These authors contributed equally to the study.
CD4+CD25+FOXP3+ regulatory T cells from human thymus and cord blood suppress antigen-specific T cell responses
Article first published online: 27 JUN 2005
Volume 115, Issue 4, pages 516–525, August 2005
How to Cite
Wing, K., Larsson, P., Sandström, K., Lundin, S. B., Suri-Payer, E. and Rudin, A. (2005), CD4+CD25+FOXP3+ regulatory T cells from human thymus and cord blood suppress antigen-specific T cell responses. Immunology, 115: 516–525. doi: 10.1111/j.1365-2567.2005.02186.x
- Issue published online: 27 JUN 2005
- Article first published online: 27 JUN 2005
- Received 17 February 2005; revised 28 March 2005; accepted 30 March 2005.
- CD25+ regulatory T cells;
- cord blood;
- myelin oligodendrocyte glycoprotein;
Activation of self-reactive T cells in healthy adults is prevented by the presence of autoantigen-specific CD4+CD25+ regulatory T cells (CD25+ Treg). To explore the functional development of autoantigen-reactive CD25+ Treg in humans we investigated if thymic CD25+ Treg from children aged 2 months to 11 years and cord blood CD25+ Treg are able to suppress proliferation and cytokine production induced by specific antigens. While CD4+CD25− thymocytes proliferated in response to myelin oligodendrocyte glycoprotein (MOG), tetanus toxoid and beta-lactoglobulin, suppression of proliferation was not detected after the addition of thymic CD25+ Treg. However, CD25+ Treg inhibited interferon (IFN)-γ production induced by MOG, which indicates that MOG-reactive CD25+ Treg are present in the thymus. In contrast, cord blood CD25+ Treg suppressed both proliferation and cytokine production induced by MOG. Both cord blood and thymic CD25+ Treg expressed FOXP3 mRNA. However, FOXP3 expression was lower in cord blood than in thymic CD25+ T cells. Further characterization of cord blood CD25+ T cells revealed that FOXP3 was highly expressed by CD25+CD45RA+ cells while CD25+CD45RA− cells contained twofold less FOXP3, which may explain the lower expression level of FOXP3 in cord blood CD25+ T cells compared to thymic CD25+ T cells. In conclusion, our data demonstrate that low numbers of MOG-reactive functional CD25+ Treg are present in normal thymus, but that the suppressive ability of the cells is broader in cord blood. This suggests that the CD25+ Treg may be further matured in the periphery after being exported from the thymus.
In recent years naturally occurring CD4+CD25+ regulatory T cells (CD25+ Treg) have become established as important contributors to the maintenance of immunological self-tolerance. The majority of CD25+ Treg appear to be produced in the thymus because neonatal thymectomy of mice leads to multi-organ autoimmunity, which can be prevented by adoptive transfer of CD4+CD25+ T cells from normal mice.1 In addition, functional CD25+ Treg expressing CTLA-4, GITR and Foxp3 can be isolated from both the periphery and the thymus in mice as well as in humans.2–5 Mice lacking interleukin (IL)-2 or IL-2 receptor are deficient in peripheral CD25+ Treg6 and rats without a thyroid are devoid of thyroid-specific Treg.7 Thus, after leaving the thymus the survival of CD25+ Treg in the periphery depends on IL-2 and recognition of antigen. CD25+ Treg have a T cell receptor (TCR) repertoire as broad as CD25– T cells8,9 and they need antigen-specific stimulation via the T cell receptor to become suppressive, but once activated they suppress in an antigen non-specific fashion.10 The mechanism(s) by which CD25+ Treg suppress effector T cells remain to be fully determined. In the majority of human and murine studies the inhibition of T cell proliferation and cytokine production in vitro has been shown to be cell-contact-dependent.8,11 In addition, IL-10 and transforming growth factor (TGF)-β contribute to suppression in in vivo models of inflammatory bowel disease but not in models of gastritis.12
