A peripheral circulating compartment
of natural naive CD4+ Tregs (original) (raw)
Subsets of naive CD4+CD25+ Tregs in human peripheral blood. Human naive and memory T cells can be distinguished by the reciprocal expression of CD45RA or CD45RO isoforms (19). Staining of peripheral blood T cells with antibodies to CD45RA, CD45RO, and CD25 revealed 4 subsets of CD4+ T cells: 1 naive CD45RA+(RO–)CD25– subset, a clearly distinct CD45RA+(RO–)CD25+ population, and 2 memory subsets, 1 CD45RA–(RO+) population expressing intermediate levels of CD25, which we will define as CD45RA–CD25–, and the previously described CD45RA–(RO+)CD25high population corresponding to Tregs (Figure 1B; compare with Figure 1A). Both CD45RA+CD25– and CD25+ cells expressed high levels of CD62L and CCR7, which were also expressed by a large proportion of Tregs. CD45RA+CD25+ cells and Tregs expressed higher ex vivo levels of CTLA-4 as compared with their corresponding CD25– counterpart populations (Figure 1C; Table 1). To assess the ex vivo expression of FOXP3 in the 4 defined CD4+ subsets, highly purified naive and antigen-experienced CD4+CD25+ and CD25– T cell populations were obtained using a procedure that involved magnetic isolation of CD4+ T cells from circulating peripheral lymphocytes followed by separation of the 4 populations by FACS cell sorting. Semiquantitative PCR analysis of the sorted populations revealed that detectable ex vivo expression of FOXP3 was confined to the CD45RA+CD25+ cell and Treg subsets (Figure 1D). It is noteworthy that the ex vivo expression level of FOXP3 in the CD45RA+CD25+ population, as assessed by real-time PCR, was comparable to that found in Tregs.
Expression of CD25 and CD45RA defines 4 distinct subsets of peripheral blood CD4+ T cells. Peripheral blood lymphocytes were stained with monoclonal antibodies to CD3, CD4, CD25, and CD45RA, which identified 4 major subsets. (A) Gated on lymphocytes. (B) Gated on CD4+ T cells. (C) CD4+ T cell subsets were further analyzed for the expression of CD45RO, CCR7, CD62L, and CTLA-4. Numbers indicate percentages of CCR7- or CD62L-expressing cells. (D) FOXP3 gene expression was assessed by conventional and by quantitative real-time PCR on the corresponding 4 sorted populations. For the PCR, an example from a representative donor is shown. Δ Rn is the difference between the Rn (emission intensity of the reporter dye/emission intensity of the passive reference) of the sample (PCR reaction with cDNA template) and the Rn of a control (PCR reaction without cDNA template or early cycles of a real-time reaction). Real-time PCR data are shown as the mean values obtained from 3 independent donors.
Phenotype of CD4+ T cell subsets defined by expression of CD25 and CD45RA
Together, these results define a subset of peripheral CD4+CD25+ T cells exhibiting a naive phenotype, which we named natural naive Tregs (NnTregs). To evaluate this newly defined T cell subset, we examined the percentage of NnTregs in relation to age. In cord blood, up to 7% of total CD4+ T cells displayed this phenotype (n = 7, mean = 5.8 ± 1.1). In adults, the population accounted for 0.2 to 3.3% of the total circulating CD4+ T cells in a panel of healthy individuals (n = 47, mean 1.4 ± 0.7). We found a significant inverse correlation between the percentage of NnTregs and the age of the donors (Figure 2A) and a significant direct correlation between the percentage of NnTregs and that of total CD4+ naive T cells (Figure 2B). In contrast, no significant correlation was found between age and proportion of Tregs (n = 47, mean 1.7 ± 0.9) (Figure 2C) or between relative proportions of the NnTreg and Treg subsets (Figure 2D).
Variation in the frequency of circulating NnTregs and Tregs during aging. Dots correspond to individual samples tested. Linear regression was calculated either taking into account all samples (solid lines) or only adult blood samples (dashed lines). A significant inverse correlation was found between the frequency of NnTregs and age (A), and a significant positive correlation was found between the frequency of NnTregs and total naive CD4+ T cells (B). In contrast, no significant correlation was found between the frequency of Tregs and age (C) or between the frequency of NnTregs and Tregs (D).
