Suppressive properties of human CD4+CD25+ regulatory T cells are dependent on CTLA-4 expression



It has been demonstrated that T cells with regulatory properties are present within the peripheral blood CD4+CD25+ T cell compartment. Here, we describe an original method to purify human CD4+CD25+CD152+ T lymphocytes as living cells by forcing the exportation of CTLA-4 molecules stored in intracellular vesicules at the cell surface. By doing so, we demonstrate that CD4+CD25+ T cells contain a smaller and more homogeneous population enriched in cells with in vitro regulatory activity. Moreover, we show that this enrichment in regulatory T cells is associated with an increased expression of Foxp3 and that CD4+CD25+CD152+ T lymphocytes display a much stronger suppressive activity in controlling in vitro proliferation of alloantigen-specific T cells than CD4+CD25+CD152 T lymphocytes purified in parallel. Lastly, by purifying such cells expressing CTLA-4, we demonstrate that indeed CTLA-4 is involved in CD4+CD25+CD152+ T cell regulatory activity, while suppressive cytokines are not.


Graft-versus-host disease


CTL antigen-4


Glucocorticoid-induced TNF receptor






Reverse transcription

1 Introduction

CD4+CD25+ T cells, considered as naturally occurring regulatory T cells, have been described to intervene in the control of animal models of autoimmune diseases 14. Decreased frequency and deficient effector function of such peripheral blood CD4+CD25+ T cells has been described in human autoimmune diseases, i.e. type I diabetes and multiple sclerosis 5, 6. Additionally, recent data from murine studies have provided evidence that CD4+CD25+ T cells from the graft would contribute to the tolerance to alloantigens and allow the modulation of the graft-versus-host disease (GVHD) 79.

Stanzani et al. 10 have recently pointed out the limits of human natural regulatory T cell definition by usage of only two markers, CD4 and CD25. They investigated the potential correlation between the frequency of CD4+CD25+ T cells in the allograft and the development of GVHD. They concluded that GVHD patients had received donor grafts containing significantly more CD4+CD25+ T cells than non-GVHD patients. However, they probably counted not only regulatory cells but also pre-activated T lymphocytes, highlighting the importance of further characterizing CD4+CD25+ regulatory cells (reviewed 11). Foxp3 might be a much more specific marker for regulatory T cells, since it has been shown to be selectively expressed by naive mice and human CD4+CD25+ regulatory T cells 12, 13. However, no convenient tools for detection of its expression at the cellular level are yet available. Meanwhile, CTLA-4 (CD152), which has been described as consitutively expressed in murine or human regulatory T cells 2, 14, 15, could be considered as a suitable third marker. Even if its expression is induced by activation, it is quickly down-regulated on activated naive T cells, while its expression is stronger and remains longer on regulatory T cells 16, 17. Moreover, involvement of CTLA-4 in the regulation mediated by CD4+CD25+ T lymphocytes has been described on mouse models 2, 14. It was not confirmed by all 18 and was, at first, not verified on human CD4+CD25+ regulatory T cells 15, 17, 19. However, a recent report by Annunziato et al. 20 is adding some new insights into this controversial topic by demonstrating that anti-CD152 neutralizing antibodies can partially suppress CD4+CD25+ T cell regulatory activity.

In this report, we have described an original method to obtain purified CD4+CD25+CD152+ T lymphocytes as living cells. By doing so, we could demonstrate that human CD4+CD25+ T lymphocytes contain a smaller and more homogeneous population enriched in cells with in vitro suppressive activity. Interestingly, the regulatory T cell enrichment observed in the CD4+CD25+CD152+ subpopulation is associated with a strong increase in Foxp3 expression. Isolation of such cells expressing CTLA-4 has allowed us to demonstrate that indeed CTLA-4 is involved in their regulatory activity.

