Age-driven declined abilities of antigen trafficking, DC migration, and activation, and DC-mediated T cell activation are ameliorated by PGA/Alum
For inducing the adaptive immunity, the DCs should facilitate antigen trafficking to dLNs, antigen processing and presentation, and upregulation of the co-stimulatory molecules and cytokines [9]. Similar to previous studies [12, 24], we compared the in vivo antigen trafficking in the aged (18-months) and young (6-weeks) mice (n = 3 per group) using OVA via the near-infrared (NIR) fluorescence imaging system. Mice were subcutaneously administered IRDye800-labeled OVA (IR800-OVA), and in vivo fluorescent signals were observed at 1, 3, 6, 12, and 24 h post-injection. As shown in Figure 1A, IR800-OVA-young mice exhibited robust fluorescent signals in dLNs at 1, 3, and 6 h post-injection, and the fluorescent signals gradually decreased until 24 h post-injection. In dLNs of IR800-OVA-aged mice, however, the fluorescent signals were weak at 1 h and peaked at 3 h post-injection only. The mean fluorescence intensities (MFIs) used for quantitative measurements were also highest in the dLNs of IR800-OVA-young mice 1 h post-injection (215,625 MFI), whereas the MFIs in the dLNs of the IR800-OVA-aged mice were 4-fold lower (53,916 MFI) than those of the young mice 1 h post-injection (P < 0.001), suggesting delayed antigen trafficking in the aged mice compared to the young mice. To determine whether PGA/Alum restores antigen trafficking in aging, aged mice were subcutaneously administered IR800-OVA or combined with γ-PGA (IR800-OVA-γ-PGA), alum (IR800-OVA-Alum), or PGA/Alum (IR800-OVA-PGA/Alum). IR800-OVA-PGA/Alum-aged mice had at least 2-fold higher MFIs in the dLN than other aged groups at all the time points (P < 0.05). Notably, MFIs of the IR800-OVA-PGA/Alum-aged mice were similar to those of the IR800-OVA-young mice, 1 h post-injection. The robust MFIs were significantly 2-fold higher than those of the IR800-OVA-young mice until 3 to 24 h post-injection.
To determine whether PGA/Alum-increased antigen trafficking in aged mice was due to the migration of antigen-loaded DCs, aged mice (n = 3 per group) were intramuscularly (i.m.) injected with Alexa Fluor 647-OVA (Fluor-OVA) alone or in combination with γ-PGA (Fluor-OVA-γ-PGA), alum (Fluor-OVA-Alum), and PGA/Alum (Fluor-OVA-PGA/Alum). The total DC and Fluor-OVA+ DC numbers were determined in injected muscle and dLNs 3 h post-injection via flow cytometry, and the Fluor-OVA-young mice were used for comparison. As expected, significantly lower numbers of DCs and Fluor-OVA+ DCs were observed in injected muscle of the Fluor-OVA-aged mice (206 and 106 cells) than the Fluor-OVA-young mice (624 and 401 cells) (P < 0.01) (Fig. 1B). In the aged mice, the Fluor-OVA-PGA/Alum group exhibited 2-fold higher numbers of DCs and Fluor-OVA+ DCs (473 and 289 cells) than the other groups (206 and 106 cells in the Fluor-OVA; 160 and 104 cells in the Fluor-OVA-γ-PGA; 160 and 66 cells in Fluor-OVA-Alum) (P < 0.05). In addition, in the dLNs, significantly lower numbers of migrated DCs and OVA+ DCs were observed in the Fluor-OVA-aged mice (3,170 and 313 cells) than the Fluor-OVA-young mice (5,936 and 1,003 cells) (P < 0.05) (Fig. 1C). However, the Fluor-OVA-PGA/Alum-aged mice showed increased numbers of DCs and OVA+ DCs (6,076 and 877 cells) compared to the other aged mice (3,170 and 313 cells in Fluor-OVA; 3,612 and 595 cells in Fluor-OVA-γ-PGA; 3,050 and 454 cells in the Fluor-OVA-Alum). Such effects were also observed in the Fluor-OVA-PGA/Alum-aged mice 12 h post-injection compared to the other aged mice (P < 0.05) (Supplementary Fig. S1A and B). These results indicate that PGA/Alum can restore the reduced DC migration ability observed in the aged mice.