Ex vivo phenotypic analysis and replicative history of NnTregs. To investigate further the characteristics of NnTregs, we analyzed the ex vivo expression of a panel of molecular markers by this population in comparison with that of other subsets of CD4+ T cells (Table 1). In addition to the expression of CD45RA, CCR7, and CD62L and the lack of expression of CD45RO by the large majority of NnTregs, this population shared other features with CD4+ naive T cells, including exhibiting no significant expression of HLA-DR, which was expressed by a considerable fraction of Tregs. Chemokine receptors CCR4 and CCR5 were expressed by a large fraction of Tregs. However, only a minor fraction of NnTregs expressed CCR4, whereas no significant expression of CCR5 was detected. Interestingly, NnTregs expressed lower levels of CD45RA as compared with CD25– naive T cells. They expressed levels of CD38 (ADP-ribosyl cyclase/cyclic ADP-ribose hydrolase) that were lower than those of CD25– naive T cells and comparable to those of the antigen-experienced CD25– population. In addition, they included lower proportions of CD31-expressing (PECAM-1–expressing) cells as compared with other naive CD4+ T cells, but much higher proportions compared with both the antigen-experienced CD25– and Treg fractions. In addition to evaluating expression of CTLA-4, we addressed the ex vivo expression by NnTregs of 2 other molecules that have been reported as involved in Treg functions, TGF-β1 and glucocorticoid-induced TNF receptor (GITR). In mice, expression of membrane TGF-β1 is clearly detectable using monoclonal antibodies against active TGF-β1. In contrast, only marginal amounts of TGF-β1 can be detected at the surface of human cells (20). However, surface expression of latent TGF-β can be detected in sorted and activated human CD4+CD25high T cells by staining with an antibody specific for human TGF-β1 latency-associated peptide (LAP). Using this antibody, we detected ex vivo a level of TGF-β1 LAP expression in Tregs that was only moderately increased as compared with that in antigen-experienced CD25– T cells (Table 1). The increase was even smaller (at the limit of detection) for NnTregs as compared with naive CD25– T cells. Only a small increase in the level of GITR expression was detectable ex vivo for Tregs, as compared with that of antigen-experienced CD25– T cells, whereas no significant increase was detectable for NnTregs as compared with naive CD25– T cells.
We next evaluated the replicative history of NnTregs by assessing their telomeres’ length and T cell receptor excision circle (TREC) content in comparison with that of the other populations defined by expression of CD45RA and CD25, following ex vivo sorting. The analysis revealed that NnTregs have longer telomeres than both antigen-experienced CD25– cells and Tregs, but that their telomeres are similar in length to those of other naive CD4+ T cells (Figure 3, A and B). The TREC content of NnTregs was also similar to that of naive CD25– T cells and higher than that of the antigen-experienced CD25– T cell fraction. In contrast, the TREC content of Tregs was lower than that of antigen-experienced CD25– T cells and below the detection limit of the assay (Figure 3C). These results are consistent with the concept that NnTregs are in an early differentiation stage comparable to that of CD45RA+CD25– naive CD4+ T cells.
Replicative history of CD4+ T cell populations defined by the expression of CD45RA and CD25. (A) Ex vivo–sorted CD4+ T cell subsets were assessed for telomere length using telomere-specific peptide nucleic acid–fluorescent in situ hybridization (FISH) followed by flow cytometry analysis. Histograms shown are from 1 representative donor. (B) Telomere length data are shown as the mean values obtained for 3 independent donors. (C) Quantification of TREC content of ex vivo–sorted CD4+ T cell subsets. Genomic DNA of sorted subsets was isolated, and the number of TRECs was determined by quantitative real-time PCR. Data are shown as the mean values obtained for 3 independent donors. *Below detection limit.
