Hallmarks of CD4 T cell immunity against influenza

Abstract

The mechanisms responsible for heterosubtypic immunity to influenza virus are not well understood but might hold the key for new vaccine strategies capable of providing lasting protection against both seasonal and pandemic strains. Memory CD4 T cells are capable of providing substantial protection against influenza both through direct effector mechanisms as well as indirectly through regulatory and helper functions. Here, we discuss the broad impact of memory CD4 T cells on heterosubtypic immunity against influenza and the prospects of translating findings from animal models into improved human influenza vaccines.

Introduction

Influenza viruses pose a grave and unique threat to human health. Due to the high communicability of the virus, a broad host-range, and remarkable antigenic flexibility, influenza represents a ubiquitous yet constantly evolving pathogen that has plagued human populations for several hundred if not thousands of years [1, 2]. Current estimates suggest that influenza epidemics lead worldwide to millions of cases of severe illness and up to 500,000 deaths annually [3]. The global significance of influenza is evidenced by the World Health Organization's establishment of the Global Influenza Surveillance Network (GISN) in 1952 and an international vaccine development program on a scale not comparable to any other infectious disease. The GISN carries out constant characterization of influenza isolates that lead to recommendations for the formulation of annual vaccines designed primarily to produce neutralizing antibodies aimed against predicted circulating strains [4].

The potential for the sudden emergence of pandemic influenza strains represents an incessant threat on an even larger scale. For example, the 1918 ‘Spanish flu’ pandemic has been estimated to have caused 40-50 million deaths worldwide [5]. A major limitation of current annual vaccines is that antibodies generated against circulating influenza strains most often do not effectively neutralize emergent pandemic strains. Given the 6-8 month timeline required for the large-scale production of seasonal vaccines, the feasibility of vaccination strategies designed specifically against a pandemic strain presents several challenges [6]. A vaccination strategy providing broad protection against diverse influenza isolates could eliminate the need for annual vaccines and provide protection against emergent pandemic strains. We suggest that one effective approach towards this goal could involve vaccines that generate strong memory T cells against multiple internal core proteins, in addition to antibody.

Although heterosubtypic immunity has been extensively studied in animal models, a full understanding of the mechanisms by which such protection works is not yet clear. This undoubtedly reflects the complex interplay of multiple cellular populations during successful heterosubtypic responses, as well as difficulty in synthesizing data from many varied experimental systems. For example, using different combinations of virus to prime and challenge, changing the interval between priming and subsequent heterosubtypic challenge, altering the doses of virus, and host genetic factors can all have a substantial impact on the strength and duration of heterosubtypic protection in mouse models. Additionally, individual effector mechanisms have been shown to play either major or minor roles in different experimental models, suggesting several distinct forms of heterosubtypic protection may exist. Finally, translating lessons learned from animal models to the clinic has been discouraged by the finding that the potency of heterosubtypic immunity in human population seems limited, but the reason for this is unclear.

Here we discuss the potential for memory CD4 T cells to contribute to heterosubtypic immunity against influenza. In the human, basic correlates of protection in the humoral immune response against influenza have proven relatively easy to define while our understanding of protective mechanisms employed by T cells have accrued more slowly. The mouse model of influenza has proven a powerful tool for dissecting how memory T cells contribute to protective antiviral responses. It has become clear over 45 years of experimentation in the mouse that optimal heterosubyptic immunity requires a dynamic, multifaceted immune response and that influenza-specific memory CD4 T cells can act as helper cells in generating optimal B and T cell responses, as regulators of innate immunity, and as direct effectors of protection. Translating these observations into novel vaccine strategies could represent an important step toward providing a vaccine-induced long-term protection against seasonal and, more importantly, pandemic influenzas.

