Protective immunity to Mycobacterium tuberculosis (Mtb) is commonly ascribed to a Th1 profile; however, the involvement of Th17 cells remains to be clarified. Here, we characterized Mtb-specific CD4+ T cells in blood and bronchoalveolar lavages (BALs) from untreated subjects with either active tuberculosis disease (TB) or latent Mtb infection (LTBI), considered as prototypic models of uncontrolled or controlled infection, respectively. The production of IL-17A, IFN-γ, TNF-α, and IL-2 by Mtb-specific CD4+ T cells was assessed both directly ex vivo and following in vitro antigen-specific T-cell expansion. Unlike for extracellular bacteria, Mtb-specific CD4+ T-cell responses lacked immediate ex vivo IL-17A effector function in both LTBI and TB individuals. Furthermore, Mtb-specific Th17 cells were absent in BALs, while extracellular bacteria-specific Th17 cells were identified in gut biopsies of healthy individuals. Interestingly, only Mtb-specific CD4+ T cells from 50% of LTBI but not from TB subjects acquired the ability to produce IL-17A following Mtb-specific T-cell expansion. Finally, IL-17A acquisition by Mtb-specific CD4+ T cells correlated with the coexpression of CXCR3 and CCR6, currently associated to Th1 or Th17 profiles, respectively. Our data demonstrate that Mtb-specific Th17 cells are selectively undetectable in peripheral blood and BALs from TB patients.
Tuberculosis (TB) represents one of the major threats to human health due to a single etiologic agent, Mycobacterium tuberculosis (Mtb) . Mtb is a facultative intracellular organism, classified as acid-fast Gram-positive bacterium, infecting usually lungs by the aerogenic route. Mtb is phagocytosed by alveolar macrophages [2, 3] and killed into the lysosomal compartment, leading to Ag presentation by MHC class II to CD4+ T cells .
However, a century after BCG vaccine development, and after immunizing more than three billion people with BCG, very little is known about protective immunity to Mtb.
The current paradigm of human cellular immunity indicates that functionally distinct CD4+ T-cell populations are specifically involved against a variety of pathogens. In this model, Th1 cells intervene against viruses and intracellular pathogens, Th2 cells against parasites and Th17 cells against extracellular pathogens [5, 6] including bacteria and fungi [6, 7].
Consistently, the classical model suggests that Mtb-specific CD4+ T cells harboring a Th1 profile control Mtb infection. In particular, the production of cytokines such as IFN-γ or TNF-α contribute to the recruitment of monocytes and granulocytes and activate the antimicrobial activity of macrophages . Of note, we recently identified  distinct Mtb-specific Th1 signatures associated either with latent Mtb infection (LTBI) or active TB, considered as prototypic models of controlled or uncontrolled infection, respectively. Relatively recently, the implication of Th17 cells in the control of Mtb infection has been suggested but remains highly controversial [10-19].
We therefore extended our initial analyses and undertook a comprehensive characterization of Mtb-specific CD4+ T-cell responses in blood and tissues in LTBI and TB individuals. As a control, we also analyzed CD4+ T-cell responses against extracellular pathogens (Gram-positive bacteria (Staphylococcus aureus, Staphylococcus pneumoniae), Gram-negative bacteria (Pseudomonas aeruginosa, Klebsiella pneumonia, and Escherichia coli), and fungi (Candida albicans), and systematically compared those to T-cell responses against Mtb.
We investigated the ability of pathogen-specific CD4+ T cells to produce IL-17A in addition to IFN-γ, TNF-α, and IL-2, both directly ex vivo and following in vitro Ag-specific T-cell expansion. We also analyzed Mtb-specific T-cell responses in bronchoalveolar lavages (BALs) and, as control, assessed extracellular pathogen-specific CD4+ T-cell responses in gut mucosal tissues, i.e. at the relevant sites of pathogen exposure.
Our data demonstrate that in TB patients, Mtb-specific Th17 cells (but not Th1 cells) were undetectable in blood and BALs, both directly ex vivo (immediate) and following in vitro T-cell expansion. Conversely, in LTBI subjects, Mtb-specific CD4+ T-cell responses lack immediate IL-17A effector function, but acquire the ability to produce IL-17A following Mtb-specific T-cell expansion in 50% of cases. Interestingly, IL-17A acquisition by Mtb-specific CD4+ T cells correlated with the coexpression of the chemokine receptors CXCR3 and CCR6, currently associated to Th1 or Th17 profiles, respectively [6, 7].
