Influenza virus infections pose a major public health threat, accounting for 3.5 million severe infections and more than 400,000 deaths globally each year (1). Most seasonal vaccines consist of inactivated influenza virus components, which induce antibody responses against immunodominant epitopes in the viral hemagglutinin (HA) and neuraminidase (NA) proteins. The genes that encode HA and NA undergo continuous changes (antigenic drift), which necessitates annual reformulation and revaccination, leading to reduced vaccine coverage. Vaccine effectiveness thus varies depending on the accuracy of preseasonal predictions, and inactivated seasonal influenza vaccines generally provide insufficient protection against pandemic viruses (2). On page 869 of this issue, Wang et al. (3) explore an unconventional strategy to overcome these shortcomings by complementing inactivated influenza virus vaccines with an adjuvant that triggers mucosal immune responses to elicit rapid protection against a variety of influenza virus strains in mice and ferrets.
Current strategies for the development of such universal flu vaccines mainly focus on the generation of broadly protective antibodies directed against conserved but immunosubdominant viral surface epitopes that are accessible to antibody binding, such as the stalk region of HA and, recently, the active site of NA (4, 5). In contrast to antibody responses (which are produced by B cells), virus-specific CD8+ T cells generated in response to natural influenza virus infection may provide broad protection against infections with numerous virus subtypes (heterosubtypic protection) (6).
2′,3′-cyclic guanosine monophosphate–adenosine monophosphate (cGAMP) is a second messenger produced in response to viral infections and a potent activator of the innate immune sensor stimulator of interferon genes (STING) (7). To mimic natural influenza infection and to elicit CD8+ T cell–mediated immunity, Wang et al. used cGAMP as an adjuvant to an inactivated influenza virus vaccine. Pulmonary surfactant (PS) is a phospholipoprotein complex produced by alveolar epithelial cells (AECs) to reduce surface tension and prevent alveolar collapse. Because PS is recognized by lung-resident alveolar macrophages, the authors used lipid components of PS to encapsulate cGAMP. They found that intranasally administered PS-GAMP nanoparticles were readily taken up by alveolar macrophages in mice. cGAMP was transferred from alveolar macrophages to AECs, where STING was subsequently activated.
Nasal delivery of inactivated H1N1 influenza virus and the adjuvant PS-GAMP leads to uptake by alveolar macrophages. cGAMP is transferred to AECs, where it activates STING. This stimulates DC differentiation and maturation, increases antibody production, and leads to a robust effector CD8+ T cell response and TRM cells. Together, this provides strong heterosubtypic immunity in mice and ferrets.
GRAPHIC: A. KITTERMAN/SCIENCE
Intranasal application of PS-GAMP nanoparticles together with an inactivated H1N1 influenza virus vaccine provided robust heterosubtypic protection—including against seasonally circulating H3N2, influenza B virus (IVB), and highly human-pathogenic avian H5N1 and H7N9—in mice and ferrets. Protection was observed as early as 2 days after vaccination and was maintained for up to 6 months. The PS-GAMP adjuvant vaccine elicited robust virus-specific CD8+ T cell responses days after immunization, and high-antibody titers were detected 2 weeks after vaccination.
Heterosubtypic protection after live viral infection has been linked to the presence of cross-reactive T cells (8, 9). Recent studies have uncovered broad cross-reactivity among human influenza virus–specific CD8+ T cells (10, 11), and cross-reactive CD8+ T cells have been associated with protection against heterosubtypic symptomatic influenza in humans (12), thus making T cells an appealing target for universal influenza vaccines. Animal studies have indicated that local immunity—in particular, tissue-resident memory T cells (TRM cells) in the lung—are critical determinants of protection against heterosubtypic influenza virus infection (13, 14). Hence, successful T cell–based universal vaccine strategies will likely require the generation of cross-reactive TRM cells in the respiratory mucosa.
The only live attenuated influenza vaccine (LAIV) licensed for use in humans is administered as a nasal spray. It induces cross-reactive T cells and generates CD8+ TRM cells in respiratory tissues in mice, which is generally not observed with commonly used inactivated influenza vaccines given by intramuscular or subcutaneous injection (14). Both mucosal delivery and viability of LAIV were required for the generation of protective TRM cells in mice. Wang et al. found that intranasal application of inactivated H1N1 with PS-GAMP led to an early increase in natural killer cells (which have antiviral functions) and pulmonary dendritic cells (which bridge innate and adaptive immunity by presenting antigens to T cells), followed by an accumulation of CD8+ T cells with a typical TRMphenotype in the lungs of vaccinated mice. Further experiments in mice revealed that STING activation in AECs orchestrated dendritic cell recruitment and subsequent CD8+ TRM cell generation (see the figure).
These results point to a central role of the alveolar epithelium in this protective multicellular cross-talk. AEC secretion of the cytokine granulocyte-macrophage colony-stimulating factor (GM-CSF) has been found to enhance the antigen-presenting capacity of lung dendritic cells, resulting in accelerated CD8+ T cell–mediated clearance of influenza viruses (15). Intranasal PS-GAMP administration also transiently increased GM-CSF and the cytokine interferon-β (IFN-β) in the lung, but the exact mechanisms by which cGAMP-STING–activated AECs expand lung dendritic cells and promote CD8+ T cell responses remain to be explored.
Promising advances toward universal influenza vaccines have been made in recent years, and several candidates are currently undergoing clinical testing (4, 5). Recent strategies for universal influenza vaccines have centered on the generation of broadly protective antibodies, whereas the approach of Wang et al. elicited T cell–mediated heterosubtypic immunity. However, heterosubtypic protection induced by intranasal LAIV or previous influenza virus infection in humans is generally less effective compared with experimental mouse models (5, 14). This discrepancy may be caused by interspecies variations and general differences between controlled experimental models and clinical reality. Although STING ligands have recently attracted attention as potential immunotherapeutics in cancer, their role in human T cell and vaccine responses remains to be investigated. It will therefore be critical to evaluate the efficacy of cGAMP as an adjuvant for mucosal influenza vaccines and their effect on cross-protective T cells in humans and other natural hosts for influenza viruses, such as pigs. Ultimately, effective adjuvants and targeted delivery systems combined with broadly protective vaccine antigens to elicit both cross-reactive CD8+ T cells and cross-protective antibodies may represent the most effective approach for urgently needed universal influenza vaccines.
(원문: 여기를 클릭하세요~)