Functional studies of human adult peripheral blood have revealed that only CD4+ T cells with the brightest expression of CD25 are regulatory13 and the CD25+ Treg are confined to the memory T cell compartment as they express CD45RO.11 Because CD25+ Treg are not a discrete cell population and activated T cells and Treg could have overlapping CD25 expression levels, it is difficult to obtain pure CD25+ Treg from adult peripheral blood. In order to acquire a more homogeneous population of human CD25+ Treg devoid of recently activated/memory T cells we and others have investigated the potential use of CD25+ T cells derived from cord blood as a source of human CD25+ Treg.3,14–16 Cord blood CD25+ T cells are different from adult CD25+ T cells as they express markers of naive T cells, including CD45RA, CD38 and CD62L.14 Freshly isolated adult, cord blood and also thymic CD25+ T cells suppress T cell proliferation induced by polyclonal stimuli.3,13,16–18 Thus, the memory phenotype of adult CD25+ Treg is a sign of antigen experience rather than a characteristic of Treg. CD25+ Treg from adults have been shown to suppress T cell responses to autoantigens as well as exogenous antigens such as tetanus toxoid and allergens.3,19–21 However, little is known about how cord blood CD25+ Treg respond to specific antigens and no attempt has been made to study this in thymus. We have shown previously that cord blood CD25+ Treg are less able to suppress T cell proliferation induced by the neural self-antigen myelin oligodendrocyte glycoprotein (MOG) when compared to adult CD25+ Treg.3 It was unclear whether this observation was due to a lower frequency of antigen-specific CD25+ Treg in cord blood compared to adults or if they were functionally immature in the newborn child. To address this question we have analysed in this study the antigen-specific suppressive ability of human thymic CD25+ Treg compared to peripheral cord blood CD25+ Treg.
Materials and methods
Tissue collection and cell preparation
Thymus and 6 ml of peripheral blood were obtained from 13 children, aged 2 months to 11 years (median age 6 month), undergoing surgery for congenital heart defects. The peripheral blood was used for the preparation of antigen-presenting cells (APC, described below). A single-cell suspension from the thymus was obtained by gently passing the tissue through a 70 µm nylon sieve (Becton Dickinson, Le Pont De Claix, France). Cord blood from 16 healthy full term neonates was collected in heparin-containing tubes immediately after delivery. Blood lymphocytes were separated by Lymphoprep™ (Nycomed, Oslo, Norway) gradient centrifugation. Freshly isolated cells from each sample were examined for the capacity to proliferate and produce cytokines and the expression of FOXP3 was analysed in four thymic samples and nine cord blood samples. In addition, four samples of cord blood were used solely for the analysis of FOXP3 expression. Informed consent was obtained from the parents and the study was approved by the human Research Ethics Committee of the Medical Faculty, Göteborg University.
