|Home | About | Journals | Submit | Contact Us | Français|
The major γδ T cell subset in the human peripheral blood expresses the Vγ9δ2 TCR and recognizes non-peptidic prenyl pyrophosphate antigens such as isopentylpyrophosphate (IPP). Upon activation the γδ T cells rapidly secrete antiviral cytokines similar to classical memory αβ T cells. Here we have investigated the ability of γδ T lymphocytes from human PBMC to become activated by influenza A virus infection. Vγ9Vδ2 T lymphocytes rapidly upregulate expression of CD25 and CD69 and produce IFN-γ following influenza infection of PBMC. Moreover, the recognition is cross-reactive between various subtypes of influenza, but not with vaccinia virus. Vγ9Vδ2 T cell responses are potently reduced by the HMG-CoA reductase inhibitor mevastatin, which inhibits the mevalonate pathway and IPP synthesis. Our results indicate that influenza virus infection induces the rapid activation and function of Vγ9Vδ2 T lymphocytes in the peripheral blood via a mechanism that depends on the mevalonate pathway.
Development of a safe and effective influenza vaccine is of great importance as shown by the 2009 H1N1 pandemic. Induction of influenza-specific memory T cell responses would likely provide cross-reactive subtype protection [1; 2; 3; 4], however generation of a vaccine that induces memory αβ T cells has proven difficult. Vγ9Vδ2 T cells are one of the most prevalent memory T cell populations specific for a particular antigen in the human blood  and were recently expanded in vitro and found to lyse influenza-infected cells . However, little is known about γδ T cell effector functions specific for influenza in humans.
Although the antigens that activate most γδ T cell populations remain unknown, the antigens that activate the Vγ9Vδ2 TCR have been well characterized. The molecules identified are phosphoantigens including prenyl pyrophosphate antigens, which rapidly induce signal transduction pathways associated with the TCR [7; 8]. Other antigens, such as alkylamines, have been shown to act indirectly via the accumulation of isopentenyl pyrophosphate (IPP), a metabolite in the mevalonate pathway [9; 10; 11]. Although a presenting molecule has been hypothesized, Vγ9Vδ2 T cells are not HLA restricted. Several groups have reported increases in γδ T cells in the PBMC after viral infections such as HIV and EBV, and have isolated γδ T cell clones that recognize and kill virus-infected APCs [12; 13; 14].
A role for γδ T cells in recovery from influenza infection has been described in the mouse . Specifically the number of γδ T cells resident in the lung increased in the bronchoalveolar lavage (BAL) fluid after influenza A virus infection resolved (days 10-13) . On day 10 post-infection there was an expansion of the resident Vγ4 population, while Vγ2 cells took over after day 13 . In addition studies have been performed to evaluate γδ T cell memory responses to influenza. Mice were initially infected with A/HKx31 (H3N2 subtype with many of the internal components of A/PR/8/34) influenza virus and were then given a H1N1 (A/PR/8/34) secondary influenza A infection one month later. γδ T cells were detected in the BAL by day 5, exhibiting a more rapid response than after challenge with a divergent influenza B virus . This ability to recognize influenza viruses of differing subtypes was also described in work on influenza virus X-31 (H3N2)-immune TCRβ-/- mouse T cell hybridomas. They were shown to produce IL-2 in response to cells infected with influenza A or B viruses, but not cells infected with other viruses including vaccinia virus. Interestingly, no responses could be detected to any of the influenza A virus proteins expressed by recombinant vaccinia virus suggesting the γδ T cells were not responding to a viral protein . Recently, γδ T cells isolated from human PBMC were also expanded in culture and found to lyse influenza-infected macrophages but the mechanism of action is not understood .
Inducing cross-reactive immunity to influenza viruses of various subtypes is the ultimate goal of influenza vaccine development. Here we examine whether Vγ9Vδ2 T cells are able to elicit rapid cross-reactive responses to influenza virus infection without prior stimulation and culture in vitro. We monitor the activation and IFN-γ production by γδ T cells following exposure to virus and investigate the mechanism of activation. Our results indicate that γδ T cells in the PBMC can be rapidly activated by influenza virus infection in the absence of expansion in culture and that this response is dependent on the infectivity of the virus and the mevalonate pathway. The responding γδ T cell population expresses the Vδ2 chain of the γδ TCR, unlike previous reports indicating Vδ1-expressing cells primarily respond to viruses. In addition, the γδ T cells exhibit cross-reactive responses across H1N1, H2N2, and H3N2 virus subtypes. These cells may provide a first line of defense in the initial immune response to a virus infection due to their ability to rapidly produce antiviral cytokines such as IFN-γ.