2 Results

2.1 Human CD4+CD25+CD152+ and CD4+CD25+CD152 T lymphocytes are phenotypically different

The phenotype of CD4+CD25+CD152+ and CD4+CD25+CD152 peripheral blood T cells defined by using the gates shown in Fig. 1 was investigated. This analysis revealed several differences. First, we observed that CD4+CD25+CD152+ T lymphocytes, which represent 6.3±2.1% of CD4+ T cells (n=20; data not shown), are almost devoid of CD45RA+ cells, while naive T cells represent 30% of CD4+CD25+CD152 T lymphocytes (Fig. 2). All CD4+CD25+CD152+ and CD4+CD25+CD152 T cells express CD45RB. However, CD4+CD25+CD152+ T cells include more highly differentiated CD45RBlow cells. CD4+CD25+CD152+ T lymphocytes are characterized by a significantly higher TNFRII expression than CD4+CD25+CD152 T cells (p=8.2×10–4). Expression of glucocorticoid-induced TNF receptor (GITR) was also investigated, and it was shown that less than 10% of cells were GITR-positive among either CD4+CD25+CD152+ or CD4+CD25+CD152 T lymphocytes. CD4+CD25+CD152+ T lymphocytes are characterized by a significantly higher expression of CD122 (IL-2 receptor β chain) (p=8.8×10–5). This variation of expression was not observed for the IL-2 receptor γ chain (data not shown). We have performed staining in order to measure the expression of membrane-bound TGF-β (or mLAP). None of the two populations expressed mLAP. Lastly, among various chemokine receptors investigated, only CCR6 and CCR4 were shown to be differentially expressed on CD4+CD25+CD152+ and CD4+CD25+CD152 T lymphocytes. While a small increase of CCR6 expression was detected on CD4+CD25+CD152+ T cells when compared to CD4+CD25+CD152 T cells, the difference in CCR4 expression was greater. Almost all CD4+CD25+CD152+ T lymphocytes express CCR4. In comparison, only half of CD4+CD25+CD152 T cells do express this receptor and with a lower intensity. This variation in CCR4 expression was significant, as was shown in six independent experiments (p=1×10–4). All the other markers tested (CD11a, CD27, CD28, CD38, CD44, CD49d, CD58, CD62L, CD69, CD70, CD71, CD95, CD103 and CD154) were not differentially expressed.

Figure 1.

Gating of CD4+ T lymphocytes and determination of their CD25 and CD152 expression. PBMC stained first with FITC-labeled anti-CD4 and PE-labeled anti-CD25 mAb or isotypic control were fixed with paraformaldehyde (PFA)-containing buffer. Fixed cells were then incubated with Cy-Chrome-conjugated anti-CD152 or mouse IgG2a mAb. CD4+CD25+CD152+ T lymphocytes were defined by their inclusion in gates R1, R2 and R3, while CD4+CD25+CD152 T lymphocytes belonged to R1 and R2 and R4.

Figure 2.

Phenotype of CD4+CD25+CD152+ and CD4+CD25+CD152 T lymphocytes. Staining of cell surface antigens was performed as described in Sect. 4. CD152 intracellular staining was then done after PFA fixation and permeabilization of the cells. The results are presented as histograms of fluorescence intensity of the antigen-specific mAb fixation (bold line) as compared to the isotypic control fixation (dotted line). The figures present one representative experiment. The frequencies of positive cells and the p values obtained when results of the six experiments were analyzed by using the paired one-sided Student's t-test are indicated.

Thus, phenotypic analysis has shown that CD4+CD25+CD152+ T lymphocytes are mainly memory T cells that can be differentiated from CD4+CD25+CD152 T lymphocytes by their higher level of expression of CCR4, TNFRII, CD122 and CCR6.

2.2 Treatment of CD4+CD25+ cells with ionomycin allows isolation of purified CD4+CD25+CD152+ T lymphocytes without modifying CD4+CD25+ T cell properties

CD152 is primarily an intracellular membrane protein, and its expression on the cell surface is barely detectable on resting T cells. To isolate living cells based on the co-expression of the three markers CD4, CD25 and CD152, we used an activation protocol derived from Linsley et al. 21. Purified CD4+CD25+ T lymphocytes were first stained for CD4 and CD25 expression. Cells were then incubated at 37°C with an anti-CD152 mAb in the presence of ionomycin (Io) in order to detect recycling CD152 and sorted (Fig. 3A). We routinely obtained 90 to 95% pure populations.

Figure 3.

Io treatment of CD4+CD25+ T cells allows purification of CD4+CD25+CD152+ T lymphocytes without influencing the properties of CD4+CD25+ T cells. (A) Purification of CD4+CD25+CD152+ and CD4+CD25+CD152 T cells. Enriched CD4+CD25+ T cells were stained with anti-CD4 and anti-CD25 mAb. Membrane expression of the CD152 molecule was then revealed by incubating the cells with anti-CD152 mAb at 37°C in the presence of 10 µg/ml Io. Such stained cells were sorted by using the indicated gates. The purity of the sorted populations obtained for one representative experiment out of eight is shown. (B) Io-treated CD4+CD25+ T lymphocytes express as much CTLA-4 as non-treated ones. Enriched CD4+CD25+ T lymphocytes were incubated in the presence or absence of Io. Intracellular staining for CTLA-4 expression was then performed. The staining of untreated CD4+CD25+ cells (shaded histogram) and Io-treated CD4+CD25+ cells (unshaded bold-bordered histograms) was compared to control staining (unshaded dotted-bordered histogram). The histograms are representative of eight experiments with similar results. (C) Io-treated and non-treated CD4+CD25+ T lymphocytes have the same suppressive activity. PBMC were co-cultured with allogeneic PBMC (S) in the presence or absence of CD4+CD25+ T cells or Io-treated CD4+CD25+ T lymphocytes (2:1 ratio). Proliferation was determined by [3H]thymidine incorporation. Data are means ± SD of triplicate determinations. Values with identical symbols differ significantly from each other. Results from one representative experiment are shown. (D) Io-treated and non-treated CD4+CD25+ T lymphocytes have the same ability to proliferate in response to polyclonal activation. Io-treated CD4+CD25+ and non-treated CD4+CD25+ T cells (5×102 each) were stimulated with immobilized anti-CD3 and soluble anti-CD28 mAb (1 µg/ml each).

