Research Article

Functional heterogeneity of human memory CD4+ T cell clones primed by pathogens or vaccines

See allHide authors and affiliations

Science  23 Jan 2015:
Vol. 347, Issue 6220, pp. 400-406
DOI: 10.1126/science.1260668

For T cells, variety is the spice of life

CD4+ helper T cells come in a variety of flavors. This allows them to respond in a manner that is tailored to the pathogen they encounter. Becattini et al. wondered whether multiple “flavors” of human CD4+ T cells respond to specific stimuli or if just one flavor dominates. To find out, they stimulated human memory CD4+ T cells with a fungus, a bacteria, or a vaccine antigen. Multiple helper cell subsets participated in each response. T cell receptor sequencing revealed that in some cases, T cells with the same specificity acquired different helper cell fates. Thus, there is more heterogeneity in human T cell responses than previously appreciated.

Science, this issue p. 400

Abstract

Distinct types of CD4+ T cells protect the host against different classes of pathogens. However, it is unclear whether a given pathogen induces a single type of polarized T cell. By combining antigenic stimulation and T cell receptor deep sequencing, we found that human pathogen- and vaccine-specific T helper 1 (TH1), TH2, and TH17 memory cells have different frequencies but comparable diversity and comprise not only clones polarized toward a single fate, but also clones whose progeny have acquired multiple fates. Single naïve T cells primed by a pathogen in vitro could also give rise to multiple fates. Our results unravel an unexpected degree of interclonal and intraclonal functional heterogeneity of the human T cell response and suggest that polarized responses result from preferential expansion rather than priming.

The functional diversity of CD4 T helper (TH) cells has evolved as a means to provide the immune system with the capacity to mount the appropriate type of defense against different classes of pathogens (1, 2). The response to viruses and intracellular bacteria is dominated by TH1 cells that produce the cytokine interferon (IFN)–γ and activate macrophage effector function. TH17 cells that produce the cytokines interleukin (IL)–17 and IL-22 and activate neutrophils dominate the response to fungi and extracellular bacteria. The response to parasites and venoms is dominated by TH2 cells that produce IL-4, IL-5, and IL-13 and activate eosinophils and other innate cells. Classical observations in mice and humans with defects in T cell polarization demonstrate the importance of generating the right type of response against each class of pathogen (37).

TH1, TH2, and TH17 are considered alternative fates of differentiating CD4 T cells, which are instructed by signals from the T cell receptor (TCR), costimulatory molecules, and cytokines (810). However, it is increasingly appreciated that most polarized T cells are not terminally differentiated but rather maintain flexibility to change the type, or increase the range, of cytokines produced (11). Although there is evidence for such cytokine flexibility from in vitro experiments and animal models (1214), it is unclear whether and how this mechanism contributes to the heterogeneity of the human memory T cell pool.

In principle, a functionally heterogeneous population of memory T cells may derive from the priming of multiple clones, each polarized to a distinct fate, or from the generation of multiple fates within the progeny of the same T cell clone (15). The “one cell, multiple fate” model has been documented in mouse experiments: Single naïve CD8 T cells were found to give rise to a heterogeneous progeny comprising T central memory (TCM) and T effector memory (TEM) cells (1618), and single naïve CD4 T cells were found to give rise to TH1 and T follicular helper (TFH) cells (19). However, both TCM and TFH cells can be considered intermediate stages of differentiation rather than alternative fates. Therefore, it remains to be determined whether the “one cell, multiple fate” paradigm might apply to alternative differentiation states, such as TH1, TH2, and TH17.

Functional heterogeneity and clonal composition of Candida albicans–specific CD4 T cells