Functional analysis of NnTregs. Naturally occurring CD4+CD25+ Treg populations have been previously shown to be composed of anergic cells, i.e., cells unable to significantly proliferate or produce cytokines following TCR-mediated stimulation. To assess NnTreg function in comparison with that of other subsets of CD4+ T cells defined by ex vivo expression of CD45RA and CD25, sorted populations from each subset were stimulated with anti-CD3/CD28 antibodies in the presence of irradiated autologous APCs. Both antigen-experienced and naive CD25– T cells showed a vigorous proliferative response. In contrast, no significant proliferation was observed for Tregs, and only a low level of proliferation, clearly inferior to that of naive CD25– T cells, was observed for NnTregs under these experimental conditions (Figure 4A). The 4 CD4+ T cell populations were also assessed with respect to their cytokine secretion profiles following stimulation with PMA/ionomycin. Antigen-experienced CD25– T cells produced both IL-2 and IFN-γ, whereas naive CD25– T cells produced IL-2 and very little IFN-γ (Figure 4B). In contrast, both Tregs and NnTregs produced only low levels of IL-2 and IFN-γ. No detectable levels of IL-4, IL-10, or TGF-β1 were produced by any of the populations under these test conditions (data not shown).
Proliferation, cytokine production, and suppressor functions of NnTregs. (A) NnTregs and other CD4+ T cell subsets were isolated by FACS cell sorting and stimulated with plate-bound anti-CD3/CD28 antibodies in the presence of APCs. Cell proliferation was assessed on day 7 after stimulation. Results of duplicate cultures are shown for 2 independent donors. (B) T cell subsets were stimulated by PMA/ionomycin. The concentration of cytokines in the culture supernatants was assessed by ELISA 24 hours after stimulation. Results of duplicate cultures are shown for 2 independent donors. No detectable cytokines were produced by any of the populations in the absence of stimulation (data not shown). (C) Ex vivo–sorted NnTreg and Treg populations were cocultured with responders (CFSE-labeled CD4+CD25– T cells) in the presence of PHA at the indicated suppressor/responder cell ratio. As internal control, CFSE-responders were cocultured with unlabeled responders or with the indicated suppressor population in transwell plates. The growth of CSFE-labeled responders was assessed by flow cytometry at day 5 after stimulation. Results are shown as mean values from 3 independent donors.
To assess the suppressor capacity of NnTregs with respect to the previously defined Treg subset, ex vivo–sorted NnTreg and Treg populations were cocultured with responder CD25– T cells labeled with 5- and 6-carboxyfluorescein diacetate succinimidyl ester (CFSE), in the presence of phytohemagglutinin (PHA) and autologous APCs. On day 5 after stimulation, cell division was assessed by measurement of CFSE dilution (Figure 4C). A clear inhibition of responder T cell expansion was detected upon coculture with NnTregs, comparable to that obtained with Tregs. Consistent with previous reports on the mechanisms of suppression by naturally occurring Tregs, inhibition of responder T cell proliferation by both Tregs and NnTregs involved cell-cell contact mechanisms, as no suppression occurred when the experiment was similarly performed using transwell culture plates, in which responder and suppressor populations are separated by a permeable membrane that prevents cell-cell contact but allows the passage of soluble factors (Figure 4C).
Proliferative capacity of NnTregs and assessment of in vitro–stimulated populations. To further examine the relative proliferative capacity of NnTregs, sorted populations from each subset were labeled with CFSE and stimulated with PHA or with anti-CD3/CD28 antibodies and irradiated autologous APCs either in the absence or presence of recombinant IL-2 at different concentrations. On day 5 after stimulation, cell division was assessed by measurement of CFSE dilution (Figure 5A). Cell growth was also evaluated by assessing total recovery of living cells and percentages of undivided cells in the cultures (Figure 5, B and C). Growth of CD25– naive and antigen-experienced T cell populations varied depending on the mode of stimulation and was higher for naive cells in the case of stimulation with PHA and for antigen-experienced cells in the case of stimulation with anti-CD3/CD28 antibodies. In the absence of IL-2, only little proliferation was observed for both NnTregs and Tregs under both stimulation conditions. However, in the presence of IL-2 at intermediate and high doses (10 and 100 IU/ml, respectively), significantly increased growth was detected for NnTregs as compared with Tregs after stimulation with both PHA and anti-CD3/CD28 antibodies.