Isolation of cells
Thymic single positive CD4+ cells were obtained after depletion of CD8+ cells. Briefly, cells were incubated with CD8 antibody (Ab) (OKT-8) for 15 min at 4°C. Excess antibody was removed by centrifugation followed by CD8 cell depletion according to the manufacturer's instructions using Pan Mouse IgG (Dynal Biotech ASA, Oslo, Norway). Cord blood, and in some additional control experiments, also thymic CD4+ cells were purified with Dynabead® CD4 Positive Isolation Kit (Dynal Biotech ASA, Oslo, Norway). CD4+ cells were incubated with anti-CD25 beads and CD25– and CD25+ fractions were recovered using a LS-magnetic column according to the manufacturer's instructions (Miltenyi Biotech, Germany). The percentage of CD8+ thymocytes after CD8 depletion was less than 3% and remaining cells consisted of 70–90% CD4+ cells, with 10–30% CD4–CD8– cells. The purity after CD4+ selection were > 95% for both thymus and cord blood CD4+ cells; the purity of thymic and cord blood CD25+ T cells were routinely above 80% and 90%, respectively. CD25+CD45RA+, CD25+CD45RA– and CD25–CD45RA+ and CD25–CD45RA– were obtained from cord blood after initial separation with a Dynabead® CD4 Positive Isolation Kit (Dynal Biotech ASA, Oslo, Norway), followed by cell sorting using a FACSVantage SE (BD, San José, CA) after staining with FITC- anti-CD4 (SK3), APC- anti-CD25 (2A3) and PE-anti-CD45RA (HI100) from Becton Dickinson (Erembodegem, Belgium). Peripheral blood mononuclear cells (PBMC) or cord blood mononuclear cells (CBMC) were depleted of CD3+ cells (Dynabeads®), γ-irradiated (25 Gy) and used as APC. Purity of cell populations was determined by flow cytometry using the following monoclonal antibodies (mAb) obtained from Becton Dickinson (Erembodegem, Belgium): PerCP-anti-CD4 (SK3), APC-anti-CD3 (SK7) and PerCP-, APC- or phycoerythrin (PE)-isotype control IgG1, and PE-anti-CD25 (clone M-A251) obtained from Pharmingen (San Diego, CA).
Analysis of FOXP3 mRNA levels by quantitative PCR
Total RNA was extracted from 1 × 105 previously purified and frozen CD25– or CD25+ T cells using the RNeasy Micro kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions and contaminating genomic DNA was removed by treatment with DNA-free (Ambion, Austin TX). cDNA was prepared in a random hexamer-primed Superscript RT reaction (Invitrogen, Stockholm, Sweden) according to the manufacturer's protocol. FOXP3 mRNA levels were measured in duplicate with reagents from the LightCycler FastStart DNA Master SYBR Green I kit using the LightCycler instrument (Roche Diagnostics, Mannheim, Germany) and the following primers: 5′-CAG CAC ATT CCC AGA GTT CCT-3′ (forward) and 5′-CG TGT GAA CCA GTG GTA GAT-3′ (reverse). GAPDH was used as endogenous reference gene for relative quantification and detected by the primers: 5′-GGC TGC TTT TAA CTC TGG-3′ (forward) and 5′-GGA GGG ATC TCG CTC C-3′ (reverse). Primers were ordered from Tib Molbiol (Berlin, Germany) and designed not to amplify genomic DNA. A melting curve analysis was performed in each run to ensure specificity of the primers. Data was collected using LightCycler Data Analysis Software and the LightCycler Relative Quantification Software was used for calibrator normalized relative quantification.
The suppressive capacity of the cells was determined by culturing 200 × 103 CD4+CD25– T cells with 200 × 103 CD4+CD25+ cells together with 50 × 103 APC for 6 days in flat-bottomed 96-well plates in X-Vivo 15 medium (BioWhittaker, Walkersville, MD). Cultures were stimulated with 10 µg/ml recombinant human myelin oligodendrocyte glycoprotein extracellular region (MOGIgD abbreviated MOG), 50 µg/ml purified bovine beta-lactoglobulin (beta-LG) (Sigma-Aldrich, St Louis, MO) or 10 µg/ml tetanus toxoid (TT, Seruminstitute Copenhagen, Denmark). In addition, 10 × 103 CD4+CD25– responder cells were cultured with 10 × 103 CD4+CD25+ cells in the presence of 100 × 103 APC in flat-bottomed 96-well plates and stimulated with 0·2 µg/ml staphylococcal enterotoxin B (SEB, Sigma-Aldrich, Stockholm, Sweden) for 4 days. At the end of the culture period cells were pulsed overnight with 1 µCi [3H]-thymidine (Amersham Pharmacia Biotech, Little Chalfont, UK). Plates were harvested with a Filtermate 196-cell harvester (Packard) and [3H]-thymidine incorporation was measured with a Matrix 96 Direct Beta Counter (Packard). We accepted only experiments in which there was no overlap between the proliferation counts ± 2 SD from unstimulated and stimulated CD25– T cells.