The influence of Io treatment on the properties of CD4+CD25+ T lymphocytes was investigated. We first compared overall CTLA-4 expression on treated and non-treated CD4+CD25+ cells. Fig. 3B shows that intracellular staining of both cell types gave the same profile of CTLA-4 expression. Thus, Io does not induce de novo synthesis of CTLA-4, as was anticipated due to its intrinsic characteristics. Then, in order to determine whether or not Io could influence their functional properties, treated or non-treated CD4+CD25+ T cells were added to mononuclear cells stimulated by allogeneic irradiated PBMC (Fig. 3C). Both cell types induced a partial inhibition of the proliferation of allogeneic T cells, and the level of proliferation observed was not significantly different. These data demonstrate that pretreatment of CD4+CD25+ T cells with Io does not modify their suppressive capabilities. Lastly, we tested the potential of Io-treated CD4+CD25+ T cells to respond to polyclonal activation. As shown in Fig. 3D, there was no significantly different proliferative response after stimulation of CD4+CD25+ T lymphocytes, treated or not treated with Io, with immobilized anti-CD3 and soluble anti-CD28 antibodies. No difference in secreted cytokines was noted either (data not shown).

Thus, even though CTLA-4 is primarily an intracellular protein, its expression at the cell surface can be forced by treatment with Io. Such treatment does not modify human CD4+CD25+ T lymphocyte properties and is required to purify CD4+CD25+CD152+ T lymphocytes as living cells.

2.3 Human CD4+CD25+CD152+ T lymphocytes are enriched in regulatory T cells

Purified CD4+CD25+CD152+ cells were added to co-culture experiments to determine whether or not they have regulatory properties. PBMC were mixed with CD4+CD25+, CD4+CD25+CD152+ or CD4+CD25+CD152 T cells at a 2:1 ratio. Data from one experiment are shown in Fig. 4A. As shown earlier (Fig. 3C), addition of purified CD4+CD25+ T cells itself induced an inhibition of the proliferation of alloantigen-specific T cells. However, this inhibition was stronger when CD4+CD25+CD152+ T cells, at the same number, were used instead of non-purified CD4+CD25+ T lymphocytes. The level of inhibition due to CD4+CD25+CD152+ T cells varied from 89% to 93% (88±5%; n=5), while inhibition induced by CD4+CD25+ T cells varied from 38% to 68% (58±13%; n=4). The alloresponse was also reduced in the presence of CD4+CD25+CD152 T cells, meaning that the CD4+CD25+CD152 T cells contain regulatory T cells as well. However, the strength of the regulation induced by CD4+CD25+CD152 T lymphocytes was never as high as was observed with CD4+CD25+CD152+ T lymphocytes and reached 58±21% of inhibition (n=5).

Figure 4.

Purified CD4+CD25+CD152+ T cells demonstrate a stronger suppressive activity and higher Foxp3 expression than CD4+CD25+ T cells. (A) PBMC were co-cultured with allogeneic PBMC (S) in the presence or absence of CD4+CD25+ T cells, CD4+CD25+CD152+ T cells or CD4+CD25+CD152 T cells. Values with identical symbols differed significantly from each other in the unpaired one-sided Student's t-test. (B) PBMC and purified CD4+CD25, CD4+CD25+, CD4+CD25+CD152+ or CD4+CD25+CD152 T lymphocytes were prepared. cDNA samples obtained from these cell suspensions were tested for Foxp3 expression by real-time RT-PCR using 18S mRNA expression as reference for normalization. Results are expressed as Foxp3 expression relative to PBMC. Data from one representative experiment out of four are presented.