Influenza A virus and antigenic variation

Influenza A is an enveloped virus belonging to the family Orthomyxoviridae which causes a range of morbidity and mortality that is determined both by the subtype of virus and infected host. Infection typically results in an accute respiratory tract infection characterized in humans by a sudden onset of fever, myalgia, headache, and non-productive cough. All subtypes in humans and in animal models infect the respiratory epithelium from the nasal passages to bronchioles, however, more pathogenic viruses also tend to infect pneumocytes and intraalveaolar macrophages [7].

Influenza A viruses contain a genome composed of eight segments of negative-sense RNA coding for 11 proteins. The surface glycoproteins hemagglutinin (HA) and neuraminidase (NA) are highly variable and define the viral subtypes: there are currently 16 subtypes based on HA analysis and 9 based on NA [8]. Seasonal vaccine strategies target the HA and NA proteins of predicted circulating strains in order to generate neutralizing antibody responses. The ability of the virus to modify genes encoding HA and NA through mutation (antigenic drift) and through the replacement of these proteins with those of another subtype (antigenic shift) limit the timeframe of effectiveness for a vaccine targeting specific HA and NA combinations and leave immunized individuals at considerable risk in the face of a pandemic outbreak.

In marked contrast to the tremendous variability in HA and NA, highly conserved sequences in the viral PB1, PB2, PA, NP, and M1 proteins have been identified in comparisons of over 36,000 sequences [9]. It is likely that this disparity reflects at least in part more rigorous functional constraints on internal proteins, such as the viral polymerases. The internal and external viral proteins are also under different selection pressures within infected hosts: while external viral proteins are exposed to recognition by antibody, which will effectively select those that cannot be recognized, internal viral proteins are recognized by T cells only after their presentation on individual MHC/HLA molecules, when viruses have already established a foothold by infecting and replicating in epithelial cells. This dichotomy between immune recognition of external and internal viral proteins is reflected in the disctions between homotypic and heterosubtypic immunity to influenza.

Homotypic and heterosubtypic immunity against influenza

Homotypic immunity, the protection against influenza infection afforded by previous exposure to an influenza of the same serotype, was first described in the 1930's [10]. Homotypic protection is dependent on the generation of circulating neutralizing antibodies, and thus, could be passively transferred to naïve animals via convalescent serum from mice primed with the same influenza strain [11, 12]. Gerhard's laboratory characterized the critical components of homotypic immunity as IgG antibodies directed primarily against the viral HA, and showed that transfer of monoclonal HA-specific antibodies provided a strong degree of homotypic immunity even in SCID hosts that otherwise lack an adaptive immune system [13]. While an initial virus-specific IgM antibody is generated after influenza infection in the absence of CD4 T cell help, virtually no virus-specific IgG antibody-secreting B cells develop, and what antibody is seen is short-lived [14, 15]. Thus, CD4 T cell help is critical for the generation of long-term homotypic immunity to influenza.

Heterosubtypic immunity, the protection against severe disease caused by previous infection with an influenza virus of a different serotype, was first described in 1965 [16]. While heterosubtypic protection could not be transferred from immune animals to naïve hosts via serum, a substantial decrease in viral titer was shown after transfer of cytotoxic T cells obtained from the spleens of immune mice to naïve mice that were then challenged with a virus that expressed a different HA and NA [17]. Recognition of conserved T cell epitopes, almost exclusively derived from internal viral proteins, and presented by MHC molecules, underlies heterosubtypic immunity. In contrast to homotypic immunity, heterosubtypic protection does not block initial infection, resulting in relatively high viral titers in the lung during the first few days after challenge, and typically some signs of disease, such as weight-loss [18, 19].

Cellular requirements for heterosubtypic immunity to influenza

It is commonly thought that a strong virus-specific memory CD8 T cell compartment is critical for heterosubtypic immunity in intact virus-primed individuals [18-21]. Heterosubtypic immunity has been correlated with enhanced CD8 T cell responses in model systems involving vaccination regimes with live virus [22-25], inactivated virus [26], or with strategies encoding cross-protective viral proteins [27-30]. Furthermore, adoptive transfer of various CD8 T cell populations specific for influenza to otherwise naïve animals can provide protection against lethal infection [31-34]. Similar studies in TCR transgenic animals expressing only CD8 T cells specific for influenza found that CD8 T cells provide strong protection in the absence of CD4 T cells and B cells. However, the virus-specific CTL response can also cause substantial pathology [35].