Enzyme-linked immunosorbent assay (ELISA)
Supernatants were collected from cultures with specific antigens on day 6 and from cultures with SEB on day 4 of culture and stored at −20° until analysis. A standard ELISA protocol was performed as described elsewhere.22 All antibodies and standards were purchased from BD Pharmingen, San Diego, CA. Costar plates (Invitrogen, San Diego, CA) were coated with the following antihuman capture mAb: interferon (IFN)-γ (NIB42), IL-13 (JES10–5A2) and IL-10 (JES3–9D7). Standard curves were generated using recombinant human IL-13, IFN-γ and IL-10, respectively. The following biotinylated mAb were used: IFN-γ (4S.B3), IL-13 (B69-2) and IL-10 (JES3–12G8). Samples, standards, biotinylated antibodies and streptavidin–horse radish peroxidase (HRP) were diluted in high-performance ELISA dilution buffer (Sanquin, Amsterdam, the Netherlands). Some samples were analysed using the Biocore Multiplex system (Biorad, Hercules, CA) according to the manufacturer's instructions.
Data were analysed by Wilcoxon's signed-rank test except for data in Fig. 5, which were analysed by a two-tailed Mann–Whitney U-test using GraphPad Prism 3·00 (GraphPad, San Diego, CA). P-values ≤ 0·05 were regarded as significant, *P < 0·05, **P < 0·01, ***P < 0·001.
Thymic CD25+ Treg do not suppress antigen-specific T cell proliferation
We first investigated if CD25+ Treg directly isolated from human thymus were able to inhibit antigen-specific proliferation of thymic CD25– T cells after developing an optimized proliferation assay for thymic cells. Figure 1 shows the level of proliferation of freshly isolated thymic CD4+CD25– T cells after stimulation with myelin oligodendrocyte glycoprotein (MOG), beta-lactoglobulin (beta-LG) and tetanus toxoid (TT) in the presence of autologous peripheral antigen-presenting cells (APC). Although the counts are generally low, as expected for a primary antigen-specific response, they are specific as cell culture wells with twice the number of T cells gave a twofold higher proliferative response (Fig. 1d–f), and cultures without the antigen resulted in only background proliferation. Only experiments in which there was no overlap between the proliferation counts ± 2 SD from unstimulated and stimulated CD25– T cells were included in the study. Cultures of CD4+CD25+ T cells alone did not induce proliferation to the chosen antigens. The addition of an equal number of CD25+ Treg to the CD25– T cell cultures resulted in suppression of proliferation in only a minority of individuals after MOG (4/13) and TT (1/8) stimulation and no suppression at all was detected after stimulation with beta-LG (Fig. 1a–c). The suppression of proliferation to MOG and TT was detected in different individuals. The level of proliferation or the ability to suppress proliferation was not correlated with the age of the child. In summary, there was no significant suppression of proliferation to any of the antigens when the results of the individuals were summarized (Fig. 1d–f).