To dissect the heterogeneity of human memory CD4 T cells, we isolated four memory TH cell subsets, distinguished by chemokine receptor expression, from the peripheral blood mononuclear cells (PBMCs) of healthy donors (fig. S1) (20, 21): TH1 (CXCR3+CCR4CCR6), TH2 (CCR4+CXCR3CCR6), TH17 (CCR4+CCR6+CXCR3), and nonconventional TH1 (CXCR3+CCR6+CCR4, hereafter defined as TH1*) that produced IFN-γ and low levels of IL-17 (22). Cells were labeled with carboxyfluorescein succinimidyl ester (CFSE) and stimulated with C. albicans in the presence of autologous monocytes. In all donors analyzed, CFSElo proliferating cells were found in high proportions in the TH17 and TH1* subsets but were also detected in the TH1 and TH2 subsets (Fig. 1, A and B) (23), consistent with previous reports (22, 24, 25). The pattern of cytokines produced in C. albicans–stimulated cultures was distinct and characteristic of each subset. C. albicans–specific TH17 cells produced IL-17A, IL-22, and IFN-γ; TH1* cells produced high amounts of IFN-γ and low amounts of IL-17A and IL-4; TH2 cells produced IL-4, IL-5, and IL-13; and TH1 cells produced IFN-γ (Fig. 1C). These findings indicate that human memory CD4 T cells responding to C. albicans are functionally heterogeneous in terms of migratory properties and effector function.

Fig. 1 Functional heterogeneity and clonal composition of C. albicans–specific CD4 T cells.

Human memory CD4 T cell subsets were isolated according to the expression of chemokine receptors, labeled with CFSE, and stimulated with heat-killed C. albicans in the presence of autologous monocytes. (A) CFSE profiles on day 6 and percentage of CFSElo proliferating cells in a representative donor. (B) Median values with 25th and 75th percentiles of CFSElo cells in 20 donors, with whiskers representing the highest and lowest values. ***P < 0.0001, **P < 0.001, *P < 0.05, as determined by nonparametric Friedman test. (C) Cytokine production measured on day 6 in the supernatants of the C. albicans–stimulated cultures. Values represent means ± SEM (n = 5). For comparison, the amounts of cytokines produced by polyclonally stimulated naïve T cells are reported in (39). (D) Number of unique TCRβ clonotypes detected by deep sequencing in C. albicans–specific T cells isolated from the four memory subsets. Each symbol represents a different donor. Lines represent mean values. (E) Frequency distribution of clonotypes in C. albicans–specific T cells from the four subsets isolated from donor CA-01. Dotted line represents the 1% frequency threshold. Numbers indicate total clonotypes in the subset and clonotypes present at frequencies >1% with their cumulative frequency (percentage of reads). Disparity index was close to 1 in all subsets. (F and G) Percent of clonotypes with frequencies >1% (F) and their cumulative frequencies (percentage of reads) (G) in five donors (means ± SEM). (H) Comparison of clonotype frequency distribution of C. albicans–specific memory CD4 T cells (y axis) and total memory CD4 T cells sequenced directly ex vivo after sorting (x axis). Dots outside the dotted lines represent clonotypes that were found in only one subset and that were assigned an arbitrary frequency value for graphical purposes. Shown is the Spearman correlation and the significance analyzed by paired t test.

To analyze the TCR repertoire of antigen-specific memory cells within each TH subset, we isolated CFSElo cells from C. albicans–stimulated cultures from five healthy donors and used multiplex polymerase chain reaction (PCR) assays on genomic DNA for quantitative assessment of TCRβ clonotypes by deep sequencing [quantification and correction procedures are described in fig. S2 and (20, 26)]. Even though the numbers of T cells recovered from the TH17 and TH1* subsets were higher than for TH1 or TH2 by a factor of 5 to 20 (table S1), the number of clonotypes detected in the four subsets was comparable within each donor tested, with an average of 976, 595, 830, and 696 unique clonotypes in TH1, TH2, TH1*, and TH17 subsets, respectively (Fig. 1D). Furthermore, the frequency distribution of individual clonotypes was comparable among the four TH subsets and varied widely, ranging from 20% to 0.01% of total reads (Fig. 1E). In all subsets and donors, fewer than 20 clonotypes (representing 3 to 4.5% of total clonotypes) were present at frequencies above 1% and together contributed 34 to 43% of total reads (Fig. 1, F and G, and table S1).

