Infectious diseases take a massive toll on the well being of both the elderly and the HIV-infected population [1,2]. The situation is particularly concerning with regard to vaccine-preventable diseases, with up to 1,000 times greater risk of death in older adults compared to vaccine-aged children [3]. Indeed, the elderly (>65 yrs) account for 90% of the >35,000 affected by annual influenza epidemics [4,5]. HIV-infected people are at a significantly higher risk than the general population at all ages for acquiring seasonal influenza infection, despite vaccination and virologic control with combination antiretroviral therapy (cART) [6-8]. Seasonal influenza vaccination is recommended for elderly as well as HIV-infected individuals to reduce influenza-related morbidity and mortality [9], but immune response to influenza vaccination is frequently impaired in both these high-risk populations [10-13]. With the considerable increase in life expectancy of HIV-infected persons with cART, and the increasing incidence of new HIV infections at older ages, it is estimated that by 2015, 50% of HIV-infected population will be ≥50 years of age [14]. Given the independent detrimental effects of aging and of HIV infection on the immune system [15-17], it is important to investigate the cumulative effects of HIV and aging on immunity, e.g. as assessed by responsiveness to seasonal influenza vaccines.

Although influenza vaccines elicit both cellular and humoral responses [18], immune protection is largely correlated with post-vaccination serum antibody (Ab) titers [19]. An essential step in the generation of vaccine induced Ab-secreting B cells is the interaction of antigen-primed B cells and T follicular helper cells (Tfh) in the germinal center reaction, where Tfh cells provide essential helper function for B cells to undergo proliferation, isotype switching and somatic hypermutation (reviewed in [20]). Tfh cells are characterized by surface expression of the CXC chemokine receptor 5 (CXCR5), that promotes their homing to lymphoid germinal centers [21], and by abundant production of the cytokine interleukin (IL)-21, that plays a major role in inducing B cell differentiation and proliferation [22] and in preserving plasma cells [23,24]. Recently, a CD4 T memory cell subset in peripheral blood bearing functional and partial phenotypic similarity to lymph node Tfh has been identified and has been designated as peripheral (p)Tfh [25-27]. The pTfh cells represent approximately 15% of circulating CD4 T cells in humans [27], express CXCR5 and provide critical help to B cells for antibody secretion in an IL-21-dependent manner [26]. In a study of vaccine responses to the pandemic H1N1/09 influenza vaccine in HIV infected young patients, impaired vaccine responses were associated with defective function of pTfh and in the IL-21/IL-21R system [27,28]. In physiologic aging, impaired Ab responses to seasonal influenza vaccination have been largely attributed to intrinsic B and T cell defects [29-33], but pTfh have not been investigated.

A common feature of both HIV/AIDS and aging is the associated inflammation and immune activation of varying degrees [34,35], which is nevertheless greater in HIV-infected individuals than in age-matched uninfected controls, as we recently demonstrated in a cohort of post-menopausal women [36]. Tumor necrosis factor (TNF)α, one of the first pro-inflammatory factors that was described in HIV infection [37], is elevated also in the elderly [38,39], and its increased levels have been associated with disease progression and treatment failure in HIV-infected patients [40-42], and with increased risk for atherosclerosis in the aging population [43]. Intriguingly, this cytokine has been reported to decrease both T and B cell responses in vitro [44-46]. Excessive immune activation is known to favor disease progression and to hinder CD4 T cell immune reconstitution in HIV-infected persons [47-49]. In this study we have demonstrated that immune activation and the proinflammatory cytokine TNFα can have a detrimental effect on influenza vaccine responses in aging HIV-infected post-menopausal women.

Influenza infection is associated with significant morbidity and mortality in the elderly and HIV-infected individuals, underscoring the importance of effective vaccination-induced protection for these populations. It is well established however that both these high risk groups can have impaired influenza vaccine responses [10,13,29-33]. Our present study in post-menopausal women implies that the burden of HIV infection on top of aging is likely to increase the risk for deficient influenza vaccine responses in comparison to HIV uninfected aging women. We describe the novel association of impaired vaccine response to ongoing immune activation. In particular, we demonstrate that an activated state of peripheral T follicular helper cells and decreased vaccine-induced plasma IL-21, the signature cytokine of Tfh, are associated with impaired influenza-vaccine induced antibodies. We provide novel experimental data implicating TNFα as a major contributor for these deficiencies.