Influenza A viruses, A/Puerto Rico/8/34 (H1N1) and A/Japan/305/57 (H2N2) were kindly provided from the Division of Virology (Bureau of Biologics, FDA, Bethesda, MD). Influenza A viruses of more recent origin, A/Johannesberg/33/94 (H3N2) and B/Harbin/97 (Influenza type B) were obtained from Pasteur Merieux Connaught (Toronto, Ontario). Experiments with A/Japan/305/57 were performed while the virus was still classified as a BSL-2 agent. Influenza A viruses were propagated in 10-day-old, embryonated chicken eggs. Infected allantoic fluids were harvested 2 days after infection with virus, aliquoted, and stored at -80°C until use. For heat-inactivation of influenza, virus was incubated at 95°C for 15 minutes. Vaccinia virus was propagated and titrated in CV-1 cells as previously described .
PBMC specimens were obtained from normal, healthy, vaccinia virus-immune adult American donors. PBMC were purified by Ficoll-Hypaque (Pharmacia, Piscataway, NJ) density gradient centrifugation (18). Cells were resuspended at 2 × 107/ml in RPMI 1640 with 20% FBS (Sigma) and 10% dimethyl sulfoxide and cryopreserved until use.
All antibodies were obtained from Becton-Dickinson, except anti-CD8β (AbD Serotec). Briefly cells were washed and stained with fluorochrome-conjugated antibodies for 30 minutes at 4°C. Cells were washed, fixed, and analyzed on a flow cytometer. Analysis was performed using WinMDI or FlowJo software.
2-4 × 106 PBMC were washed two times and exposed for four hours to A/PR/8/34 (H1N1) influenza A virus at an multiplicity of infection (moi) of 5 in 0.5-1 ml RPMI 1640 with 10% FBS at 37°C. In some cases mevastatin was added 2-48 hours prior to infection at a concentration of 10μM. After four hours, 5-10μg/ml Brefeldin A (Sigma) was added to each sample and incubated for four hours at 37°C. The cells were fixed/permeablized with Cytofix/Cytoperm kit (B-D) for 20 minutes at 4°C. Cells were stained with fluorescent antibodies and analyzed on a flow cytometer.
Upon activation, γδ T cells secrete cytokines, become cytolytic, and home to the site of infection . Recently, it has been shown that γδ T cells from the PBMC can be cultured in vitro with stimulating phosphoantigen and develop cytotoxic responses to influenza infected cells . In order to determine whether γδ T cells in human PBMC are able to rapidly respond to influenza virus infection without prior stimulation, we infected PBMC ex vivo and examined the ability of γδ T cells to become activated. PBMC were infected in vitro with influenza A A/PR/8/34 (H1N1) for 8-10 hours and surface expression of CD25 and CD69 was examined on γδ T cells by flow cytometry. Following exposure to influenza A virus, CD25 expression was upregulated with 24-33% of γδ T cells expressing this early activation marker (Fig. 1A,B). Influenza infection induced a corresponding increase in CD69 expression by γδ T cells. In some donors γδ T cells in the PBMC already expressed elevated levels of CD69 prior to infection perhaps suggesting the γδ T cells had been recently activated. In donors 1 and 2 nearly all γδ+ T cells expressed high levels of CD69 after influenza exposure in vitro (Fig. 1A,B). γδ T cell activation in response to influenza was observed in all six donors examined. These results indicate that prior stimulation with IPP is not necessary for γδ T cells to become activated by influenza infection.
We next examined whether this rapid activation of γδ T cells was cross-reactive between different strains and subtypes of influenza. Since the virus strain used in previous experiments, A/PR/8/34, is an H1N1 subtype, we tested H2N2 and H3N2 subtypes of influenza for the ability to activate γδ T cells. We also examined an influenza type B virus, which is very different from the influenza type A viruses. We found that both an H2N2 influenza A virus (A/Japan/305/57) and an H3N2 subtype of influenza A virus (A/Johan/97) were able to stimulate γδ T cell upregulation of CD25 (Fig. 2). In contrast, fewer γδ T cells responded to infection with the influenza B virus (B/Harbin/97) (Fig. 2). These results indicate that human γδ T cells exhibit heterotypic responses to influenza virus.