To investigate whether the most frequent TCRβ clonotypes detected after in vitro stimulation were already expanded in vivo, we compared the clonotype composition of in vitro expanded C. albicans–specific CD4 T cells and total memory CD4 T cells immediately after isolation from peripheral blood. In one donor, we detected 300,945 clonotypes in 5 × 106 total memory cells, whereas we detected 2350 clonotypes in C. albicans–specific cells (Fig. 1H). Of the latter, 1426 (corresponding to 60.7% of unique clonotypes and 86.7% of total reads) were also found in unstimulated total memory cells. Furthermore, among the shared clonotypes, we observed a significant positive correlation between the frequency in the C. albicans–specific cell line and that in the total memory pool. Collectively, the above analysis indicates that memory CD4 T cells specific for C. albicans are highly polyclonal and that individual clonotypes show a broad range of frequency distribution that is comparable in different memory subsets and reflects their frequency in the total memory pool.

Unique and shared clonotypes among C. albicans–specific memory T cell subsets

The diversity and functional heterogeneity of the C. albicans–specific memory pool could be due to different clones each polarized to a single fate, or to clones that had acquired multiple fates through a process of intraclonal diversification. A comparative analysis of the TCRβ repertoires of C. albicans–specific T cells from a representative donor (Fig. 2A) showed that of the 395 clonotypes detected in the TH17 subset, 143 were also found in the TH1* subset, 53 in the TH2 subset, and 8 in the TH1 subset. Furthermore, among the 143 clonotypes shared between TH17 and TH1*, 27 were also found in the TH2 subset, and 3 in both TH2 and TH1 subsets (Fig. 2B). Several shared clonotypes were found to be highly expanded in both TH17 and TH1* subsets. Moreover, several expanded clonotypes were found only in one subset, excluding the possibility that clonotype sharing was due to contamination during cell sorting. Similar patterns of clonotype sharing among memory subsets were found in the additional four donors analyzed (fig. S3). From the compiled analysis of the five donors, the highest sharing of clonotypes and reads was found between the TH17 and TH1* subsets, followed by the TH17 and TH2 subsets, whereas the lowest was between the TH17 and TH1 subsets (Fig. 2C). Collectively, the above findings indicate that the functional heterogeneity of C. albicans–specific memory T cells rests both on clones that are uniquely polarized to a single fate and on clones that have generated multiple and distinct T cell fates through a process of intraclonal diversification.

Fig. 2 Unique and shared clonotypes among C. albicans–specific memory T cell subsets.

(A) Comparison of clonotype frequency distribution in samples of T cells isolated from C. albicans–stimulated TH1, TH2, TH1*, or TH17 subsets from donor CA-02. Frequencies are shown as percentage of total reads. The total number of clonotypes in each sample is indicated on x and y axis. Values in the lower right corner represent the number of shared clonotypes between the two samples. The Spearman correlation and paired t test values are shown when significant. (B) Venn diagrams showing the number of unique and shared clonotypes in the four subsets of donor CA-02. (C) Bar histograms showing the percentage of clonotypes (upper panel) and the percentage of reads (lower panel) that are shared by the indicated subsets in the five donors. Data are means ± SEM. Each dot represents a different donor. ***P < 0.0001, **P < 0.001, *P < 0.05, as determined by nonparametric Friedman test.

Different patterns of clonotype sharing among Mycobacterium tuberculosis–specific T cell subsets

The high extent of clonotype sharing between TH17 and TH1* observed in the response to C. albicans could have been predicted on the basis of some commonalities between these two subsets as well as the reported plasticity of TH17 cells (1214). To investigate whether clonotype sharing is a general property of these memory subsets, we analyzed memory T cells specific for a different pathogen. Three CCR6+ T cell subsets were sorted from PBMCs of healthy donors (fig. S4) (27). M. tuberculosis–specific T cells were found in high numbers in the TH1* subset (CXCR3+CCR4) and at lower numbers in the TH17 subset (CCR4+CXCR3) (28), as well as in a subset of CXCR3+CCR4+ T cells, which, similarly to TH1*, produced IFN-γ and low amounts of IL-17 (fig. S5, A and B). As observed in the response to C. albicans, despite the difference in proliferating cells (27), the TCRβ diversity of M. tuberculosis–specific T cells was comparable in the three subsets, with an average of 152, 158, and 142 clonotypes in TH1*, TH17, and CXCR3+CCR4+ T cell subsets, respectively (fig. S5C). However, in contrast to what we observed for C. albicans, very few clonotypes were shared between M. tuberculosis–specific TH1* and TH17, whereas several were shared between TH1* and CXCR3+CCR4+ T cells (fig. S5, C and D). These findings indicate that the main patterns of clonotype sharing between different TH subsets are characteristic for different antigens and suggest that TH1* cells may be generated indirectly from plastic TH17 cells as well as directly without transition through a TH17 stage.