For selecting a clear functional state of aging, we performed this study in post-menopausal women and thus avoided any influence of perimenopausal hormonal fluctuations on inflammation and immune activation. The study participants were recipients of the 2011-2012 seasonal influenza vaccine, which contained the same strains that had been included in the 2010-2011 vaccine composition [containing an A/California/7/2009 (H1N1)-like virus, an A/Perth/16/2009 (H3N2)-like virus and a B/Brisbane/60/2008-like virus]. In our study cohort, the pre-vaccination influenza Ab titers were in the protective range (≥1:40) in 58% of HIV^– and 31% of HIV⁺ aging women, presumably due to persisting or cross-reactive antibodies resulting from previous infection or vaccination. Although seroconversion rates were similar between HIV-infected and uninfected women, they differed in the response to vaccination. In the HIV⁺ women, antibody titers at wk4 and the fold increase in titers from t0 were lower, with several showing <4 fold increase in comparison to the HIV^– control group, pointing to less immune competence in the HIV-infected women.

To elucidate factors contributing to immune impairment, we investigated the relationship between Ab responses and cellular immune activation and inflammatory cytokines. Several studies have previously shown that activated CD8 T cells are associated with disease progression in HIV infection [48,49,57]. Interestingly, for the influenza vaccine-induced Ab response, it was CD4 T cell activation rather than CD8 T cell activation that showed this inverse relationship, suggesting that immune activation may have different outcomes that are specific for each cell type. Since among CD4 T cells, the pTfh subset promotes B cell proliferation and maturation [25-27], we investigated this subset and noted that the presence of activated pTfh cells prior to vaccination had a strong negative correlation with magnitude of post-vaccination Ab titers, both in HIV-uninfected and HIV-infected donors. This observation suggests that in an activated state, pTfh cells are less competent in providing the needed help to B cells to differentiate into Ab-producing cells, ultimately leading to impaired Ab response to influenza vaccine.

Both aging and HIV infection can lead to excessive inflammation. The inflammation that accompanies physiologic aging has disparate etiologies including, but not limited to, hormonal changes, diet, environmental and genetic causes. We have recently reported that in the post-menopause state several measures of immune activation and inflammation, notably T cell activation and exhaustion, monocyte/macrophage activation, and proinflammatory cytokines IL-6, IL-8 and TNFα, were higher in HIV⁺ women as compared to HIV-uninfected controls [36]. The drivers of immune activation in HIV disease in virally suppressed individuals are not well understood, but may include microbial translocation due to persistent virus-induced injury in gut-associated lymphoid tissue [58]. However, no association between markers of microbial translocation (circulating levels of LPS, monocyte/macrophage activation marker sCD14) and magnitude of antibody response to the influenza vaccine was observed in our study cohort (data not shown). Levels of plasma TNFα had a significant negative relationship with vaccine-specific Ab titers at week 4 post-vaccination in both HIV^– and HIV⁺ aging women. Prolonged TNFα exposure is known to impair T cell responses (namely TCR-induced proliferation and cytokine production) by attenuating T cell receptor signaling [44,45], and anti-TNF therapy in patients with psoriasis and inflammatory bowel disease results in improved responses to TCR stimulation ex vivo [59]. We investigated the influence of TNFα experimentally by adding it to cell cultures and noted that it could lead to upregulation of activation markers on CD4 T cells including pTfh cells. TNFα blockade in co-cultures of purified memory CD4 T cells and B cells enhanced Ab production supporting the contention that TNFα has a negative influence on Ab responses. Additionally, TNFα can potentially inhibit B cells directly as suggested by studies in mice [46] and by the observation that B cells of aging individuals can secrete TNFα with intrinsic inhibition of their function [60].