Recent reports have demonstrated that human T lymphocytes expressing Foxp3 have a regulatory phenotype 13, 22. We thus measured Foxp3-specific mRNA in CD4+CD25, CD4+CD25+, CD4+CD25+CD152+ and CD4+CD25+CD152 T cell preparations by quantitative reverse transcription (RT)-PCR and compared the results to the level of Foxp3 expressed in non-purified PBMC. High levels of Foxp3 mRNA were expressed in all three subpopulations: CD4+CD25+, CD4+CD25+CD152+ and CD4+CD25+CD152 T cells (Fig. 4B). However, its expression was considerably enriched in CD4+CD25+CD152+ T lymphocytes (52 times greater than in PBMC) as compared to CD4+CD25+ T lymphocyte preparations (19 times greater than in PBMC) and CD4+CD25+CD152 T lymphocyte preparations (eight times greater than in PBMC).

To further prove that CD4+CD25+CD152+ T cell preparations are enriched in regulatory T cells, suppressive assays under the condition of allogeneic stimulation using decreasing numbers of CD4+CD25+ or CD4+CD25+CD152+ T lymphocytes were performed. As shown in Fig. 5, the mean percentage of inhibition induced by CD4+CD25+CD152+ T cells was always higher than the inhibition induced by CD4+CD25+ or CD4+CD25+CD152 T cells, the ratio of responder to regulatory T cells varying from 2:1 to 32:1. However, the difference in the percentage of inhibition of proliferation was not anymore significant below a ratio of 8:1. Even considering this interval of significance (2:1 to 8:1), these data demonstrate that at least four times more non-purified CD4+CD25+ or CD4+CD25+CD152 T cells are needed to reach the suppressive activity due to sorted CD4+CD25+CD152+ T lymphocytes.

Figure 5.

Purified CD4+CD25+CD152+ T cell suspensions are more enriched in lymphocytes with regulatory activity than CD4+CD25+ T cell suspensions. The mean percent inhibition of the proliferative alloresponse induced by CD4+CD25+ (—▵—), CD4+CD25+CD152+ (—▪—) or CD4+CD25+CD152 (…□…) T cells in four different experiments was plotted. PBMC were co-cultured with allogeneic PBMC in the presence of decreasing numbers of autologous CD4+CD25+, CD4+CD25+CD152+ or CD4+CD25+CD152 T cells. p values, obtained when the percentages of inhibition due to CD4+CD25+CD152+ T cells versus CD4+CD25+ T cells and CD4+CD25+CD152+ T cells versus CD4+CD25+CD152 T cells (underlined) observed in the four experiments were subjected to Mann-Whitney test, are indicated.

Overall, these data demonstrate that the CD4+CD25+CD152+ T lymphocytes we are now able to purify after Io treatment have a higher suppressive activity than CD4+CD25+ T lymphocytes. This intense regulatory activity is correlated to an increased expression of Foxp3 and a higher frequency of cells with suppressive activity.

2.4 Suppressive cytokines are not involved in human CD4+CD25+CD152+ T lymphocyte regulatory activity, but CTLA-4 is

To further characterize the CD4+CD25+CD152+ T lymphocytes we compared their proliferative response after polyclonal stimulation. CD4+CD25+CD152+ T lymphocytes did not proliferate (Fig. 6A), as has been previously described for human and murine CD4+CD25+ regulatory T cells 15, 17, 19, 23. Conversely, CD4+CD25+CD152 T cells were not unresponsive. They proliferated in response to stimulation via their TCR, even though the proliferation detected was weaker than the proliferation of whole PBMC or purified CD4+CD25 T lymphocytes. The same results were observed when the proliferative assay was reduced to 3 days (data not shown). We then compared the cytokines produced by the various populations. Low amounts of IFN-γ, IL-10 and IL-4 were released by polyclonally activated CD4+CD25+CD152+ T lymphocytes (Fig. 6B). TGF-β was undetectable. On the contrary, IFN-γ, IL-10 and IL-4 were produced by CD4+CD25+CD152 T lymphocytes upon anti-CD3/anti-CD28 antibody stimulation. High amounts of IL-10 and some IL-4 were present in 7-day CD4+CD25+CD152 T cell culture supernatants, but were not yet detectable 3 days after the beginning of the stimulation (data not shown). Since CD4+CD25+CD152 T cells produced high amounts of IL-10 after polyclonal activation, we anticipated that this cytokine could be involved in the inhibition of proliferation induced by CD4+CD25+CD152 T cells. Thus, anti-IL-10 neutralizing antibodies were added to PBMC cultured in the context of an allogeneic stimulation. Even if a trend towards an increased proliferation of allogeneic PBMC was observed in wells containing anti-IL-10 mAb, no significant blocking of the inhibition induced by CD4+CD25+CD152 T cells could be demonstrated (Fig. 7A). Potential blocking of CD4+CD25+CD152+ T lymphocytes was also examined. As anticipated since no IL-10 was produced when CD4+CD25+CD152+ T cells were polyclonally activated (Fig. 6B), proliferation measured when anti-IL-10 antibody was added to the suppressive assay with CD4+CD25+CD152+ T cells was not modified at all (Fig. 7A). Targeting IL-4 or TGF-β with neutralizing antibodies was tested as well. Neither anti-IL-4 nor anti-TGF-β antibodies used were able to block all or part of both CD4+CD25+CD152+ and CD4+CD25+CD152 regulatory T cell activities (data not shown).