These data demonstrate that Mtb-specific IL-17A–producing CD4+ T cells are not detectable in active or latent infection, but are successfully acquired in 50% of subjects with controlled Mtb infection upon in vitro T-cell expansion. This mechanism might be analogous to the one observed following influenza virus infection, where virus-specific CD8+ T cells lack direct cytotoxic potential but consistently acquire this function upon expansion [20, 21]. These data shed light on the controversial results on Mtb-specific Th17 cells and Mtb control obtained in human and mouse models [13, 22-28].
In the present study, we have investigated the presence of IL-17A–producing Mtb-specific CD4+ T cells in 10 TB patients and 28 LTBI subjects. Since IL-17A–producing CD4+ T-cell responses are commonly ascribed to extracellular pathogen-specific CD4+ T-cell responses, we have systematically compared Mtb- with extracellular pathogen-specific CD4+ T-cell responses obtained from 30 healthy individuals.
Lack of IL-17A production in Mtb-specific CD4+ T cells directly ex vivo
The functional profiles of Mtb-specific and extracellular pathogen-specific CD4+ T-cell responses were investigated ex vivo by intracellular cytokine staining according to the gating strategy shown in Supporting Information Fig. 1A. In particular, the ability of pathogen-specific CD4+ T cells to produce IL-17A in addition to IFN-γ, TNF-α, and IL-2 was assessed. In-depth analysis of pathogen-specific CD4+ T-cell responses demonstrated that the global functional profile of T-cell responses against the Mtb in LTBI or TB subjects and for extracellular pathogens in healthy subjects were all significantly different from each other (all p < 0.05 (except TB versus E. coli; p > 0.05)). In particular, Mtb-specific Th17 cells were not detected ex vivo in LTBI (n = 28) or in TB individuals (n = 7). Consistently with our previous study , Mtb-specific CD4+ T-cell responses were mostly composed of triple TNF-α/IFN-γ/IL-2 or of single TNF-α–producing cells in LTBI or in TB subjects, respectively (Fig. 1). In contrast, extracellular pathogen-specific CD4+ T-cell responses were dominated by single TNF-α–producing cells (76% for C. albicans, 73% for S. aureus [Gram-positive bacteria] and 57% for E. coli (Gram-negative bacteria)) (Fig. 1B and C). However, in contrast to Mtb-specific CD4+ T-cell responses, Th17 cells were frequently detected in response to extracellular pathogens (ranging from 20 to 50%), but represented a consistent but minor component of the responding CD4+ T cells (Supporting Information Fig. 2A and B). Of note, IL-17A production was also undetectable in CD3 negative cells, CD8+ T cells and CD3+CD4−CD8− cells following Mtb-specific stimulations, in both LTBI and TB individuals (Supporting Information Fig. 3). Conversely, IL-17A was consistently detected in CD4+ T cells in all subjects following polyclonal stimulations (Fig. 1B).
Mtb-specific CD4+ T cells from BAL of TB patients lack immediate IL-17A effector function
Since preferential accumulation of Ag-specific T cells at the site of pathogen-replication/exposure is well established , we investigated whether Mtb-specific IL-17A–producing CD4+ T cells in TB patients could be present in lung tissue. To address this issue, Mtb-specific CD4+ T-cell responses were assessed on cells isolated from both peripheral blood and bronchoalveolar fluids. As control, extracellular bacteria-specific T-cell responses were assessed on cells isolated from peripheral blood and gut mucosal biopsies from healthy subjects. As shown in the representative flow cytometric profiles, bacteria-specific IL-17A–producing CD4+ T cells were consistently detected in both blood and gut mucosal tissues from healthy individuals (Fig. 2A), whereas IL-17A–producing Mtb-specific CD4+ T-cell responses from cells isolated from either peripheral blood or BAL from TB patients were not detected (Fig. 2B). Cumulative analyses confirmed the lack of Mtb-specific, as compared with bacteria-specific, Th17 responses in the relevant tissue (i.e. BAL versus gut mucosa) (p = 0.0027; Fig. 2C). Consistently, the frequencies of bacteria-specific Th17 cells in gut biopsies were also significantly increased as compared with Mtb-specific Th17 cells in BAL (p = 0.001; Fig. 2D). These data indicate that Mtb-specific CD4+ T cells of TB patients lack immediate IL-17A effector functions also in BAL.