Extensive clonotype sharing among TT-specific memory T cell subsets

The heterogeneity of memory T cells observed for C. albicans and M. tuberculosis could be attributed to the multiplicity of antigens and innate receptor ligands present in these microbes. To address whether a single protein antigen would induce a less diverse and more polarized response, we analyzed memory T cells primed by the tetanus toxoid (TT) vaccine, which is given with alum as adjuvant. PBMC-sorted, CFSE-labeled TH1, TH2, TH1*, and TH17 cells were stimulated with TT in the presence of autologous monocytes (29). Among the four subsets, the extent of cell proliferation was comparable but the pattern of cytokine production was distinct (Fig. 3, A to C). For example, TT-specific TH1 cells produced IFN-γ, as expected, but also low levels of IL-4, IL-5, and IL-13.

Fig. 3 Extensive clonotype sharing among TT-specific CD4 T cell subsets.

(A) CFSE profiles and percentage of CFSElo proliferating cells in memory T cell subsets stimulated with TT in one representative donor. (B) Median values with 25th and 75th percentiles in four donors; whiskers represent the highest and lowest values. (C) Cytokine production in day 6 supernatants of TT-stimulated cultures. Values represent means ± SEM (n = 4). (D) Frequency distribution of clonotypes in TT-specific T cells from the four subsets isolated from donor TT-01. Dotted line represents the 1% frequency threshold. Numbers indicate total clonotypes in the subset and clonotypes present at frequencies >1% with their cumulative frequency (percentage of reads). Disparity index was close to 1 in all subsets. (E and F) Percent of clonotypes with frequencies >1% (E) and their cumulative frequencies (percentage of reads) (F) in four donors (mean ± SEM). (G) Comparison of clonotype frequency distribution in samples of T cells isolated from TT-stimulated TH1, TH2, TH1*, or TH17 subsets from donor TT-01. Frequencies are shown as percentage of total reads. The total number of clonotypes in each sample is indicated on the x and y axis. Shown are the number of clonotypes shared between the two samples and the Spearman correlation and paired t test value. (H) Bar histograms showing for four donors the percentage of clonotypes (upper panel) and percentage of reads (lower panel) that are shared by the indicated subsets. Data are means ± SEM. Each dot represents a different donor.

TCRβ sequencing revealed a high and comparable number of clonotypes in the four subsets, with a broad distribution of frequencies and a few dominant clonotypes accounting for a large fraction of total reads (Fig. 3, D to F). Surprisingly, there was a high level of clonotype sharing among all four TT-specific subsets (Fig. 3, G and H), with several clonotypes being present in three and even all four subsets (fig. S6). These results indicate that, in contrast to our expectation, the response to TT is characterized by a higher TCR diversity and clonotype sharing as compared to the response to complex microbes.

Isolation and characterization of sister clones from different memory subsets

To provide further evidence that memory T cells expressing the same αβ TCR can acquire in vivo different phenotypes and functional properties, we characterized T cell clones isolated from the CFSElo C. albicans–specific cells that were frozen as backup. From donor CA-01 we isolated 242 C. albicans–specific T cell clones, of which 80 were sister clones found in more than one subset (Fig. 4A). Consistent with the deep sequencing data, sister clones were frequently isolated from the TH17 and TH1* subsets and from the TH17 and TH2 subsets (Fig. 4B). Sequencing of the TCRα chain confirmed that the sister clones isolated from different subsets carried the same αβ TCR (Table 1).

Fig. 4 Isolation of sister clones from different memory subsets.

T cell clones were isolated from cryopreserved samples of C. albicans–stimulated CFSElo TH1, TH2, TH1*, and TH17 cells from donor CA-01. TCRβ sequences were obtained by Sanger sequencing on amplified cDNA. (A) For each subset the plot indicates the total number of clones isolated (center), the number of clones unique to that subset (white sector), and the number of clones whose TCRβ was found also in a sister clone isolated from another subset (color-coded sectors). (B) Frequency distribution (based on deep sequencing data) of the clonotypes corresponding to the clones that were found in a single subset (white circles) and in multiple subsets (colored circles). (C and D) Characterization of sister clones isolated from different memory subsets. Shown are cytokine production (C) and expression of transcription factors and CCR6 (D). (E and F) Characterization of two pairs of sister clones isolated from TH1* and TH17 subsets (E) and from TH2 and TH17 subsets (F).