Figure 6.

CD4+CD25+CD152+ T lymphocytes are unresponsive when polyclonally activated. (A) PBMC, CD4+CD25, CD4+CD25+CD152+ or CD4+CD25+CD152 T cells (5×103 each) were stimulated with anti-CD3+anti-CD28 mAb. Proliferation was determined by [3H]thymidine incorporation after 7 days. One of eight independent experiments with similar results is shown. (B) Culture supernatant (50 µl/well) was removed 18 h before the end of the culture, and cytokines produced were measured by CBA technology (IFN-γ, IL-10 and IL-4) and ELISA (TGF-β). Results from one representative experiment are shown.

Figure 7.

Neither CD4+CD25+CD152+ nor CD4+CD25+CD152 T lymphocyte regulatory activity can be blocked by anti-IL-10 mAb, but regulatory activity of CD4+CD25+CD152+ T lymphocytes can be partially blocked by anti-CD152 mAb. (A) PBMC were co-cultured with allogeneic irradiated PBMC (S) in the presence or absence of autologous purified CD4+CD25+CD152+ or CD4+CD25+CD152 T lymphocytes and 10 µg/ml anti-IL-10 mAb or isotype-matched control. (B) PBMC were co-cultured with allogeneic PBMC (S) in the presence or absence of CD4+CD25+CD152+ or CD4+CD25+CD152 T cells and 10 µg/ml anti-CTLA-4 or isotype-matched mAb. For both (A, B), values with identical letters differ significantly from each other by Student's t-test. Results from one representative experiment are shown.

To investigate the potential role of CTLA-4 in CD4+CD25+CD152+ T cell regulatory activity, blocking experiments with anti-CTLA-4 mAb were also conducted. When anti-CTLA-4 mAb was added to PBMC activated by irradiated allogeneic cells in the presence of CD4+CD25+CD152+ T lymphocytes, the regulatory mechanism due to CD4+CD25+CD152+ T cells was partially, but significantly abolished (Fig. 7B). This effect induced by anti-CTLA-4 mAb was not observed if allogeneic stimulation of PBMC was performed in the presence of CD4+CD25+CD152 T lymphocytes.

In conclusion, the CD4+CD25+CD152+ T cells we were able to isolate after Io treatment are fully anergic and do not produce cytokines in vitro in response to polyclonal activation. However, they are characterized by a strong suppressive activity, and the mechanism they use to induce this suppression involves the CTLA-4 molecule.

3 Discussion

The data presented here demonstrate that regulatory T cells can be purified from human CD4+CD25+ T lymphocyte suspensions based on cell surface expression of CTLA-4. Such CD4+CD25+CD152+ T lymphocytes are anergic when stimulated via their TCR. They are characterized by a high level of Foxp3 expression and can develop a strong regulatory activity in vitro that involves at least partially CTLA-4. Thus, distinction of CD4+CD25+ T lymphocytes on the basis of the expression of a third marker, i.e. CTLA-4, allows to define a more homogeneous cell population enriched in regulatory T cells. We have also shown that even though the CD4+CD25+CD152 T lymphocyte subpopulation contains cells with suppressive activity, these T cells have, on the contrary, a weaker suppressive activity, express much less Foxp3 than CD4+CD25+CD152+ T cells, and we were not able to block this suppressive activity by targeting either CTLA-4, IL-10 or IL-4.

The existence of naturally occurring regulatory T cells and their suppressive function have previously been shown in mice 13, 23, 24 as well as in humans 1517, 19, 20. Sorting of CD4+ T cells based on the expression of the IL-2 receptor α chain (CD25) was, until now, a usual and convenient way to prepare suspensions enriched for cells with regulatory activity 1, 25. However, expression of CD25 is not an exclusive property of regulatory T cells and characterizes also activated T cells. Recent data from a correlation study between the frequency of CD4+CD25+ T lymphocytes in allografts and GVHD 10, in which Stanzani et al. enumerated most likely not only regulatory T cells but also pre-activated T lymphocytes, show us how confusing the sole usage of CD4 and CD25 could be. Moreover, Levings et al. 26 have indicated that even the small fraction of CD4+CD25bright cells is not a homogeneous population of suppressor cells. To overcome such confusion between regulatory and activated T cells, we thought that the addition of a third marker, i.e. CTLA-4, which had been shown to be constitutively expressed by a fraction of the human CD4+CD25+ regulatory T cells 1517 and whose expression is different on regulatory T cells or activated naive T cells 16, would give a more precise definition of the naturally occurring regulatory T cells.