Mtb-specific CD4+ T cells in LTBI subjects but not in TB patients acquire IL-17A effector function
We then assessed whether Mtb-specific Th17 cells would become detectable after Ag-specific in vitro T-cell expansion. For that purpose, CFSE-labeled PBMCs from LTBI subjects or TB patients were stimulated with Mtb-derived Ags for 6 days. As a control, CFSE-labeled PBMCs from healthy individuals were also stimulated with extracellular pathogens for 6 days. Cell cultures were then restimulated with PMA/ionomycin to assess the cytokines profile of cultured CD4+ T cells. CD4+ T-cell responses were analyzed according to the gating strategy shown in Supporting Information Fig. 1B. Following in vitro expansion, the proportion of subjects with detectable Ag-specific (i.e. CFSElow) IL-17A–producing CD4+ T cells was significantly increased for C. albicans and Gram-positive bacteria (p = 0.001 and 0.007, respectively) but not for Gram-negative bacteria (Supporting Information Fig. 2A and B). Furthermore, the frequency of IL-17A–producing cells among the total responding CD4+ T cells (i.e. CFSElow cells) also significantly increased after in vitro T-cell expansion for C. albicans and Gram-positive bacteria (p = 0.0001 and 0.004, respectively) but not for Gram-negative bacteria (Supporting Information Fig. 2A and C). Interestingly, the proportion of subjects with detectable Mtb-specific (i.e. CFSElow) IL-17A–producing CD4+ T cells was significantly (p = 0.0002) increased in LTBI but not in TB individuals (Fig. 3A and C). The frequency of Mtb-specific IL-17A–producing CD4+ T cells among the total Mtb-specific CD4+ T-cell responses (i.e. CFSElow) was also significantly increased after in vitro T-cell expansion as compared with that of direct ex vivo assessment in LTBI subjects (p = 0.0027) but not in TB patients (Fig. 3A, B, and D). Of interest, C. albicans-specific (i.e. CFSElow) IL-17A–producing CD4+ T cells were consistently detected following T-cell expansion in 21 out of 22 LTBI subjects and in five out of six TB patients. The proportion of individuals with detectable IL-17A–producing CD4+ T cells in response to C. albicans and the frequencies of C. albicans-specific IL-17A–producing CD4+ T cells in LTBI subjects and TB patients were neither significantly different from each other nor from healthy subjects (data not shown). Of note, the magnitude of Mtb-specific CD4+ T-cell proliferation (i.e. the % of CFSElow CD4+ T cells) was not different between LTBI and TB individuals (data not shown).
Therefore, Mtb-specific Th17 cells were enriched following in vitro T-cell expansion in LTBI subjects but not in TB patients, with regards to both the proportion of responders with detectable Th17 cells (p = 0.04, Fig. 3C) and the frequency of Mtb-specific Th17 cells (p = 0.01, Fig. 3D).
Mtb-specific IL-17A effector function in LTBI subjects correlates with CCR6 and CXCR3 expression
Th1 and Th17 cells are characterized by the expression of specific chemokine receptors [6, 30]. In this model, Th1 cells express CXCR3 (and/or CCR5), while Th17 cells express CCR6, either alone or in combination with CCR4 [6, 30]. We therefore postulated that acquisition of IL-17A effector function by Mtb-specific CD4+ T cells from LTBI subjects might be related to the expression of CXCR3 and/or CCR6. To address this hypothesis, the expression of CXCR3 and CCR6 was assessed by flow cytometry directly ex vivo on Mtb-specific and C. albicans-specific (internal control) CD4+ T cells from LTBI subjects. The detection of Mtb-specific or C. albicans-specific CD4+ T cells was based on the expression of membrane-bound TNF-α (mTNF-α) . Of note, the percentage of Mtb-specific CD4+ T cells assessed by mTNF-α expression directly correlated with the frequencies of TNF-α–producing Mtb-specific CD4+ T cells assessed by intracellular staining (p = 0.02, data not shown).
Representative flow cytometric profiles as well as cumulative data show that C. albicans-specific CD4+ T cells were more represented in the single CCR6+ (CXCR3−) CD4+ T-cell subset than Mtb-specific CD4+ T cells (p < 0.05), while Mtb-specific CD4+ T cells were more represented in the single CXCR3+ (CCR6−) and in the dual CXCR3+/CCR6+ CD4+ T-cell subsets than C. albicans-specific CD4+ T cells (both p < 0.05) (Fig. 4A–C). Interestingly, the acquisition of IL-17A effector function by Mtb-specific CD4+ T cells directly correlated with the initial (ex vivo) proportion of Mtb-specific CD4+ T cells coexpressing CXCR3 and CCR6 (p = 0.002) (Fig. 4D). In contrast, no correlation was observed with any other T-cell subsets defined by the expression of CXCR3 and CCR6 (p > 0.05; data not shown). Of note, acquisition of IL-17A effector function was not related (p > 0.05) to the frequency of Ag-specific CD4+ T-cell proliferation (i.e. the percentage CFSElow CD4+ T cells), but directly correlated with the amount of IL-17A production in the supernatants of cell cultures following in vitro expansion (p < 0.0001; Supporting Information Fig. 4).