Next, DC functions were examined using splenic CD11c+ DCs purified from the aged and young mice (n = 4 per group). Flow cytometry showed that the co-stimulatory molecules (CD40 and CD86) levels were significantly lower in the DCs from the aged mice (P < 0.001), but PGA/Alum triggered highly their levels on DCs from the aged mice compared to PBS, γ-PGA, and alum (P < 0.01) (Fig. 1D and E). In DCs from the aged mice, production of the inflammatory cytokines (IL-6, IFN-γ, TNF-α, IL-1α, MCP-1, and IL-1β) were also increased by PGA/Alum compared to PBS, γ-PGA, and alum (P < 0.05) (Supplementary Fig. S2), suggesting that PGA/Alum effectively enhances the DC activation of the aged mice. Moreover, the antigen uptake and processing of DCs were tested using FITC-OVA or DQ-OVA, which are well-established models for antigen uptake and processing, respectively. The percentages of FITC-OVA+ DCs and DQ-OVA+ DCs were lower in the DCs from the aged mice than in those from the young mice, and PGA/Alum increased the percentages compared to the γ-PGA or non-treatment when using DCs from the aged mice (P < 0.05) (Supplementary Fig. S3). However, the small fold changes, even if statistically significant, were observed in the DCs of all the age groups. Since the cross-presentation ability of DCs is a unique function to directly activate the CTLs capable of clearing the viral infection [9], the DC-mediated CD8+ T cell activation was assessed using the OVA-specific MHC class I (H-2Kb)-restricted OT-I CD8+ T cells. CD11c+ DCs were incubated with OVA alone or combined with γ-PGA (OVA-γ-PGA), alum (OVA-Alum), and PGA/Alum (OVA-PGA/Alum) for 6 h, followed by co-culture with the OT-I CD8+ T cells. The T cell activation was analyzed using the CFSE-diluted profiles and IFN-γ production via flow cytometry. As shown in Figure 1F, the percentages of CFSE− CD8+ T cells were 0.7-fold lower in co-culture with OVA-exposed DCs from the aged mice than in the young mice but were significantly 1.7-fold higher in the co-culture with the OVA-PGA/Alum-exposed DCs than the OVA, OVA-γ-PGA, or OVA-Alum-exposed DCs (P < 0.001). The percentages of the IFN-γ+CD8+ T cells were also 0.8-fold lower in the co-culture with OVA-exposed DCs from the aged mice than in those from the young mice (P < 0.001), but the age-driven low percentages were significantly 1.2-fold higher in the co-culture with OVA-PGA/Alum-exposed DCs than in the DCs exposed to OVA or OVA-γ-PGA (P < 0.01) (Fig. 1G). These results suggest that PGA/Alum can enhance the age-driven decline of antigen trafficking, DC migration and activation, and DC-mediated T cell activation.
PGA/Alum robustly increases the antigen-specific CTL activity effectively in the aged mice
Since adaptive immune responses are essential for inducing vaccine efficacy [25], PGA/Alum was investigated to determine whether it improves the antigen-specific cellular and humoral immunity in the aged mice (n = 4 per group) by i.m. administering OVA alone or in combination with γ-PGA (OVA- γ-PGA), alum (OVA-Alum), or PGA/Alum (OVA-PGA/Alum) on days 0, 14, and 21. Seven days after the last immunization, splenocytes were stimulated with the MHC class I-restricted OVA257−264 peptide and subjected to the ELISPOT assay to evaluate the CD8+ CTL activity. As shown in Figure 2A, significantly higher IFN-γ+ spot-forming units (SFUs) were observed in the OVA-PGA/Alum group (133 ± 74 SFUs) than in the OVA (33 ± 16 SFUs), OVA-γ-PGA (20 ± 9 SFUs), and OVA-Alum (14 ± 6 SFUs) groups (P < 0.05). Additionally, the IFN-γ+ SFUs from the aged and young mice were compared finding that the degree of IFN-γ+ SFUs was similar between OVA-PGA/Alum-immunized aged and young mice (Supplementary Fig. S4A). Flow cytometry also showed that the percentages of the IFN-γ+ CD8+ T cells were significantly higher in the OVA-PGA/Alum-aged mice (3.9 ± 0.3%) than in the OVA- (2.9 ± 0.2%), OVA-γ-PGA- (2.6 ± 0.3%), and OVA-Alum- (2.8 ± 0.3%) aged mice (Fig. 2B). Moreover, the generation of OVA257−264 tetramer+ CD8+ T cells was higher in the OVA-PGA/Alum-aged mice (1.4 ± 0.1%) than in the aged mice immunized with OVA (0.7 ± 0.3%), OVA-γ-PGA (0.4 ± 0.1%), and OVA-Alum (0.9 ± 0.2%) (Fig. 2C), suggesting that the CTL activity in aged mice could be raised, similar to that of young mice, by PGA/Alum.