Surprisingly, a critical role for B cell responses in mediating heterosubtypic immunity has also been shown [36, 37]. It is likely that virus-specific memory B cells enhance heterosubtypic protection primarily through the generation of cross-reactive antibodies [38, 39]. Contributions from antibodies directed against non-neutralizing internal proteins, such as NP [40], as well as neutralizing antibodies directed against conserved sequences of HA [41, 42] have been described. A number of studies also support an important contribution from virus-specific CD4 T cells in mediating heterosubtypic protection [19, 24, 28, 43-45]. In fact, a strong degree of cross-protection has been observed in influenza-primed mice lacking CD8 T cells entirely [36, 37, 43, 46].

That memory CD8 T cells, B cells, and CD4 T cells have all been shown to mediate protection individually suggests that heterosubtypic immunity is compsed of multiple components which may synergize to provide the most robust protection. Thus, primed individuals might differ in the relative strength of discrete protective mechanisms and reliance on particular cellular subsets (Figure 1). In the human population, such differences could be influenced through multiple variables. For example, the length of time between exposure to influenza viruses could have a dramatic impact on the relative strength of CD8 versus CD4 T cell immunity as numerous studies have found that memory CD4 T cells contract more rapidly than memory CD8 T cells [47]. Differences between individuals in terms of antigen-specific clonal diversity [48] may also impact the ability of memory CD4 and CD8 T cells to optimally respond to heterosubtypic challenge. Furthermore, infection with other pathogens can alter the repertoire of T cells recognizing influenza through stimulation of virus-specific memory T cells with cross-reactive epitopes [49], with important consequences for protective immunity as well as immunopathology [50]. Finally, age has a significant and broadly negative impact on immune function, and particularly compromises the effectiveness of influenza vaccination [51].

Figure 1.

Following influenza challenge or vaccination, the magnitude of individual components of the adaptive anti-viral immune response differentially peak and contract (a). Within a given population, several factors might impact the overall strength of heterosubtypic immunity, and the relative magnitude of individual components thereof, resulting in both protective and non-protective states characterized by distinct contributions from virus-specific CD4 T cells, CD8 T cells, and B cells (b).

How memory CD4 T cells contribute to heterosubtypic immunity is not fully understood. Recent studies have shown that memory CD4 T cells act not only as helper cells for the development of optimal B cell and CD8 T cell responses, but also as regulators of innate immunity and as potent effector cells. Thus, the potential for memory CD4 T cells to improve protective immune responses may not be restricted to one particular time-point or function. Below, we discuss how influenza-specific memory CD4 T cells can contribute both at early and later stages of infection to protective heterosubtypic responses (Figure 2).

Figure 2.

Memory CD4 T cell impact multiple stages of the anti-viral response. Lung-resident memory CD4 T cells (blue) recognizing influenza up-regulate innate immune responses within 48 hours of infection, preceding memory CD4 T cell division. Helper activities mediated by memory CD4 T cells are evident in secondary lymphoid organs by 5 days post-infection, during which time memory CD4 T cells divide rapidly and begin to expand as a population. Finally, memory CD4 T cells develop into large and potent effector populations that migrate to the lung at later timepoints and directly control virus through perforin-mediated cytotoxicity and other mechanisms.

