC B cell divisions, and are thus essential for the generation of high-affinity matured antibodies (26, 27). The induction of T_FH cells in early life is challenging: we previously reported that both alum-adsorbed (10) and MF59^®-adjuvanted (6) vaccines failed to generate T_FH cells in neonates. CD4⁺ CXCR5^highPD-1^high T_FH cells were thus measured in the draining LNs after neonatal or adult (control) immunization. T_FH cell responses increased significantly upon HA/CAF01 immunization in both age groups, reaching similar numbers in adults and neonates (Figures 2A,B).

Unexpectedly, despite the poor neonatal primary antibody responses to HA/GLA-SE and HA/IC31^®, a significant increase in CD4⁺CXCR5^highPD-1^high T_FH cells, reaching adult levels, was also observed in HA/GLA-SE- and HA/IC31^®-immunized neonates (Figures 2A,B). Neonatal T_FH cell responses elicited by HA/GLA-SE peaked at day 10 but dropped down relatively rapidly over time (Figure S2A in Supplementary Material). In contrast, less than 0.05% of CD4⁺ CXCR5^highPD-1^high T_FH cells were repeatedly observed in naïve infant mice (not shown). Thus, all tested adjuvants passed the challenge of inducing T_FH cells in neonates.

As GC reactions are essential determinants of the magnitude and the quality of antibody responses and humoral memory, we next assessed whether these adjuvants differed in their capacity to induce GCs. B220⁺GL7⁺CD95⁺ GC B cells were quantified by flow cytometry in draining LNs at day 10 or 12 upon immunization, as indicated in figure legends. In adults, the three adjuvants significantly increased the numbers of GC B cells when compared with unadjuvanted HA (Figures 3A,B). The few GC B cells in the draining LNs of neonates that received unadjuvanted HA were comparable with that observed in non-draining LNs (not shown), representing background levels. In neonates, adjuvantation with CAF01 strongly augmented both the number (Figure 3B) and proportion of GC B cells compared with unadjuvanted HA (HA/CAF01: 1.99% ± 0.22 vs HA: 0.30% ± 0.05, P < 0.0001). In contrast, only few GC B cells were observed following HA/IC31^® or HA/GLA-SE neonatal immunization. Neonatal GC B cell responses induced by HA/GLA-SE remained low even later after immunization (Figure S2B in Supplementary Material). In contrast, in adult mice higher GC B cell responses were observed at day 12 following vaccination with HA/GLA-SE (Figure S2B in Supplementary Material). Thus, a single dose of HA/CAF01-induced significantly higher numbers of neonatal GC B cells than unadjuvanted HA, HA/GLA-SE, or HA/IC31^®, in consistency with the observed higher antibody titers (Figure 1).

To explore whether the GC B cells identified by flow cytometry were organized in bona fide GC structures, we imaged follicular B cells (IgD, green), GC B cells [peanut agglutinin (PNA), red] and CD4⁺ T cells (CD4, purple) in draining LN sections post-prime (Figure 3C). In adults, highly organized IgD⁻PNA⁺ GC structures were observed. These GCs were very few in adult mice that received unadjuvanted HA and increased following immunization with the adjuvanted HA, irrespective of which adjuvant (Figure 3C and data not shown). Vaccination of neonatal mice with unadjuvanted HA, HA/IC31^®, or HA/GLA-SE did not generate bona fide GC structures. In contrast, well-organized GCs were observed in HA/CAF01-immunized neonates (Figure 3C), albeit PNA⁺ GCs induced by HA/CAF01 remained fewer (1 week: 1.85 ± 0.29 vs adults: 3.44 ± 0.46, P ≤ 0.01 per LN section) and smaller (1 week: 23,883 ± 9,973 µm² vs adult: 42,888 ± 9,810 µm²) than in adults. Thus, a single dose of HA/CAF01 elicited bona fide GCs even in neonates.

Early life immunization in general induces reduced titers of antibodies and these are of lower avidity compared with responses achieved by immunization in adults (1, 28). We thus asked whether a single dose of HA/CAF01 also affected these hallmarks of neonatal B cell responses. First, serum IgG antibody responses were measured for up to 9 weeks post-prime. In adults, HA/CAF01-induced significantly higher antibodies than unadjuvanted HA and these responses reached a plateau already 3 weeks after prime (Figure 4A). A similar enhancement of antibody responses was observed in neonates, although with slower kinetics (peak at 5 weeks, Figure 4A) and, as previously reported, lower magnitude (Figure 1C). HA/CAF01-induced antibodies persisted for at least 9 weeks post-prime (last time point assessed, Figure 4A) in both neonates and adults.

CAF01 (29) and IC31^® (15) both induce a depot effect which may allow a slow antigen release from the injection site and contribute to sustained antibody responses. However, antibodies induced by HA/IC31^® in neonates remained very low over time (Figure S3 in Supplementary Material). Thus, a depot effect is not sufficient to eventually elicit antibody responses in early life.