Table 1 T cell clones with identical TCR isolated from different memory T cell subsets.

Complementarity-determiningregion 3 (CDR3) sequences were obtained for both TCRα and TCRβ chains. Amino acid abbreviations: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; Y, Tyr.

View this table:

The sister clones isolated were further characterized phenotypically and functionally. With the exception of IFN-γ, which was produced by virtually all clones as a consequence of the in vitro expansion, cytokine production was consistent with the origin of the cells. IL-4 was produced at high levels by sister clones isolated from the TH2 subset, whereas IL-17 and IL-22 were produced by sister clones isolated from the TH17 and TH1* subsets (Fig. 4C). In addition, CCR6 expression and RORγt mRNA were detected only in sister clones derived from TH17 and TH1* (Fig. 4D). As examples, sister clones of clonotype 11, isolated from the TH17 and TH1* subsets, and of clonotype 64, isolated from the TH17 and TH2 subsets, showed the characteristic polarized cytokine production and expression of transcription factors (Fig. 4, E and F). Sister clones with different functional properties (i.e., TH1 and TH2, or TH2 and TH17) were also isolated from different subsets of TT-specific memory cells.

Generation of multiple fates by in vitro priming of naïve CD4 T cells

The above analysis shows that memory CD4 T cells primed in vivo by pathogens or vaccines are highly heterogeneous, both at the population level and at the clonal level. Given the generation of multiple naïve T cells with identical TCRs in the thymus and their possible expansion in the periphery, one cannot rule out the possibility that some of the heterogeneity observed at the clonal level may reflect spatially and temporally distinct priming events of sister naïve T cells. Furthermore, heterogeneous fates may derive from the differential secondary stimulation of memory T cell clones. To test whether one round of stimulation could imprint heterogeneous fates within the progeny of a single naïve T cell, we primed in vitro a relatively small number of highly purified naïve CD4 T cells with C. albicans in the presence of autologous monocytes (30). As shown in Fig. 5 for one representative donor, proliferating CSFElo T cells recovered on day 15 produced IFN-γ, IL-17, IL-22, and IL-4 in various combinations (Fig. 5A). Using the cytokine secretion assay, we sorted IL-17+ (IFN-γ, IL-4), IFN-γ+ (IL-17, IL-4), and IL-4+ (IL-17, IFN-γ) T cells (31), which were directly cloned by limiting dilution and further expanded in bulk cultures with IL-2 for clonotypic analysis (Fig. 5B). As expected, the expanded populations produced primarily the original cytokine (i.e., IL-17, IFN-γ, or IL-4) (fig. S7). However, some IFN-γ+ and some IL-17+ sorted cells acquired the capacity to produce IL-4, consistent with flexibility in cytokine gene expression. TCRβ deep sequencing performed on the expanded cultures revealed a strikingly high level of clonotype sharing among the three cytokine-secreting cell populations (Fig. 5C). High frequencies of shared clonotypes and shared reads were found in three independent experiments performed with naïve T cells from different donors (Fig. 5D).

Fig. 5 Generation of multiple fates by in vitro priming of naïve T cells.

(A) Left: CFSE profile of CFSE-labeled naïve CD4 T cells primed in vitro by C. albicans in the presence of autologous monocytes and analyzed on day 15. Right: Intracellular cytokine staining on CFSElo cells stimulated with phorbol myristate acetate and ionomycin. (B) Schematic outline of the experimental approach: After priming in vitro, viable cytokine-secreting cells were enriched using the cytokine secretion assay (Miltenyi). Three populations sorted as IL-17+ (IFN-γ IL-4), IFN-γ+ (IL-17 IL-4), IL-4+ (IL-17 IFN-γ) were further expanded in IL-2 and used for TCRβ deep sequencing and T cell cloning. (C) Frequency distribution of TCRβ clonotypes in the indicated cytokine-secreting T cell populations from donor N-01. Frequencies are shown as percentage of total reads. The total number of clonotypes in each population is indicated in parenthesis on x and y axis. Shown are the number of clonotypes shared between two samples and the Spearman correlation and paired t test values. (D) Bar histograms showing the percentage of clonotypes (left) and percentage of reads (right) that are shared between the indicated cytokine-secreting T cell populations. Data are means ± SEM (n = 3). Each dot represents a different donor.