To our knowledge, this is the first time that the suppressive activity of purified CD4+CD25+CD152+ T lymphocytes was investigated. CTLA-4 is localized mostly in lysosomes 27 and almost undetectable at the cell surface of resting T cells, rendering their purification impossible as such. However, when pre-enriched CD4+CD25+ T cells were incubated with anti-CD152 mAb in the presence of a calcium ionophore that can induce the degranulation of secretory vesicles without modifying CD4+CD25+ T cell properties, as we demonstrated (Fig. 3B–D), cells expressing CTLA-4 at their surface were labeled and could be sorted out. The purity of the cell suspensions was always high, but the efficiency of purifications was rather low with 0.3×106–1.4×106 cells obtained for 400 ml blood or for up to 8×108 PBMC processed (data not shown), limiting extremely the culture conditions we could test. The results we have provided show that purified CD4+CD25+CD152+ T lymphocytes indeed control the proliferation of alloantigen-specific T cells. Our experiments have also consistently shown that the CD4+CD25+CD152 T cell compartment contains regulatory cells as well. However, the suppressive activity of these cells was always weaker than the suppressive activity provided by CD4+CD25+CD152+ T cells used under the same conditions.

CD4+CD25+CD152+ T cells are much more efficient in suppressing alloantigen-specific T cell proliferation, because this population is enriched in regulatory T cells compared to non-purified CD4+CD25+ T cells, as was shown by doing a titration experiment (Fig. 5). And in complete agreement with these results, quantification of Foxp3-specific mRNA expression revealed that Foxp3 is considerably more expressed in CD4+CD25+CD152+ T lymphocytes than in non-purified CD4+CD25+ T lymphocytes. Thus, our data clearly demonstrate that CTLA-4 expression allows to define a more restricted population included in the CD4+CD25+ population and characterized by an increased frequency of cells with suppressive activity.

Investigation of the phenotype and the functional abilities of CD4+CD25+CD152+ T lymphocytes has shown that they are truly different from CD4+CD25+CD152 T lymphocytes. CD4+CD25+CD152+ T lymphocytes are highly differentiated memory T cells (CD45RA, CD45RBlow). The expression of CD122, TNFRII and CCR4 on CD4+CD25+ or CD4+CD25high regulatory T cells has previously been shown 19, 20, 28. We demonstrated that these markers are not only expressed by CD4+CD25+ T cells, but that their levels of expression are significantly higher on CD4+CD25+CD152+ T lymphocytes than on CD4+CD25+CD152 T lymphocytes. The high and homogeneous expression of CCR4 on CD4+CD25+CD152+ T lymphocytes is of particular interest and could enable these cells to preferentially migrate towards mature dendritic cells producing MDC/TARC, whereas regulatory T cells contained in the CD4+CD25+CD152 T cell compartment would be less efficiently attracted. The CD4+CD25+CD152+ T cells we have described might in fact be the CD25+ Ts cells that preferentially express CCR4 as shown by Iellem et al. 28.

GITR has been reported as preferentially expressed on murine CD4+CD25+ regulatory T cells, and its engagement by anti-GITR antibody abolished both in vitro and in vivo CD4+CD25+ regulatory T cell activity 29, 30. Higher expression of GITR-specific mRNA on human CD4+CD25+ T cells compared to CD4+CD25 T cells has been reported as well 22. By flow cytometry, we found that only a small fraction of human CD4+CD25+CD152+ or CD4+CD25+CD152 T lymphocytes expressed GITR (8% and 10%, respectively). A similar frequency of CD4+CD25+GITR+ T cells (15%) has recently been reported by Li et al. 31 for healthy controls, while it reaches 40% for patients with uveitis. On the contrary, our data showing only 8% of cells expressing GITR among CD4+CD25+CD152+ T lymphocytes do not confirm the direct correlation between CTLA-4 and GITR expression on the suppressive CD4+CD25+ T cell clones recently described 26. These differences might be due to experimental procedure: while our staining was performed on PBMC that were manipulated not more than was necessary for their isolation from blood, results reported by Levings et al. 26 were issued from cloned cells, and culture had most likely induced a bias towards some level of selection. The transcription factor Foxp3 might be a much more specific regulatory T cell marker since its expression has been correlated to regulatory activity even on human CD4+CD25+ T cells derived from activated CD4+CD25 T cells 13. However, convenient tools like Foxp3-specific antibodies fit for use in flow cytometry in order to quantify and follow CD4+CD25+Foxp3+ T cells are not yet available. Moreover, with Foxp3 being a nuclear protein, it will be impossible to purify Foxp3+ T cells in order to prove in vitro that they are able to control the proliferation of naive T cells.