The classical model ascribes protective Mtb-specific CD4+ T-cell responses with a typical Th1 profile [8, 9]. However, contrasting evidences obtained in both the mouse and the human models suggested that Mtb-specific Th17 cells might contribute to the control of Mtb infection [11-13, 22-27]. These discrepancies, however, may result from the difficulty to compare in vitro versus ex vivo investigations or from the use of Ags discriminating T-cell responses induced by Mtb infection versus BCG vaccination.
In this context, we assessed the frequencies of IL-17A–producing Mtb-specific CD4+ T cells in both LTBI and TB individuals, isolated from peripheral blood and from BAL, directly ex vivo and following in vitro T-cell expansion. Since, Ag-specific Th17 cells are primarily associated with extracellular pathogens , we systematically compared the profiles of Mtb-specific with extracellular pathogens-specific Th17 cells. Of note, the frequency of IL-17A was also assessed in CD3-negative cells, CD8 T cells and CD3+CD4−CD8− cells.
In the present study, we demonstrate that the global functional profile of T-cell responses against Mtb in both LTBI and TB subjects and for extracellular pathogens in healthy subjects were all significantly different from each other. In contrast to previous studies performed in mice [22-26] or humans [13, 27], we found no evidence of IL-17A immediate effector function in response to Mtb in subjects with latent infection or in patients with TB disease, regardless of the T-cell population assessed (CD3-negative cells, CD4+ T cells, CD8 T cells, or CD3+CD4−CD8− cells) and the Ags used (ESAT-6, CFP-10, or PPD; data not shown). We then addressed the issue of the potential accumulation of Mtb-specific Th17 cells at the site of pathogen replication in the tissues, i.e. lung, and confirmed the lack of Mtb-specific IL-17A–producing CD4+ T cells in BALs. However, we cannot formally exclude that Mtb-specific Th17 cells might be sequestered in the active lesions of TB patients, and may not exudate into the BAL fluid. In addition, these conclusions only pertain to the Ags analyzed.
Of note, we observed that bacteria-specific Th17 cells were more frequently detected than C. albicans-specific Th17 cells in healthy individuals. These results are contrasting with the data published by Van de Veerdonk et al.  who showed that C. albicans were a more potent inducer of IL-17A as compared to Gram-negative bacteria. This discrepancy might be explained by the distinct experimental settings used (direct ex vivo stimulation versus in vitro T-cell expansion).
Also, we assessed whether extracellular pathogen-specific or Mtb-specific Th17 cells could be enriched following Ag-specific in vitro T-cell expansion. We found that C. albicans- and Gram-positive bacteria- but not Gram-negative bacteria-specific Th17 cells were enriched with regards to both, the proportion of responders with detectable Th17 cells and the proportion of IL-17A–producing cells within the global CD4+ T-cell responses. Of interest, we found that, upon Ag-specific T-cell expansion, Mtb-specific Th17 cells were only detected in 50% of LTBI subjects but not in TB patients, irrespectively of the magnitude of T-cell proliferation. Finally, acquisition of IL-17A effector function by Mtb-specific CD4+ T cells directly correlated with the coexpression of CXCR3 and CCR6, currently associated to Th1 or Th17 profiles, respectively [6, 7]. Two main scenarios could explain the detection of Mtb-specific Th17 cells only after expansion: (i) Mtb-specific Th17 cells represent a very rare CD4+ T-cell population which expands following Ag-specific stimulation or (ii) IL-17A is acquired by Mtb-specific Th1 following Ag-specific stimulation. Although we cannot formally exclude the first hypothesis, the fact that Mtb-specific CD4+ T cells coexpress CXCR3 and CCR6, and that the proportion of Mtb-specific Th17 among total Mtb-specific CD4+ T cells correlates with CCR6/CXCR3 expression, support the second hypothesis. Thus, the ability of Mtb-specific CD4+ T cells to acquire IL-17A seems to be absent in patient with uncontrolled Mtb infection. This mechanism might be analogous to the one observed following influenza virus infection, where virus-specific CD8+ T cells lack direct cytotoxic potential but acquire this function upon expansion [20, 21].