Memory CD4 T cells regulate early innate responses upon influenza infection

One consequence of influenza infection is a profound inflammatory response in the lung involving many innate cell populations as well as the lung epithelium. Influenza stimulates several pathogen-associated molecular pattern (PAMP) receptors directly [52-54], as well as triggers activation of inflammasomes [55, 56]. We have recently demonstrated that virus-specific memory CD4 T cells can also directly regulate innate inflammation within 48 hours following heterosubtypic challenge, independently of PAMP recognition [57]. Transfer of virus-specific memory CD4 T cells to otherwise naïve animals before influenza infection dramatically upregulates pulmonary and systemic levels of several innate cytokines and chemokines including IFNγ, IL-12, IL-1, IL-6, CXCL9, CXCL-10, and CCL2 [57], as well as type I interferons (unpublished observations). We believe that the enhanced inflammatory response mediated by memory CD4 T cells upon influenza challenge is an important component of heterosubtypic immunity for several reasons (Figure 3).

Figure 3.

Memory CD4 T cells recognizing influenza antigen presented by APC drive upregulation of costimulatory molecules and inflammatory cytokines and chemokines independently of PAMP signals. Activated APC can migrate to draining lymph nodes to more quickly initiate adaptive responses against the virus. Inflmmatory mediators produced by the initial interaction between memory CD4 T cells and APC can also have important consequences. First, chemokine gradients that are established can call in diverse elements of the innate immune system and also memory CD8 T cells into the infected lung. Second, the enhanced inflammatory response can act on lung epithelial cells to drive further inflammation, creating an antiviral state that controls virus during the initial days of infection. However, it is likely that some elements of the enhanced inflammatory response also cause a degree of immunopathology.

First, enhanced inflammation mediated through memory CD4 T cell regulation of innate immunity early after influenza infection correlates with a significant level of control on viral titers 3 and 4 days post-infection [57]. While recognition of antigen by CD4 T cells is required for regulation of inflammation, viral control does not require T cell recognition. This suggests that elements of the enhanced inflammatory response induced through memory CD4 T cell recognition of antigen have direct anti-viral properties. This hypothesis is consistent with studies correlating enhanced pulmonary inflammatory responses and protection against influenza [58-60]. These results might also explain findings of a strong degree of protection mediated by CD4 T cells in mice not expressing MHC II on lung epithelial cells [61], as we observed induction of early inflammation and viral control in mice only expressing MHC II on CD11c⁺ cells [57]. Further supporting a beneficial role of memory CD4 T cell-induced inflammation, we observed enhanced inflammatory responses upon transfer of Th1- and Th17-polarized memory CD4 T cell populations specific for influenza, both of which can protect otherwise naïve hosts against a high dose challenge of virus, but not after transfer of Th2- or Th0-polarized memory cells, neither of which affords protection [57]. We suggest that a major consequence of the early capping of viral titers by memory CD4 T cell-induced inflammation is the ‘buying of time’ for the development of subsequent effective innate and adaptive responses against the virus by reducing the period of ‘stealth’ influenza replication that precedes initiation of adaptive responses in the naïve state [62].

In addition to buying time by restricting early replication, several of the elements of the enhanced inflammatory response induced by CD4 T cells may play important roles in actively driving the earlier development of T cell and B cell responses. This might be of special importance in aged individuals, as the proinflammatory cytokines IL-1, IL-6, and TNF can help to overcome age-related defects in T cell function [63]. Additionally, we observed that virus-specific memory, but not naïve, CD4 T cells recognizing antigen on CD11c⁺ cells in the lung caused dramatic up-regulation of MHC II as well as costimulatory molecules on dendritic cells and alveolar macrophages as early as 2 days after infection [57]. Thus, the pro-inflammatory properties of memory CD4 T cells after infection combined with early activation of APC populations might act to ‘jumpstart’ adaptive immune responses during heterosubtypic challenge.

Beyond influencing the earlier generation of immune responses against novel epitopes present in heterosubtypic viruses, the enhanced inflammatory response induced by memory CD4 T cells could have a substantial impact on CD8 T cell responses directed at conserved viral epitopes. For example, it has been shown in a model of herpes simplex virus infection, that CD4 T cells control the influx of CTL into infected tissues through the induction of local chemokine gradients [64]. Also, a recent study demonstrated that type I interferons can directly enhance the cytolytic capacity of memory CD8 T cells [65].