Isolation of purified CD4+CD25+CD152+ T lymphocytes gave us the opportunity to prove that indeed these regulatory T cells use a mechanism involving a membrane protein, i.e. CTLA-4, implying the necessity of cell-to-cell contact at some point. To our knowledge, this is the first time that a CTLA-4-dependent mechanism was attributed to a precise subpopulation of CD4+CD25+ T cells. Annunziato et al. 20 have reported that CD4+CD25+ regulatory T cell suppressive activity could be almost fully abolished by a combination of anti-CTLA-4 and anti-TGF-β antibodies. On the contrary, Levings et al. 17, 26 have shown that targeting CTLA-4 via a neutralizing antibody had no effect on regulatory activity, whereas blocking of TGF-β decreased the CD4+CD25+ T cell regulatory effect. The experiments we performed in order to block the potential action of TGF-β in the regulatory mechanisms were unsuccessful for both CD4+CD25+CD152+ and CD4+CD25+CD152 T lymphocytes. A TGF-β-dependent mechanism of regulation has been attributed to regulatory T cells generated in the periphery 32, 33. Our in vitro model where whole PBMC were seeded in culture wells as responder cells, instead of purified CD4+ T cells used by others, might not have been the best model to visualize such induction of suppressive cells 32. However, by partially inhibiting CD4+CD25+CD152+ T lymphocyte regulatory activity with an anti-CTLA-4 antibody, our results show that CTLA-4 is directly involved in their suppressive mechanism. We have observed that CTLA-4 is expressed on CD4+CD25+CD152 T lymphocytes when they are polyclonally activated (data not shown), but anti-CTLA-4 antibodies were still not able to induce any inhibition of CD4+CD25+CD152 T cell regulatory activity. So, CTLA-4 expression in itself is not sufficient, and the CTLA-4-mediated suppressive mechanism is probably an intrinsic property of the CD4+CD25+ T cells constitutively expressing CTLA-4.

In vivo studies in murine models strongly implicate cytokines in regulatory mechanisms 3436. Some in vitro data obtained with human CD4+CD25+ regulatory T cells suggested also that IL-10 and TGF-β in particular would be involved in their suppressive mechanisms 20, 26, 32, 37. Our data demonstrate that purified CD4+CD25+CD152 T cells produce high amounts of IL-10 when they are activated by anti-CD3/anti-CD28 antibodies. The capacity of human CD4+CD25+ regulatory T cells to produce large quantities of IL-10 has previously been reported by others 1517. However, none of these studies has linked IL-10 production to the regulatory activity of these cells. Such an IL-10-mediated mechanism has been shown for other types of suppressive T cells: CD4+ T cells differentiated in vitro in the presence of IL-10, termed Tr1 cells 38, and anergized CD4+CD25 T cells 37. A trend towards partial abolition of the CD4+CD25+CD152 T cell regulatory activity could be observed in the blocking experiments we performed with anti-IL-10 neutralizing antibodies. However, significance could not be reached. In order to ascertain the IL-10 involvement, we will have to further investigate the CD4+CD25+CD152 T cell compartment and figure out which of these cells have regulatory properties.

In conclusion, the results of the experiments reported here on purified CD4+CD25+CD152+ T lymphocytes obtained as living cells demonstrate that human CD4+CD25+ T lymphocytes contain a smaller and more homogeneous population enriched in cells with in vitro suppressive activity. Further studies will be needed to better characterize the suppressor mechanisms induced by these CD4+CD25+CD152+ regulatory T cells, which involve at least CTLA-4.

4 Materials and methods

4.1 Antibodies

Most of the fluorochrome-conjugated mAb were purchased from BD PharMingen, except for the antibodies against CD49d (HP2/1) and CD103 (2G5) that were from Beckman Coulter and against mLAP (27232), TNFRII (22235.3), GITR (110416), CCR1 (53504.111), CCR5 (45523), CCR6 (53503.111) and CCR7 (150503) that were from R&D Systems. The goat anti-CCR8 polyclonal antibody was purchased from Alexis Biochemicals. For T cell activation, the anti-CD3 (UCHT1) and the anti-CD28 (CD28.2) mAb were purchased from BD PharMingen. The anti-CD152 (BNI3.1; BD PharMingen), anti-TGF-β1 (9016.2), anti-IL-10 (JV12) mAb and the goat anti-IL-4 polyclonal antibody (R&D Systems) were used for blocking experiments.