Since the antimycobacterial activity of β-defensins is well established , we postulate that one mechanism by which IL-17A production by Mtb-specific CD4+ T cells might contribute to the control of Mtb growth may rely on the IL-17A-induced secretion of β-defensins by alveolar epithelial cells [34, 35].
Consistently with the previous demonstration of Khader et al.  who showed that optimal protection against mucosal Mtb infection in immunized mice required mixed Th1/Th17 response, the present study provides new evidences that acquisition of IL-17A effector function by Mtb-specific CD4+ T cells is associated with the control of Mtb in humans. According to the recent observations of Curtis et al. , the major factor accounting for CD4+ T-cell plasticity might be the number of cell division following T-cell priming, cytokine environment, and chromatin modifications during priming. The reason why acquisition of IL-17A effector function by Mtb-specific CD4+ T cells was detected in about 50% and not in all LTBI subjects remains to be established. However, one could speculate that the heterogeneity of Mtb genotype, the Ag load, and the cytokine environment at the time of priming, might account in the potential acquisition of IL-17A effector function. Furthermore, this heterogeneity in the ability to generate Mtb-specific IL-17A–producing CD4+ T cells in LTBI subjects may also be associated to positive responses to tuberculin skin test (which is typically in the range of 50% in LTBI subjects), to the diversity of LTBI individuals ranging from exposed uninfected subjects to subjects with subclinical TB [37, 38] or to the probability to reactivate Mtb and to develop an active TB disease. Finally, it would be of interest to assess the presence of IL-17A–producing CD4+ T cells in TB patients upon completion of treatment. Indeed, we previously showed that efficient TB therapy strongly restores a polyfunctional, i.e. IFN-γ+TNF-α+IL-2+ Th1 profile .
Taken together, the present study sheds light on the controversial results obtained in human and mice models [13, 22-28] and advances the delineation of the functional profile of the human adaptive T-cell responses against bacterial and fungal pathogens. Furthermore, our results indicate that protective T-cell immunity likely results from the mobilization and cooperation of multiple functionally distinct T-cell populations.
Materials and methods
Seventy-one subjects were recruited in this study. LTBI subjects (n = 28) and TB patients (n = 10) were selected based on previously described criteria  and all were IFN-γ release assay positive. LTBI subjects were asymptomatic health-care workers routinely screened and TB patients were microbiologically demonstrated and were diagnosed by a clinician after validation of these criteria associated with clinical symptoms. None of these patients was under antimycobacterial treatment at the time of the present analyses. In addition, BALs were obtained from five TB patients. Furthermore, 33 healthy volunteers were recruited in this study and 5 individuals with normal colonoscopic findings also provided gut biopsies in addition to peripheral blood. These studies were approved by the Institutional Review Board of the Centre Hospitalier Universitaire Vaudois and informed written consent was obtained from each volunteer.
Mononuclear cells isolation from blood and tissues
Mononuclear cells were isolated either from peripheral blood using ficoll-histopaque separation , or following type II-S collagenase digestion (Sigma; 0.5 mg/mL; 37°C; 90 min) of gut biopsies , or following centrifugation (10 min, 300 g) of BAL .
Bacteria- and C. albicans-derived Ags were prepared as previously described . Mtb-derived CFP-10 and ESAT-6 peptides pools are composed of 15-mers overlapping by 11 amino-acids encompassing the entire sequences of the proteins and all peptides were HPLC purified (>80% purity). Stimulations were performed with 10 μg/mL of heat-inactivated C. albicans yeast, 5 × 107 CFU/mL of bacteria, or with 1 μg/mL of ESAT-6 and/or CFP-10 peptide pools. A pool of bacteria-derived Ags (S. aureus, S. pneumonia, P. aeruginosa, K. pmeunomiae, E. coli; 5 × 107 CFU/mL) was used for gut biopsy stimulations. Staphylococcus enterotoxin B (100 ng/ml; Sigma) was used as positive control.
Assessment of ex vivo CD4+ T-cell responses
Mononuclear cells (106 cells) were stimulated (18 h, 37°C, 5% CO2) with the aforementioned Ags in 1 mL of complete RPMI (10% FCS, penicillin/streptomycin) containing Golgiplug (1 μg/mL; BD Biosciences) as described . At the end of the stimulation period, cells were washed, permeabilized (Cytofix/Cytoperm solution; BD Biosciences), and stained as previously described .