To determine whether other viruses would similarly activate human γδ T cells, we infected PBMC with vaccinia virus as opposed to influenza virus. There was no increase in CD25+γδ+ cells after exposure to vaccinia virus (Fig. 2). In fact in both donors there were fewer γδ T cells expressing CD25 after vaccinia virus infection than in the uninfected samples. These results indicate that influenza virus provokes a specialized response from γδ T cells.
To further analyze the subset of γδ T cells that respond to influenza infection, we identified the responding population TCR usage and distinguished between CD8+ and CD8- γδ T cells. The majority of γδ T cells in the blood express Vδ2, so we identified whether Vδ2+ T cells were the responding population. CD3+Vδ2+ T cells upregulated CD69 upon influenza infection (Fig. 3A). Interestingly, the Vδ2 TCR was slightly downregulated by the responding population suggesting activation via the TCR (Fig. 3A). Depending on the donor there were 1.5 to 3-fold more CD8- γδ T cells than CD8+ γδ T cells in human PBMC (Fig. 3B). Next we identified whether the CD8+ γδ T cells were CD8αα similar to the intestine or CD8αβ similar to αβ T cells in the periphery. The majority of the CD8+ γδ T cells were CD8αβ, while a minor population of CD8αα were also present (Fig. 3B). We detected increased expression of CD69 in response to influenza infection in both the CD8+ and CD8- populations of γδ T cells (data not shown).
A hallmark of T cell activation during virus infection is the production of IFN-γ which has a role in degrading antigen, regulating antigen presentation, and initiating proliferation and differentiation of lymphocytes [20; 21; 22]. To examine whether γδ T cells rapidly produce IFN-γ in response to influenza infection, we added influenza virus to PBMC and stained for IFN-γ production by intracellular cytokine staining. 8-10 hours post exposure to influenza virus γδ T cells produce IFN-γ (Fig. 4). This is similar to memory αβ T cells, which secrete IFN-γ 6-8 hours post infection . Indeed in all donors tested a small population of memory αβ T cells produced IFN-γ in response to influenza infection, suggesting that these adult donors have had previous exposure to influenza virus. IFN-γ production by γδ T cells in response to influenza was observed in all 6 donors examined, but to varying degrees (Fig. 4 and data not shown). Next we determined whether CD8+ or CD8- γδ T cells were producing IFN-γ. In all donors both the CD8+ and CD8- γδ T cell populations became activated to produce IFN-γ (Fig. 4, middle panels). In order to determine whether infectious virus is necessary for γδ T cell activation, H1N1 influenza A virus was heat-inactivated prior to addition. Heat inactivation greatly reduced the number of IFN-γ-producing γδ T cells in the PBMC (Fig. 4, bottom panels). These results indicate that active infection is necessary for influenza-specific responses by γδ T cells.
The majority of studies on human and murine γδ T cells have shown that they are non-MHC restricted . For example, Vγ9Vδ2 T cells from the PBMC in humans recognize and expand to IPP, which is an intermediate in the mevalonate pathway. We blocked the mevalonate pathway with the HMG-CoA reductase inhibitor, mevastatin. The inhibitor was added at various time-points prior to infection with influenza. Addition of mevastatin for 11 to 48 hours prior to infection inhibited γδ T cell activation (Fig. 5). These results identify a role for the mevalonate pathway in the rapid activation of Vγ9Vδ2 cells by influenza infection and suggest that influenza infection upregulates an isoprenoid in the mevalonate pathway that activates γδ T cells. Interestingly, both CD8+ and CD8- γδ T cells were very sensitive to mevastatin inhibition (Fig. 5). This suggests that although the Vγ9Vδ2 T cells exhibit differential expression of CD8, there appear to be conserved activation requirements.
Although the antibacterial and tumoricidal roles for Vγ9Vδ2 cells have been well studied, their role during viral infections is still not well understood. The semi-activated nature of γδ T cells suggests that they are rapid responders, perhaps similar to memory αβ CD8+ T cells. Indeed Vγ9Vδ2 T cells in human PBMC are able to quickly upregulate activation markers and secrete IFN-γ in response to influenza A virus infection. Although there were differences between donors in the extent of basal γδ T cell activation marker expression, there were clear increases in activated γδ T cells in the influenza virus-exposed PBMC of all six donors tested. This suggests that the γδ T cell response may be an additional early arm of the cell-mediated immune response to influenza that can be exploited for vaccination purposes.