The avidity of HA-specific IgG antibodies was measured early (3 weeks) and late (9 weeks) after a single dose of HA/CAF01 (Figure 4B). Avidity is shown as percentages of eluted HA-specific antibodies after treatment with increasing concentrations of ammonium thiocyanate (NH₄SCN) (30). At 3 weeks post-prime, a higher proportion of low-avidity antibodies (eluted with a low concentration of ammonium thiocyanate) was observed in neonates than in adults (Figure 4B). However, avidity increased rapidly following neonatal immunization with HA/CAF01, reaching adult-like levels at 9 weeks (Figure 4B).

We previously reported that neonatal immunization with aluminum adjuvants elicits abortive and rapidly terminated GC responses (10). In contrast, a single dose of HA/CAF01-induced sustained GC activity: HA/CAF01-induced GC B cells persisted in both age groups for at least 5 weeks after a single immunization, albeit at significantly lower numbers than in adults in agreement with previous results (Figure S4 in Supplementary Material). Thus, the neonatal B cell adjuvanticity of CAF01 results into a potent and sustained induction of bona fide GC B cells, which rapidly generates high-affinity and persistent primary antibody responses.

We next asked whether the HA-specific antibody responses elicited by a single dose of HA/CAF01 in neonates were sufficient to confer protection against influenza viral challenge, which is difficult to achieve. Neonatal mice received a single dose of unadjuvanted HA, HA/CAF01, or PBS (control) and were challenged intranasally 8 weeks later with a lethal dose of matching influenza A/H1N1 virus. As additional controls, we challenged HA/CAF01- (positive control) or PBS- (negative control) injected adult mice. The body weight of mice that received PBS or unadjuvanted HA declined to 80% on average within 6 days (Figures 5A,B). In contrast, the body weight of HA/CAF01-immunized mice only transiently declined by less than 10% (4–6%) in both age groups (Figures 5A,B). The protection mediated by HA/CAF01 translated into a survival rate of 100% in both age groups (Figures 5C,D). In neonates, CAF01 adjuvantation was required for protection, as only 3/8 mice that received unadjuvanted HA survived up to day 8 (Figure 5C). Among these survivors, two mice had signs of severe infection and lost considerable body weight (17.1 and 19.9% of weight loss, respectively). Similarly, in the group of neonates that received PBS, one of two surviving mice had signs of severe infection on day 8 (18.3% of weight loss). Only one mouse in each group either was not fully infected or recovered spontaneously. Thus, a single dose of HA/CAF01 is sufficient to confer protection against influenza viral challenge even when given to neonates.

The potent neonatal B cell adjuvanticity of the Mincle-activating CAF01 adjuvant suggests that CLR targeting may be critical in neonates rather than signaling through TLRs. To explore this possibility, we tested the neonatal B cell adjuvanticity of Curdlan, which binds the CLR Dectin-1. Thus, by replacing the CAF01 immunostimulator TDB with Curdlan in the cationic surfactant DDA, we generated DDA-Curdlan.

We first confirmed the adjuvanticity of DDA-Curdlan in adult mice: a single dose of HA/DDA-Curdlan strongly enhanced HA-specific IgG antibodies compared with unadjuvanted HA [anti-HA IgG (log₁₀) 4.4 ± 0.14 vs 2.1 ± 0.07, P < 0.001, 4 weeks post-prime], a similar effect to that induced by CAF01. Then, we vaccinated neonatal mice with unadjuvanted HA or formulated with either CAF01 or DDA-Curdlan. HA/DDA-Curdlan strongly increased primary HA-specific neonatal antibody responses, with no differences to HA/CAF01 (Figure 6A). Both adjuvants significantly increased HA-specific antibody titers when compared with unadjuvanted HA, at all assessed time points for up to 10 weeks post-prime. This reflected the enhancement of both T_FH and GC B cell responses by DDA-Curdlan, with no differences compared with CAF01 (Figure 6B).

Next, we evaluated the capacity of both formulations to induce antibody-secreting cells (ASCs) able to home into the bone marrow (BM), another hallmark of potent GC reactions which is rarely reached in early life (28). Neonatal mice were immunized s.c. at day 0 with HA/CAF01 or HA/DDA-Curdlan. Adult mice receiving HA/CAF01, HA only or naïve were included as control groups. Except for naive animals, all groups of mice were boosted 10 weeks later with HA only and ASCs were measured 1-week post-boost by ELISpot assay. High numbers of HA-specific ASCs were retrieved from the neonatal BM after vaccination with either HA/CAF01 or HA/DDA-Curdlan (Figure 6C). Neonatal responses were higher than those observed in adults immunized with HA only, but did not reach the levels observed in adults immunized with HA/CAF01. High numbers of ASCs were also observed in the spleens of immunized neonates (Figure 6D). Thus, the neonatal B cell adjuvanticity of CAF01 and DDA-Curdlan results in a potent and sustained induction of bona fide GC B cells, which rapidly generates high-affinity and long-lived ASCs.