Growth of CD4+ T cell subsets following stimulation with PHA or anti-CD3/CD28 antibodies in the absence or presence of IL-2 at different doses. Ex vivo–sorted CD4+ T cell subsets were labeled with CFSE and stimulated with PHA or with anti-CD3/CD28 antibodies, autologous CD4–CD8– APCs, and IL-2 at the indicated doses. (A) Example of data obtained for 1 donor following stimulation with PHA. Cell division was measured at day 5. The dotted lines indicate CFSE intensity of undivided cells. Percentage cell recovery (B) and percentage of undivided cells (C) are shown as mean values from 3 independent donors.
The phenotypic characteristics of in vitro–stimulated NnTregs as compared with cells in other CD4+ T cell subsets were assessed at day 10 after in vitro stimulation with PHA or anti-CD3/CD28 antibodies in the presence of IL-2 (100 IU/ml). Following in vitro stimulation, all populations showed significant expression of CD25 (Figure 6A and Table 2). CD25 expression levels, however, remained significantly higher for Tregs and NnTregs as compared with both naive and antigen-experienced CD25– T cell populations. In addition, in contrast to ex vivo relative expression levels, CD25 expression levels in stimulated populations were higher in NnTregs than in Tregs. The proportion of HLA-DR–expressing cells was increased in all subsets following stimulation but remained relatively low in NnTregs, close to that found in stimulated cells from the naive CD25– fraction and much lower than that found in both Tregs and antigen-experienced CD25– stimulated populations. Similarly to naive CD25– T cells, a relatively high proportion of NnTregs downregulated CD45RA, particularly after stimulation with anti-CD3/CD28 antibodies. However, a proportion of NnTregs maintained the expression of CCR7, particularly after stimulation with PHA. In addition, most NnTregs maintained expression of CD62L. Stimulated Treg populations contained a higher proportion of CCR4-expressing cells as compared with antigen-experienced CD25– T cell populations. In addition, and remarkably, a much higher proportion of NnTregs expressed CCR4 as compared with other naive CD4+ T cells. In contrast, expression of CCR5 was slightly decreased in NnTreg and Treg populations as compared with the corresponding naive and antigen-experienced CD25– T cell populations. Expression of CTLA-4, TGF-β1 LAP, and GITR was assessed in day 10 poststimulation cultures either in the resting state or 48 hours following restimulation with anti-CD3/CD28 antibodies (Figure 6B). Interestingly, in resting populations, expression of GITR was lower in both Tregs and NnTregs as compared with their counterpart CD25– populations, whereas that of LAP was moderately increased, as was that of CTLA-4. However, following restimulation, the expression of these molecules was upregulated in all populations and particularly in Tregs and NnTregs.
Phenotype of CD4+ T cell subsets following in vitro expansion. Sorted CD4+ T cell populations were stimulated with PHA or with anti-CD3/CD28 antibodies in the presence of allogeneic irradiated feeder cells and IL-2 (100 IU/ml). Ten days after stimulation, aliquots of the different cultures were labeled with the indicated antibodies and assessed using a FACSAria cell sorter. Data analysis was performed using FACSDiva software. (A) Example of data obtained from 1 donor following stimulation with anti-CD3/CD28 antibodies. Data obtained from 3 donors are summarized in Table 2. (B) Expression of GITR, TGF-β1 LAP, and CTLA-4 was assessed 48 hours later in cultures that had undergone or not undergone restimulation with anti-CD3/CD28 antibodies. Numbers are the mean of fluorescence intensity values obtained from 3 donors.