The results of our study are supported by recent findings of Seo et al., demonstrating a critical role for PAMP signaling in driving optimal primary but not secondary homotypic or heterosubtypic responses against influenza [66]. We suggest that upon re-infection, memory CD4 T cells specific for influenza, that were primed and polarized partly through signals delivered by the initial inflammatory milieu established through PAMP recognition during a primary response, act to increase the tempo and magnitude of a similar protective inflammatory response. In a sense, memory CD4 T cells can remember the inflammatory environment that they were exposed to during their generation and can play an important role in rapidly re-establishing a similar inflammatory setting upon reencounter with their cognate antigen.

Memory CD4 T cells as B cell helpers during heterosubtypic responses

CD4 T cells are necessary as helpers to generate protective heterosubtypic B cell antibody responses, as severely impaired heterosubtypic protection and very low levels of virus-specific antibodies are observed in CD4 T cell-deficient mice [36]. Recent studies from our laboratory have shown that the development of strong and long-lived IgG and IgA responses against influenza require SLAM-associated protein (SAP)-expressing CD4 T cells [15], implicating the involvement of so-called follicular helper cells [67]. Thus, CD4 T cell help during a primary response against virus is critical to facilitate the generation of sufficiently diverse cross-reactive antibodies of this kind that have been found to play important roles in heterosubtypic responses [38]. But do memory CD4 T cells impact B cell responses during heterosubtypic challenge? We have observed enhanced production of virus-specific IgG antibody upon transfer of memory CD4 T cells to otherwise naïve mice then infected with flu, correlating with protection (unpublished observations). These results are similar to studies from Gerhard's group finding that transfer of CD4 T cell clones specific for infecting virus led to protection of nude hosts through an antibody-dependent mechanism [68]. However, we and others have not found a crucial role for memory CD4 T cell B cell helper activity in mounting successful heterosubtypic immunity in intact primed mice [39]. This is perhaps because strong heterosubtypic T cell responses are capable of clearing virus by day 7 post-infection [19], before substantial antibody responses against determinants specific to the heterosubtypic virus can be raised. However, it may be possible that as this antibody-independent form of heterosubtypic protection wanes with time, or in cases in which relatively weak heterosubtypic CD8 T cell responses are generated, the role of memory CD4 T cell help for B cell responses may become more prominent. This remains to be studied but may have particular relevance for human influenza, since it would enhance generation of effective neutralizing antibody titers that can protect until new strains emerge.

Memory CD4 T cells as regulators of heterosubtypic CD8 T cells

CD4 T cells have been shown to act as helpers, via multiple mechanisms, in driving optimal CD8 T cell responses, both in terms of the magnitude of the response, and the function of CD8 T cell effectors [69]. A role for CD4 T cell helpers in quantitatively enhancing CD8 T cell responses has also been documented in mouse models of primary influenza infection [70-72]. With regards to heterosubtypic protection, the extent of the initial expansion of CD8 T cell effectors during a primary response is likely an important factor in establishing the magnitude of the memory CD8 T cell pool. Indeed, it has been shown that CD4 T cells responding during a primary influenza challenge have a dramatic role in actively supporting the generation of long-lived functional memory CD8 T cells [71-74].

Again, as is the case for B cell help during heterosubtypic responses, whether or not memory CD4 T cell enhancement of CD8 T cell responses can be an important component of heterosubtypic protection has not been carefully addressed. As already discussed, it is possible that memory CD4 T cells could enhance CD8 T cell responses through more efficient maturation of APC early after infection [57]. Indeed, it has been shown with human memory CD8 T cells specific for influenza that optimal responses, in terms of both proliferation and cytokine production, are elicited by highly activated dendritic cells [75]. It is also likely that elements of the inflammatory cytokine response induced by memory CD4 T cells, as well as production of IL-2 by memory CD4 T cells, could accelerate the response kinetics of CD8 T cells and perhaps improve trafficking of CD8 T cell effectors to critical sites within the infected lung. As for B cell helper functions, it is quite possible that the importance of such mechanisms rises as the relative strength of heterosubtypic immunity wanes.