Assessment of in vitro CD4+ T-cell expansion
Mononuclear cells were labeled with 5, 6-carboxyfluorescein succinimidyl ester (CFSE, Invitrogen; 0.25 μM), cultured in 4% human AB serum (Institut de Biotechnologies Jacques Boy) RPMI as previously described , and stimulated with Ags described earlier. After 5 days of culture, cells were washed and replated in complete medium for 18 h of rest. Cells were then restimulated for 6 h with phorbol myristate acetate (PMA; 100 ng/mL; Sigma) and ionomycin (1 μg/mL; Sigma) in the presence of Golgiplug (1 μg/mL).
Assessment of CXCR3 and CCR6 expression on Ag-specific CD4+ T cells using membrane-bound TNF-α
Mononuclear cells (106 cells/ml) were stimulated with heat-inactivated C. albicans yeast or ESAT-6 or CFP-10 peptide pools in complete RPMI containing TAP1–0 (10 μM; Calbiochem) and PE-CY7-conjugated anti-TNF-α Ab as described . At the end of the incubation period (6 h), cells were stained with Abs directed to CD3, CD4, CCR6, and CXCR3 and percentages of CXCR3 and CCR6 expression on Ag-specific CD4+ T cells were assessed.
Flow cytometry analyses
To assess Ag-specific CD4+ T-cell responses, the following Abs were used: CD4-allophycocyanin-H7 (clone SK3), CD8-PerCP-Cy5.5 (SK1), CD3-ECD (UCHT1), IFN-γ-AF700 (B27), IL-2-PE (MQ1–17H12), TNF-α-PECY7 (Mab11), IL-17A-AF647 (eBio64DEC17), CXCR3-allophycocyanin (TG1/CXCR3), and CCR6-PE (11A2). All Abs were purchased from BD Biosciences except IL-17A (eBioscience), CD3-ECD (Beckman Coulter) and CXCR3-allophycocyanin (Biolegend). Furthermore, dead cells were excluded using the violet LIVE/DEAD stain kit (Invitrogen). Data were acquired on an LSRII four-laser (405, 488, 532, and 633 nm) and analyzed using FlowJo version 8.8.6 (Tree Star Inc.). Analysis and presentation of distributions were performed using SPICE version 5.1, downloaded from http://exon.niaid.nih.gov/spice . Concerning the assessment of ex vivo CD4+ T-cell responses, an individual was considered as a positive responder for a particular Ag when at least one cytokine was positive. For blood analyses, the positivity of each cytokine was determined as follows: the cytokine frequency obtained in the sample must exceed the threshold (set as the mean of the controls for all donors + 2 SD (i.e. TNF-α: 0.032; IFN-γ: 0.017; IL-2: 0.018; IL-17A: 0.010)) and ≥3 times the value obtained in the corresponding individual's control. For tissues analyses, the positivity of each cytokine was determined as follows: the cytokine frequency obtained in the sample must be ≥3 times the value obtained in the corresponding individual's control. Concerning in vitro expanded CD4+ T-cell responses, the percentage of proliferating CD4+ T cells, i.e. CFSElow cells, was determined in the CD3+CD4+ T-cell population. The criteria for scoring as positive the proliferating cell cultures included: (i) percentage of CFSElow cell >1% after background subtraction (percentage of CFSElow cells in unstimulated cell cultures) and (ii) stimulation index (SI) >3. The SI is calculated as the ratio between stimulated versus unstimulated cell cultures. The positivity of each cytokine was determined as follows: The cytokine frequency obtained in the sample must be ≥3 times the value obtained in the corresponding individual's control.
p Values were derived from either χ2 analyses, for comparison of positive proportions, or one-way ANOVA (Kruskal–Wallis test), followed by Student's t-test. When applicable, Bonferroni correction for multiples analyses was applied. Regarding SPICE analyses of the flow-cytometry data, comparison of distributions was performed using a Student's t-test and a partial permutation test as described .
The research leading to these results was supported by the Swiss Vaccine Research Institute and by the Swiss National Science Foundation and has received funding from the European Community's Seventh Framework Programme ((FP7/2007–2013), (FP7/2007–2011)) under EC-GA n° (241642) and (FP7-20072013 under grant agreement HEALTHF2-2010-260338ALLFUN). We are grateful to John and Aaron Weddle for their assistance for figures preparation.
Conflict of interest
The authors declare no financial and commercial conflict of interest.