Phenotype of in vitro–stimulated CD4+ T cell subsets defined ex vivo by expression of CD25 and CD45RA
We assessed the growing capacity of day 10 poststimulation cultures following further stimulation with anti-CD3/CD28 antibodies in the absence of IL-2. The results of these experiments, illustrated in Figure 7A, revealed that, consistent with their ex vivo anergic state, Tregs and NnTregs remained relatively refractory to TCR-mediated stimulation and proliferated significantly less as compared with T cell lines derived from the CD25– subsets. Following stimulation with PMA/ionomycin, lines derived from Tregs produced IFN-γ, although at lower levels than lines derived from antigen-experienced CD25– T cells. Lines derived from naive CD25– T cells produced low levels of IFN-γ, and NnTreg-derived lines did not produce any detectable IFN-γ (Figure 7B). As expected, lines derived from naive CD25– T cells produced higher levels of IL-2 as compared with lines derived from antigen-experienced CD25– T cells. Treg-derived lines produced lower amounts of IL-2 as compared with both naive and antigen-experienced CD25– derived lines, and NnTreg-derived lines produced very little IL-2. At variance with the data obtained ex vivo, production of both IL-4 and IL-10 by day 10 poststimulation cultures was clearly detectable. Both antigen-experienced CD25– T cells and, to a lesser extent, Treg-derived lines produced high levels of IL-4, whereas lines derived from naive CD25– T cells produced lower but significant levels of IL-4 and NnTreg-derived lines produced very little IL-4. Production of IL-10 was significantly higher for Treg-derived lines as compared with lines derived from antigen-experienced CD25– T cells, undetectable in lines derived from naive CD25– T cells, and relatively low for NnTreg-derived lines. Production of TGF-β1 was below the limit of detection for all populations under these test conditions. Assessment of suppressive capacity of day 10 poststimulation cultures revealed a higher suppressive activity for NnTregs as compared with Tregs (Figure 7C).
Functional assessment of CD4+ T cell subsets following in vitro expansion. CD4+ T cell populations were stimulated with PHA or with anti-CD3/CD28 antibodies in the presence of allogeneic irradiated feeder cells and IL-2 (100 IU/ml). Cultures were assessed functionally 10 days after stimulation. (A) Aliquots of the cultures were labeled with CFSE and restimulated with anti-CD3/CD28 antibodies in the presence of APCs. Cell growth was evaluated at day 5 after stimulation. Percentage of undivided cells in the restimulated cultures along with the mean division cycles in the different populations are shown as mean values from 3 donors. (B) Cytokine production by in vitro–expanded populations was assessed 10 days after stimulation by measurement of their concentration in the culture supernatants 24 hours after stimulation with PMA/ionomycin. Results are shown as mean values from 3 donors. (C) Suppressor functions of CD4+CD25+ Treg populations expanded in vitro during 10 days in the presence of IL-2 were evaluated as detailed in Figure 4. Results are shown as mean values from 3 independent donors.
Reactivity of NnTregs and other CD4+ T cell subsets to autologous APCs. Information about the antigen specificity of naturally occurring Tregs is scarce. The current paradigm, mostly based on data obtained in TCR transgenic mice models, is that naturally occurring Tregs develop in the thymus due to TCR interactions with cognate self-peptide/MHC class II complexes within a restricted avidity range between positive and negative selection (21). Interestingly, a recent study by Romagnoli et al. (22) has demonstrated the presence, among murine Tregs, of precursors that can proliferate upon stimulation with autologous APCs. To gain initial insight into the antigen specificity of T cells in the NnTreg subset, we stimulated ex vivo–sorted and CFSE-labeled CD4+ T cells from the 4 previously defined subsets with autologous APCs, including monocytes and immature and mature DCs in the presence of IL-2. Cell growth was measured at day 7 after stimulation. As illustrated in Figure 8, no significant proliferation was observed in the absence of APCs for any subset. However, a small proportion of cells in the antigen-experienced CD25– T cell compartment and an increased proportion of cells among Tregs proliferated upon coculture with autologous monocytes. Similar results were obtained for these populations upon stimulation with both immature and mature DCs. No significant proliferation was detectable in the case of CD25– naive T cells upon culture with any of the APC populations. In sharp contrast, a high proportion of NnTregs proliferated in response to stimulation with autologous APCs, particularly in the case of both immature and mature DCs. Together these data demonstrate that, in the presence of IL-2, NnTreg populations react and proliferate in response to autologous APCs, indicating that this newly identified subset of circulating naive Tregs is highly enriched in T cells bearing self-reactive TCRs.
Reactivity of NnTregs and other CD4+ T cell subsets to autologous APCs. Ex vivo–sorted CD4+ T cell subsets were labeled with CFSE and stimulated with the indicated autologous APCs and IL-2 (100 IU/ml). Cell division was measured at day 7. (A) Example of data obtained from 1 donor. (B) Percentages of undivided cells are shown as mean values from 3 donors.