Next, the humoral immune response was confirmed by measuring the OVA-specific antibody (Ab) titer in the sera via ELISA. The OVA-specific IgG titers were almost 8-fold higher in the OVA-PGA/Alum-aged mice (452,018 ± 49,238 titer) than in the aged mice immunized with OVA (28,594 ± 23,021 titer), OVA-γ-PGA (22,215 ± 12,539 titer), and OVA-Alum (92,935 ± 62,881 titer) (P < 0.001) (Fig. 2D). Titers of OVA-specific IgG subclasses, Th1-biased Ab (IgG2b) and Th2-biased-Ab (IgG1), also showed significant 2-fold increase in the OVA-PGA/Alum-aged mice than in the other aged mice (Fig. 2E and F). The serum IgG titers were further compared between the aged and young mice immunized with OVA-PGA/Alum. Unlike the CTL activity, IgG titers were still 10-fold lower in the OVA-PGA/Alum-aged mice than in the OVA-PGA/Alum-young mice (Supplementary Fig. S4B). The IgG2b titer was also 5-fold lower in the OVA-PGA/Alum-aged mice than in the OVA-PGA/Alum-young mice, whereas the IgG1 titer was not different between the groups (Supplementary Fig. S4C). These results suggest that the use of PGA/Alum in aged mice can improve the antigen-specific CTL activity more effectively than Ab production.
The use of PGA/Alum as an adjuvant enhances the influenza pH1N1 vaccine antigen-specific immune responses in aged mice
To investigate whether the recovery of age-driven immune suppression by PGA/Alum can act similar to that of the commercially available vaccine antigens, the aged mice (n = 5–10 per group) were immunized twice with the influenza pH1N1 split-vaccine antigen alone (vaccine) or mixed with γ-PGA (vaccine-γ-PGA), alum (vaccine-Alum), or PGA/Alum (vaccine-PGA/Alum) at 2-week intervals. Two weeks after the last vaccination, the vaccine-specific adaptive immune responses were analyzed using the splenocytes and sera via the IFN-γ ELISPOT assay, hemagglutinin-inhibition, and IgG-specific ELISA assays. The ELISPOT assay was performed by stimulating the splenocytes with a UV-inactivated pH1N1 and revealed that IFN-γ+ SFUs were significantly more than 3-fold higher in the vaccine-PGA/Alum group (81 ± 23 SFUs) compared to the vaccine (11 ± 7 SFUs), vaccine-γ-PGA (24 ± 19 SFUs), and vaccine-Alum (7 ± 3 SFUs) groups (P < 0.001) (Fig. 3A). The HI titers against the pH1N1, an indicator of the protective efficacy of the influenza vaccine, were also increased in the sera from the vaccine-PGA/Alum group [190 geometric mean titer (GMT)] compared to those from the vaccine (5 GMT; P < 0.001), vaccine-γ-PGA (48 GMT; P < 0.05), and vaccine-Alum (44 GMT) groups (Fig. 3B). In addition, a significantly higher pH1N1-specific IgG titer was observed in the sera from the vaccine-PGA/Alum group (308,201 ± 193,963 titer) than in the vaccine (4,321 ± 3,458 titer), vaccine-γ-PGA (66,197 ± 23,618 titer), and vaccine-Alum (73,167 ± 32,633 titer) groups (P < 0.001) (Fig. 3C). The Ab titers of IgG subclasses, IgG2b and IgG1, were also significantly higher in the sera from the vaccine-PGA/Alum group than in the other groups (P < 0.01) (Fig. 3D). These findings demonstrate that PGA/Alum triggers influenza antigen-specific cellular immune responses in the aged mice, accompanied by an increase in humoral immune responses.