The mechanism of Vγ9Vδ2 T cell activation by influenza infection is dependent on the infectivity of the virus and function of the mevalonate pathway. It was recently suggested that the antiviral responses of γδ T cells could be induced by treating patients with phosphoantigens . Importantly, our findings show that this is the mechanism by which γδ T cells naturally respond to influenza infection, since activation of Vγ9Vδ2 cells did not require previous stimulation with antigens in vitro. We postulate that influenza infection of an APC such as a dendritic cell or macrophage leads to the accumulation of IPP, which is an intermediate in isoprenoid synthesis and a Vγ9Vδ2 T cell antigen. Mevastatin blocks the mevalonate pathway by inhibiting HMG-CoA reductase, which reduces IPP levels. Similar mechanisms have been found for the activation of Vγ9Vδ2 cells by bisphosphonates and alkylamines [25; 26; 27]. In these studies the activation of Vγ9Vδ2 was also found to be indirect as it could be blocked by mevastatin [25; 26; 27]. It appears that the accumulation of IPP is a mechanism by which a stressed, transformed or infected cell can alert γδ T cells to respond.
It appears that not all viruses alter the mevalonate pathway since vaccinia virus infection did not induce Vγ9Vδ2 T cell activation. However, other studies have reported that vaccinia virus can inhibit Vγ9Vδ2 responses to phosphoantigens , suggesting an alternative possibility that the γδ T cells in our study were inhibited by vaccinia virus. This differs from the mouse in which γδ T cells exhibit cytotoxic responses to vaccinia virus infected cells and high levels of IFN-γ secretion . An interesting difference between our study and the previous studies examining Vγ9Vδ2 responses to vaccinia virus is that we assess the ability of the T cell to become activated immediately upon virus stimulation without prior expansion with IPP. Influenza virus may be unique in its ability to induce such a rapid cytokine response by γδ T cells.
The conserved mechanism of activation by influenza infection provides a rationale for the cross-reactive response by Vγ9Vδ2 T cells to divergent influenza strains and subtypes. A similar heterotypic response was recently reported in which Vγ9Vδ2 cells had been expanded in vitro by IPP stimulation and examined for cytotoxic responses to macrophages infected with influenza viruses . In this study, H5N1 and H9N2 influenza virus-infected cells were lysed by the cultured Vγ9Vδ2 cell lines. Similarly, in the murine system, Ponniah et al. observed that mouse γδ T cell hybridomas secreted IL-2 in response to X-31 (H3N2), A/PR/8/34 (H1N1), and B/HK/8/73 (influenza B) viruses . The cross-reactive recognition of influenza virus strains and subtypes by γδ T cells may help promote recovery from infection by newly emerging influenza virus strains.
Most γδ T cells are CD4-CD8-, but there are CD4+ and CD8+ subsets. Here we identified both CD8+ and CD8- populations of Vδ2+ cells that could respond to influenza. Functions for CD8+ γδ T cells have been identified in other models of infectious disease. For example, the PBMC of HSV-1-immune donors were stimulated with autologous HSV-1-infected cells, and lysis of infected targets was found to at least partly be due to CD8+ γδ T cells . Further studies are necessary to discriminate between these two populations of γδ T cells in influenza infection.
Attaining heterotypic immunity to influenza virus is the goal of many vaccination research programs, however it has proved a difficult task. Induction of αβ memory T cells promotes clearance of antigenically distinct strains and subtypes of virus, however identifying peptide antigens that would be effective in vaccinees of differing HLA types is complex. It would be beneficial to have γδ T cells with heterotypic immunity, since Vγ9Vδ2 T cells can respond to the same antigen in different patients. Once activated γδ T cells may help control virus infection by homing to the inflamed lung, secreting IFN-γ, killing virus infected cells, and regulating inflammation. Our results support the hypothesis that γδ T cells can provide heterotypic immunity to influenza A virus and may be exploited to promote a more rapid recovery from viral infection.
This work was supported by the National Institutes of Health (NIH) grant DK-080048 (J.J.) from the Institute of Diabetes and Digestive and Kidney Diseases and NIH grant U19 AI-057319 (M.T. and F.E.) from the Institute of Allergy and Infectious Disease.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.