Memory CD4 T cells as anti-viral effectors during heterosubtypic responses

After the induction of innate responses and provision of help, memory CD4 T cells re-expand to give rise to effector populations that migrate in large numbers to the infected lung. Early experiments established that CD4 T cell clones specific for influenza could provide a strong degree of protection when transferred to otherwise naïve hosts, and that the protection correlated with cytotoxic activity [76-79]. More recently, our laboratory has shown that in vitro-generated Th1-polarized CD4 T cell effectors specific for influenza can protect against otherwise lethal infection when transferred to B cell-deficient hosts [80]. Protection in this model is dependent on perforin-mediated cytolytic function. Interestingly, we also found that Th17-polarized CD4 T cell effectors deficient for perforin could protect B cell-deficient mice [81], suggesting that CD4 T cells are capable of multiple and distinct helper-independent mechanisms of protection to combat influenza. Additional pathways by which CD4 T cell effectors could clear virally infected cells include FAS- and TRAIL-mediated killing, both of which have been demonstrated as potent anti-viral mechanisms utilized by CD8 T cell effectors during influenza infection [82-84]. Conversely, similar to strong CD8 T cell responses, CD4 T cell effectors themselves can cause a substantial degree of pulmonary immunopathology during influenza challenge [85].

As it is well established that primed but not naïve CD4 T cell responses contribute to protection against influenza [86], the results discussed above suggest that protection afforded by effector CD4 T cell mechanisms plays a more important role in heterosubtypic protection than during primary infection. We have investigated this possibility by transferring protective memory CD4 T cells to mice and infecting with influenza, followed by in vivo depletion of the donor memory cells on day 5 post-infection, after their provision of helper function (as determined by detecting increased levels of virus-specific antibody compared to mice given naïve CD4 T cells of the same specificity). We observed that depletion resulted in a loss of protection even though it did not decrease the memory cell-dependent helped antibody response, suggesting that helper-independent effector functions of memory CD4 T cells contribute significantly to protection, even in intact mice. In further support of the hypothesis that effector functions of CD4 T cells play an important role during heterosubtypic responses, we find that secondary effectors, derived from memory CD4 T cells, are superior to primary CD4 T cell effectors in terms of multifunctional capabilities and in providing protection upon adoptive transfer (unpublished observations).

Complexity of CD4 T cell responses during heterosubtypic challenge

It is increasingly clear that protective immune responses against influenza involve multiple cellular populations and numerous independent and overlapping mechanisms of protection. The complexity of the CD4 T cell response during heterosubtypic challenge is difficult to categorize in terms of individual subsets based on surface marker expression, cytokine production, and activation status [87]. Additionally, we have observed striking differences in the phenotype and function [88], as well as in gene expression (Figure 4) of CD4 T cells responding to influenza in the lung versus secondary lymphoid organs. These findings might reflect the simultaneous development of subsets of memory cells specialized to act either as regulators, helpers, and effectors, or conversely, the heterogeneity may reflect cells along different points in a linear differentiation program. For example, whether memory CD4 T cells capable of helper functions can also act as cytotoxic effectors and visa versa is currently not known.

Figure 4.

Memory CD4 T cells responding against influenza are heterogeneous. Upon adoptive transfer to otherwise naïve hosts then infected with virus, influenza-specific memory CD4 T cells adopt distinct phenotypes, functions, and gene-expression profiles in the spleen, draining lymph nodes, and lung. Such heterogeneity makes identification of correlates of protection difficult, and raises the question of what can be learned from assessing CD4 T cells from only one source, such as the blood.