To further corroborate the deep sequencing data, we analyzed 205 T cell clones isolated from the different cytokine-secreting populations from one donor. Among the 88 unique TCRβ clonotypes identified, 21 were found in clones isolated from two and even three cytokine-secreting populations (table S2). In particular, six clonotypes (8, 32, 37, 41, 43, and 59) were found in clones isolated from IL-17–, IFN-γ–, and IL-4–secreting cells. Collectively, these findings provide evidence that a single human naïve CD4 T cell can generate a heterogeneous progeny, even in a single round of antigenic stimulation.

Discussion

By combining antigenic stimulation with deep sequencing and isolation of T cell clones, we unraveled an unpredicted degree of heterogeneity in the human CD4 T cell response to microbes and vaccines. This heterogeneity is the result of two distinct phenomena: (i) interclonal heterogeneity (i.e., different clones that give rise to a homogeneous polarized progeny), and (ii) intraclonal heterogeneity (i.e., clones that give rise to daughter cells that adopt different fates). The high level of intraclonal heterogeneity observed in this study supports the “one cell, multiple fates” model of CD4 T cell differentiation and provides evidence for T cell plasticity in the context of the human immune response. Possible mechanisms that lead to interclonal and intraclonal heterogeneity can involve the spectrum of pathogen- and damage-associated molecular patterns associated with a given pathogen or vaccine, the nature of the antigen-presenting cells in different tissues, the strength of TCR stimulation and costimulation, and the stochastic stimulation and restimulation of individual T cells within a proliferating clone (10, 3234). Another mechanism, which may be particularly relevant for complex pathogens (35), involves the stimulation of T cells by cross-reactive antigens that trigger heterologous immunity (36, 37).

Intraclonal heterogeneity was observed in all responses analyzed, but the pattern and extent of clonotype sharing appeared to be characteristic for each antigen. In the response to C. albicans, the extensive clonotype sharing between TH17 and TH1* could be explained by the plasticity of TH17 cells that have been shown to acquire IFN-γ production in appropriate stimulatory conditions (1214). However, clonotype sharing was not observed between M. tuberculosis–specific TH17 and TH1* cells of healthy donors—a finding that would be consistent with different pathways leading to the generation of TH1*. In contrast, there are no precedents to explain the substantial degree of clonotype sharing between C. albicans–specific TH17 and TH2 subsets. However, a similar pattern of intraclonal diversification was found in clones of naïve T cells primed in vitro by C. albicans, indicating that multiple fates can be rapidly induced within a single round of stimulation through mechanisms that remain to be defined. Finally, it should be noted that the clonal differentiation along multiple fates does not apply to all clones. Several expanded clones were found uniquely in a single T cell subset.

The extensive clonotype sharing among all TT-specific memory subsets was surprising and may be related to repeated exposure to the vaccine in booster immunizations. In all cases, this finding supports the notion that intraclonal heterogeneity can be generated even in response to a single protein antigen and under defined, albeit nonphysiological, stimulatory conditions. We suggest that in this case intraclonal heterogeneity may result from the initial priming of nonpolarized memory T cells that are subsequently driven to differentiate toward distinct fates by homeostatic or tissue-specific mechanisms, a phenomenon that has been shown to occur in human TCM cells in vitro (38). All considered, the extent of intraclonal heterogeneity found ex vivo and after in vitro priming suggest that intraclonal diversification is a fundamental property of the CD4 T cell response.

The finding that memory T cells are present in higher numbers in particular subsets associated with a protective response (for instance, in TH17 for C. albicans and in TH1* for M. tuberculosis) is at odds with our finding that the diversity of antigen-specific clonotypes was comparable in all subsets analyzed. This observation would be compatible with a model whereby polarized T cell responses result from the selective expansion, rather than priming, of T cell clones stimulated by the pathogen in a particular microenvironment. Accordingly, a broad priming of effector T cells endowed with different functions and migratory capacities would ensure a range of differentiated precursors to be recruited and expanded where necessary.