The helper function of Tfh and pTfh cells is mediated by cognate interaction with B cells and IL-21 secretion [25-27,55,61]. We previously demonstrated in young HIV⁺ adults who responded to influenza vaccine that their plasma IL-21 levels increased with upregulation of IL-21R in B cells, while vaccine non-responders failed to do either [28]. In the present study as well, we observed higher plasma IL-21 levels only in participants who were seroprotected at week 4, and frequency of IL-21R expression in CD4 T cells was also higher in these individuals. These findings expand the role of the IL-21/IL-21R system in regulating vaccine responses by activities not only in the B- but also in the CD4 T cell compartment. A trend of IL-21R upregulation in B cells was also observed, even though it did not reach statistical significance likely due to small sample size. Interestingly, IL-21R levels in CD8 T lymphocytes remained unaltered, implying that mechanisms other than IL-21 are involved in the generation of influenza-specific CD8 T cells. We have previously shown that the pTfh CD4 T cell subset resides almost exclusively in the memory CD4 T cell pool and that IL-21 production by pTfh is critical for influenza Ab response [27]. Here we demonstrated that addition of TNFα resulted in decreased IL-21 secretion in CD4 cell cultures, whereas IL-21 addition in our experimental model resulted in increased Ig production. As expected, IL-21 blockade resulted in suppressed Ig production based on the known critical role of IL-21 in supporting B cell function [22-24]. Taken together, our findings imply that excessive TNFα can impair pTfh function and IL-21 secretion, resulting in deficient help to B cells and poor Ab responses. Lack of clinical samples precluded us from doing experiments to investigate the mechanism of TNF alpha on adaptive immune responses in co-culture studies. We cannot exclude the possibility that in vivo, the decreases in IL-21 and antibody secretion iresulted from death of activated cells or a direct inhibitory effect of TNF-alpha on B cells.

Previous studies have implicated immune senescence for failure to respond to influenza vaccination [62,63]. Accumulation of senescent lymphocytes has also been reported for HIV-infected individuals [64,65], possibly as consequence of chronic antigenic stimulation. We observed that senescent CD4 and CD8 T cells in our cohort were in higher frequency in HIV⁺ women. Since senescent T lymphocytes are characterized by lack of the costimulatory signal transduced by CD28, decreased ability to secrete IL-2 upon activation and poor proliferative capacity (reviewed in [66]), it is not surprising that accumulation of these cells is associated with decline of the immune function and poor humoral response to influenza vaccination. B cell exhaustion and loss of memory B cells in HIV-infected adults [67] have been associated with impaired humoral response to influenza vaccination. In our cohort alterations in B cell subsets distribution and increased B cell exhaustion (FcRL4) were observed in HIV⁺ as compared to the uninfected participants. However, no significant association between vaccine-specific Ab titers and these B cell abnormalities was observed in this cohort. Moreover, the fact that no correlation between altered B cell subsets and TNFα levels was observed suggests that the overriding defect is that of TNFα-derived excessive T cell activation.

Limitations of this pilot study are the small sample size, absence of data for other causes of immune activation such as chronic co-infections e.g. with CMV, HCV [68,69], and possibility of residual HIV replication despite virus suppression [70], and the lack of a study group composed of young HIV-infected and uninfected donors, that would have allowed to dissect the effects of aging versus HIV infection on the immune response to influenza vaccine. Moreover, Ab titers were established for the whole vaccine rather than for its three individual components, possibly underestimating the contribution of the specific strains to the seroprotection levels of the donors. Nevertheless, this is the first study to focus on influenza vaccine-induced Ab responses in the context of aging and HIV infection and immune activation. The novel finding reported herein is the association of activated pTfh and impaired IL-21 production, with negative correlation of TNFα with poor Ab responses. Elevated plasma TNFα levels are present in HIV^– aging population as well as in young HIV⁺ individuals [35,71]. Our results provide a rationale to further evaluate the role of pTfh in vaccine response in elderly HIV⁺ individuals and in other proinflammatory conditions where abnormal antibody responses are observed upon vaccination. Interventions aimed to reduce chronic inflammation and activation in HIV-infected patients may improve their response to vaccination. Studies in larger cohorts may help to define predictive immunologic signatures of vaccine responses.