PGA/Alum suppresses the proportion of age-associated CD8 + T cell subset and gene expression of the inhibitory regulators within the CD8+ T cells, contributing to a decrease in the dysfunctional CD8+ T cells
To analyze the alteration of the immune profile by PGA/Alum in the aged mice in detail, the single-cell RNA analysis, a widely used tool to identify differentially expressed genes within one cell was employed. The scRNA-seq of the immune cells was analyzed using the CD45+ cells sorted from the splenocytes of the vaccine-aged mice and vaccine-PGA/Alum-aged mice (pool of n = 3 per group) on day 14 after the last immunization. To compare the cellular landscape, the CD45+ cells were sorted from the splenocytes of the vaccine-young mice. Clustering of the data was performed based on the key signature genes and standard surface markers and represented in the uniform manifold approximation and projection (UMAP) dimension reduction (Supplementary Fig. S5A and B). Similar to previous findings [26, 27], the aged mice exhibited lower proportions of the CD8+ and CD4+ T cells and a higher proportion of the regulatory T cells (Treg) than the young mice, despite immunization with the vaccine (Supplementary Fig. S5C). Since the functional CD8+ T cell subsets are key players in steering the immune responses to execute viral clearance [25], we focused on the CD8+ T cell subsets. As shown in Figure 4A, the vaccine-aged mice exhibited a 6.6-fold higher percentage of age-associated (CD44+PD-1+, 10.6%), 2.3-fold higher percentage of effector memory (CD44+PD-1−, 24.2%), and 1.33-fold lower percentage of naive (CD44−CD62L−, 65.2%) subsets than the vaccine-young mice (CD44+PD-1+, 1.6%; CD44+PD-1−, 11.4%; CD44−CD62L−, 87%), as previously reported [26, 28]. The most striking observation was a 2.4-fold lower percentage of the age-associated subset in the vaccine-PGA/Alum-aged mice (4.3%) than in the vaccine-aged mice (10.6%). However, there was little change in the percentages of effector memory and naive subsets: 1.09-fold and 1.06-fold higher in the vaccine-PGA/Alum-aged mice (26.4% and 69.3%) than in the vaccine-aged mice (24.2% and 65.2%), respectively. Within the age-associated CD8+ T cell subset, the expression of the phenotypic and transcriptional markers of senescence or exhaustion, including Pdcd1, Tox, Lag3, and Gzmk (encoding PD-1, Tox, Lag3, and Granzyme K proteins) was lower in the vaccine-PGA/Alum-aged mice than in the vaccine-aged mice (Fig. 4B). The gene expression profiles were further analyzed because the inhibitory regulators in modulating the T cell activation are important for controlling the CD8+ T cell function. As shown in Figure 4C, the CD8+ T cells of the vaccine-PGA/Alum-aged mice had significantly lower gene expression of the negative regulators of T cell activation (lgals1, ctla2a, Nfkbiz, and Bhlhe40, encoding Galectin-1, CTLA-2 alpha NF-Kappa-B inhibitor ζ, and BHE40 protein) than those of the vaccine-aged mice (P < 0.05). The PGA/Alum-altered gene levels were the median values between aged and young mice immunized with vaccine alone. These results demonstrate that PGA/Alum can decrease the proportion of age-associated CD8+ T cell subsets and gene levels of negative regulators in CD8+ T cell response, consequently resulting in the recovery of the functional CD8+ T cells in the aged mice.
Aged mice immunized with PGA/Alum-adjuvanted influenza vaccine are robustly protected against the influenza virus infection
Finally, to evaluate the adjuvanticity of PGA/Alum in the flu vaccine in aged mice, we first immunized young and aged mice (n = 5 per group) with the flu vaccine and compared the protection to the pH1N1 virus (A/California/04/09) infection. Similar to the previous reports showing that the flu vaccine efficacy decreases with advancing aging [29, 30], all vaccine-aged mice had died 6 days after the virus challenge (0% survival), whereas all vaccine-young mice survived even though the bodyweight loss occurred (100% survival) (Supplementary Fig. S6A and B). Next, the aged mice (n = 5 per group) were i.m. administered various doses of the pH1N1 vaccine antigen (0.25, 0.5, and 1 µg) alone or mixed with PGA/Alum. As shown in Figure 5A and B, the vaccine-PGA/Alum group at the highest vaccine dose (1 µg) exhibited a 100% survival rate with very little body weight loss. The vaccine-PGA/Alum group at 0.5 µg of vaccine dose also showed 100% survival despite body weight loss, whereas the group at the lowest vaccine dose (0.25 µg) was partially protected (75% survival). These findings indicate that PGA/Alum can recover the age-driven decline in the protective vaccine efficacy to a similar extent as the efficacy of the vaccine alone in young mice. To elucidate the ability of PGA/Alum as a vaccine adjuvant, the aged mice (n = 5 per group) were administered the flu vaccine alone (0.5 µg) or mixed with γ-PGA, alum, or PGA/Alum, and then challenged with the pH1N1 as described above. As expected, the vaccine-PGA/Alum group showed a 100% survival rate, whereas vaccine-γ-PGA group and vaccine-Alum group showed 50% and 40% survival rate, respectively (Fig. 5C and D). All vaccine alone group had died (0% survival, P < 0.001).
Since clearance of virus from the lung is a pivotal factor in protective vaccine efficacy, the viral titers were determined in the lung from the vaccinated aged mice (n = 3 per group) on days 3, 5, and 7 after the pH1N1 viral challenge. The vaccine-PGA/Alum group was significantly better at clearing the infected virus (Fig. 5E). On day 7 post-challenge, the vaccine-PGA/Alum group exhibited complete viral clearance (P < 0.05), whereas the other groups still had high viral loads by day 7. Moreover, the immunofluorescence staining revealed that the presence of influenza virus was very low in the lung section from the vaccine-PGA/Alum group 7 days post-challenge, whereas the virus was strongly detected in the other groups (Fig. 5F). Collectively, these results indicate that the use of PGA/Alum as an adjuvant improves the protection against the influenza virus by facilitating the viral clearance in the aged mice.