4.2 Isolation of T cell subpopulations

CD4+ T cells were prepared from healthy volunteers’ whole blood or cryopreserved PBMC obtained by leukopheresis by using RosetteSep™ or StemSep™ CD4+ T cell enrichment cocktail, respectively (StemCell Technologies). Purified CD4+CD25+ and CD4+CD25 T cells were then obtained by using anti-CD25 magnetic beads (Miltenyi Biotec). After staining with FITC-conjugated anti-CD4 and PE-conjugated anti-CD25 mAb, CD4+CD25+ T cells were incubated for 1 h at 37°C in the presence of Io (Sigma) (10 µg/ml) and Cy-Chrome-conjugated anti-CD152 mAb before isolation of the CD4+CD25+CD152+ and CD4+CD25+CD152 subpopulations was performed by sorting, using a FACSVantage SE (BD Biosciences).

4.3 T cell stimulations

PBMC (5×104 cells/well) were seeded in 96-well U-bottom plates and stimulated for 7 days with 5×104 allogeneic irradiated (2500 rad) PBMC in the presence or absence of various numbers of autologous CD4+CD25+, CD4+CD25+CD152+ or CD4+CD25+CD152 T cells, as indicated in the figure legends. For blocking experiments, 10 µg/ml of anti-CD152, anti-TGF-β, anti-IL10, anti-IL4 antibodies or appropriate isotypic controls were added. In some experiments, T lymphocytes were activated with immobilized anti-CD3 antibody: purified T cells (5×102–5×103) were distributed in 200 µl medium per well pre-coated with 1 µg/ml anti-CD3 mAb and cultured for 7 days in the presence of 1 µg/ml soluble anti-CD28 mAb. All cultures were performed in RPMI 1640 medium (Biochrom AG) supplemented with 10% pooled human serum, 100 U/ml penicillin/streptomycin (GIBCO), 2 mM L-Glutamine (GIBCO) and 1 mM sodium pyruvate (GIBCO). To determine the cell proliferation, wells were pulsed with 1 µCi/well [3H]thymidine (Amersham Pharmacia Biotech) for the last 16 h of culture, before harvesting and counting in a scintillation counter (Packard).

4.4 FACS analysis

Immunofluorescence staining for detection of surface antigens was performed by incubating cells for 20 min at 4°C with the optimal dilution of each mAb, except for recycling CD152. For determination of intracellular expression of CD152, cells already stained for CD4 and CD25 expression were fixed for 20 min at 20°C (Cytofix/Cytoperm kit; BD PharMingen). Intracellular staining was then performed at 20°C in saponin-containing perm/wash buffer and in the presence of Cy-Chrome- or allophycocyanin (APC)-conjugated anti-CD152 mAb. Samples were run on a FACSCalibur flow cytometer and analyzed using CellQuest software (Becton Dickinson). For enumeration of CD4+CD25+CD152+ T cells and determination of their phenotype, at least 2×105 events were analyzed.

4.5 Cytokine assays

Cytokine analysis was performed by quantification of cytokines released into supernatants harvested 7 days after stimulation. Production of TGF-β was determined by using a commercially available ELISA kit (Biosource Europe S.A.). IL-2, IL-4, IL-10, IFN-γ and TNF-α were measured by cytometric bead array (Th1/Th2 Cytokine CBA; BD PharMingen).

4.6 Foxp3 expression by real-time RT-PCR

Total RNA was prepared by using a ABI PRISM6100 Nucleic Acid PrepStation (Applied Biosystems, Foster City, CA), and single-stranded cDNA was synthesized (SuperScriptII, Invitrogen). Foxp3 mRNA expression was quantified by fluorescence-based real-time PCR (TaqMan technology, ABI PRISM 7700 Sequence Detector; Applied Biosystems) in a simplex assay using 18S ribosomic mRNA expression as reference for normalization. This assay has been described 39. Assay by demand kits for detection of human Foxp3 and 18S were purchased from Applied Biosystems.

4.7 Statistical analysis

Phenotypical differences observed in flow cytometry were analyzed by using the paired one-sided Student's t-test. Differences between the means of proliferative values obtained in [3H]thymidine incorporation assays were analyzed for significance by the unpaired one-sided Student's t-test. The Mann-Whitney test was used to evaluate the difference of CD4+CD25+ and CD4+CD25+CD152+ T cell inhibitory activity when plated at various ratios.


This work was supported, in part, by Ministère de la Santé (PHRC 2001, 2003), by Rennes University (BQR 2003) and by EFS


  1. 1


  2. 2


  3. 3


  4. 4


  5. 5


  6. 6


  7. 7