It is also clear that discrete mechanisms of protection employed by memory CD4 T cells play more or less importance depending on the context of re-infection (for example, depending on the presence of influenza-specific B cells or CD8 T cells). While such ablation of lympocyte populations is not likely to occur in intact mice, it is almost certain that the relative strength of virus-specific B cell, CD8 T cell and CD4 T cell memory can vary within experimental animals, and vary as time from priming increases within the same animal. This raises the possibility, as already discussed, that several forms of heterosubtypic immunity exist that rely more or less heavily on contributions from different lymphocyte subsets, or even on the individual functions brought to bear by memory CD4 T cells.

Finally, though CD4 T cells produce an array of cytokines during anti-influenza responses, including IFNγ, IL-2, IL-10, IL-17, and TNF, the role of these cytokines in directing protective immunity and/or pathology is often unclear. For example, numerous studies aimed at elucidating the role of IFNγ during influenza infection have yielded conflicting results [76, 89-92]. Similarly, the role of IL-10 production by CD4 T cells during influenza challenge has been found to be a negative [81, 93] and a postive factor [94]. Again, the design of model systems employed in these studies may have a substantial impact on conclusions drawn. For example, the finding that IFNγ contributes to resistance against flu in mice deficient for nitric oxide synthase 2 [95], but not WT mice, raises the possibility that because of the many different antiviral mechanisms that contribute clearing the virus, only reductionist models that eliminate one or more of these pathways will reveal striking roles for individual components of the CD4 T cell response.

Heterosubtypic immunity in humans

Whether or not heterosubtypic immunity plays an important role in protection against human influenza has been a matter of controversy. Before the past few years only a few examples of demonstrable heterosubtypic protection in well-controlled human studies could be found [96, 97], but recent pandemic strain outbreaks have provided several others, demonstrating that seasonal vaccines can provide a degree of protection against an emergent pandemic strain [98-101], and especially against severe disease. Furthermore, retrospective analysis of the 1918 epidemic suggests a substantial impact of heterosubtypic immunity [102].

While the protective components of heterosubtypic protection in these studies was not determined, it is likely that memory CD4 T cells play a role as studies have demonstrated reactivity against pandemic H5N1 and H1N1 viruses by CD4 T cells primed with seasonal circulating virus and vaccines [103-107]. Translating from the mouse model and investigating in the clinic elements of the memory CD4 T cell response found to correlate with protection in reductionist systems is difficult. For example, the negative impact of aging on the ability of the immune system to mount successful responses has been well-documented and is of particular relevance to influenza [108]. Furthermore, while the antigenic experience of laboratory mice can be controlled, the individual histories of patients may dramatically impact the shape of and strength of heterosubtypic immunity.

Nevertheless, determining the impact of memory CD4 T cells during heterosubtypic responses in the human population should become increasingly approachable in the coming years. Identification of more HLA-restricted influenza peptides combined with improvements in MHC class II multimer-based assays and peptide-MHC cellular microarrays should improve the scope of functional testing of influenza-specific human memory CD4 T cells [109]. Identification of surface markers associated with the memory state and functional potential of T cells combined with improvements in multi-color flow cytometry should allow for increasingly rigorous investigation of the properties of vaccine- and influenza-primed memory CD4 T cells. Furthermore, while the enumeration of influenza-specific IFNγ-producing cells has been the hallmark of studies addressing T cell responses in humans, a new appreciation of the capacity of T cells to produce multiple cytokines [110] may prove more informative correlations between memory CD4 T cell cytokine production and protection against influenza. However, that T cells isolated from peripheral blood are most often studied for function analysis raises the possibility that this subset of responding cells may not reflect accurately how T cells are responding in the lung and draining lymph nodes (Figure 4).

Concluding remarks

In contrast to some infectious diseases for which definitive correlates and mechanisms of protection have been described, our understanding of protective immunity against influenza remains clouded. Memory CD4 T cells can make important and varied contributions during heterosubtypic influenza challenge. Translating lessons learned from the study of reductionist animal models to human influenza remains an important but difficult task. A better understanding of the mechanisms employed by memory CD4 T cells should be of benefit for the design of vaccines capable of providing a broad spectrum of protection against seasonal and pandemic influenza viruses.