Supplementary Materials

www.sciencemag.org/content/347/6220/400/suppl/DC1

Materials and Methods

Figs. S1 to S7

Tables S1 and S2

References (4044)

Data files

References and Notes

  1. See supplementary materials on Science Online.
  2. T cell subsets were isolated according to the differential expression of chemokine receptors, as previously described (22). Cells in the CXCR3+CCR4CCR6 subset (defined as TH1) produced IFN-γ and expressed T-bet mRNA; cells in the CCR4+CXCR3CCR6 subset (TH2) produced IL-4 and expressed high level of GATA3 mRNA; cells in the CCR4+CCR6+CXCR3 subset (TH17) produced IL-17A and expressed RORγt; cells in the CXCR3+CCR6+CCR4 (nonconventional TH1, defined as TH1*) produced IFN-γ and low levels of IL-17A and expressed both T-bet and RORγt mRNA.
  3. In all experiments, the specificity of proliferating memory T cells was assessed by sorting CFSElo cells and isolating T cell clones by limiting dilution. On average, more than 95% of the T cell clones responded to the antigen used in the initial stimulation.
  4. T cell subsets from two healthy donors were isolated from PBMCs, labeled with CFSE, and stimulated with M. tuberculosis. For donor MT-01, input cell numbers were TH1*, 1.5 × 106; TH17, 1.5 × 106; CXCR3+CCR4+ TH, 1.2 × 106. For donor MT-02, input cell numbers were TH1*, 1.5 × 106; TH17, 1.4 × 106; CXCR3+CCR4+ TH, 1.1 × 106. Recovered CFSElo cells on day 6 for donor MT-01 were TH1*, 2.2 × 106; TH17, 0.95 × 106; CXCR3+CCR4+ TH, 2.0 × 106. For donor MT-02, these numbers were TH1*, 2.4 × 106; TH17, 0.7 × 106; CXCR3+CCR4+ TH, 1.7 × 106.
  5. T cell subsets from four donors were isolated and stimulated with TT. Input cell number for each subset was 2.5 × 106 (donor TT-01), 1.5 × 106 (donors TT-02 and TT-03), and 1.1 × 106 (donor TT-04).
  6. Naïve T cells were isolated from three donors. Input cell numbers were 3.2 × 106 (donor N-01), 6.6 × 106 (donor N-02), and 8.8 × 106 (donor N-03).
  7. In control experiments, the purity of the sorted populations was confirmed by measuring IFNG, IL4, and IL17 mRNAs by quantitative PCR immediately after sorting.
  8. Production of cytokines was measured on day 6 in the supernatants of 5 × 104 naïve CD4 T cells stimulated with plate-bound CD3 and CD28 antibodies. Mean values and SEM (n = 3) were as follows: IFN-γ, 414 ± 270.2; IL-17, 30.7 ± 3.3; IL-22, 391.6 ± 129.2; IL-4, 3.6 ± 3.6; IL-5, 9.6 ± 5.4; IL-13, 91 ± 20.5.
  9. Acknowledgments: We thank D. Jarrossay for cell sorting; O. Petrini, C. Fragoso, and A. Sette for providing valuable reagents; the Adaptive Biotechnologies team for technical support; and S. Monticelli for critical reading of the manuscript. The data presented in this manuscript are tabulated in the main paper and in the supplementary materials. All TCR sequences are available in the supplementary materials as .txt files. FASTA format files are available at downloads.adaptivebiotech.com (username: SallustoSciDec2014data; password: di2PhaiT). The sequences have been deposited in the Gene Expression Omnibus (accession numbers xxxxxx). Supported by European Research Council grant 323183 PREDICT (F.S.), Swiss National Science Foundation grants 149475 (F.S.) and 147662 (A.L.), and European Commission grants FP7-HEALTH-2011-280873, ADITEC (A.L. and F.S.) and FP7-HEALTH-2013-601958, SUPERSIST (T.N.S.). The Institute for Research in Biomedicine is supported by the Helmut Horten Foundation.
View Abstract

Stay Connected to Science

Navigate This Article