Thymic CD25+ Treg suppress antigen-specific IFN-γ production
We measured the production of IFN-γ, IL-13 and IL-10 from the thymic T cell cultures after antigen-specific stimulation with MOG, beta-LG and TT. First, it should be noted that not all thymic cultures responded with cytokine production. IFN-γ was detected in six of 13 thymi after MOG stimulation (Fig. 2a) and in three of eight thymi after beta-LG stimulation (Fig. 2c). The cytokine production was not correlated with the level of proliferation. Furthermore, cytokine production was not detected after TT stimulation despite the occurrence of a proliferative response to this antigen. In Fig. 2 we show the absolute levels of IFN-γ and IL-10 by CD25– T cells, CD25+ T cells and in the cocultures of the cell fractions after stimulation with MOG and beta-LG. We found that CD25– T cells produced both IFN-γ and IL-10 but not IL-13 after stimulation with MOG or beta-LG, and that thymic CD25+ Treg produced low levels of IL-10 but no detectable levels of IFN-γ or IL-13 (Fig. 2a–d). Cytokine production was not detected in control wells with APC stimulated alone (data not shown). Interestingly, despite the lack of suppression of proliferation, thymic CD25+ Treg potently suppressed IFN-γ production by CD25– T cells in the cocultures after stimulation with MOG or beta-LG (Fig. 2a,c). With the exception of one individual stimulated with MOG, suppression of IFN-γ was not reflected by corresponding suppression of proliferation. In five of seven donors, IL-10 was detected at lower levels in the coculture compared to CD25– T cells cultured alone after MOG stimulation (Fig. 2b). The capacity of CD25– T cells to produce cytokines in response to the antigens did not improve with increasing age of the children (data not shown). In summary, when MOG and beta-LG were able to induce IFN-γ production from thymic CD25– T cells, the CD25+ Treg irrespective of the donors' age potently suppressed the response, which suggests that human antigen-specific CD25+ Treg are present and functional in the thymus.
Cord blood CD25+ Treg suppress MOG-induced proliferation and cytokine production
In our earlier work we did not detect suppression of MOG-induced proliferation using cord blood CD25+ T cells. However, we did not analyse cytokine production, which seems to be a more sensitive way of measuring suppression than is proliferation. Therefore, we reinvestigated the capacity of cord blood CD25+ T cells to suppress responses by cord blood CD25– T cells to MOG. We employed the same suppression assay as for the thymocytes, using 200 000 T cells to increase the chance of obtaining sufficient numbers of MOG-reactive effector T cells and CD25+ Treg. In Fig. 3 we show the proliferation of CD25– T cells, CD25+ T cells and of cocultures of both cell fractions after stimulation with MOG. CD25– T cells proliferated well in response to MOG, while CD25+ T cells where hyporesponsive. In contrast to thymic CD25+ T cells, we found that cord blood-derived CD25+ Treg significantly suppressed the proliferation of CD25– T cells stimulated with MOG (Fig. 3a,b). The analysis of cytokines in the cultures revealed that cord blood CD25– T cells from all individuals produced IFN-γ, IL-13 and nine of 12 also produced detectable amounts of IL-10 (Fig. 3c–e). In contrast, CD25+ T cells did not produce detectable levels of IFN-γ and IL-13, while IL-10 was measurable. Cord blood CD25+ Treg were able to inhibit the production of IFN-γ and IL-13, but not IL-10 in the coculture wells (Fig. 3c–e). In summary, cord blood CD25+ Treg from nine of 12 individuals tested were able to suppress both proliferation and IFN-γ and IL-13 production induced by MOG in vitro.
Thymic CD25+ Treg suppress polyclonal responses
Because we did not detect suppression of proliferation to specific antigens in thymus, and others have shown that CD25+ Treg from thymus inhibit proliferation to phytohaemagglutinin (PHA) and alloantigens, we decided to test the suppressive ability of thymic CD25+ Treg after polyclonal stimulation. Indeed, CD25+ Treg from thymus significantly suppressed proliferation induced by staphylococcus enterotoxin B (SEB, Fig. 4a). We also found that the production of IFN-γ, IL-13 and IL-10 after SEB stimulation was significantly suppressed (Fig. 4b–d). In contrast to thymic T cells, stimulation of cord blood T cells with SEB resulted in vigorous proliferation of the CD25– T cells, which was not inhibited by the CD25+ T cells when added in a one-to-one ratio (Fig. 4e). In fact, CD25+ T cells also proliferated, although to a lower extent than CD25– T cells cultured alone. A similar lack of suppression was observed for the production of cytokines in the cord blood cultures. CD25+ T cells did not significantly suppress IFN-γ, IL-13 or IL-10 in cord blood after SEB stimulation (Fig. 4f–h). Interestingly, CD25+ Treg produced similar amounts of IL-13 compared to CD25– T cells, while the levels of IFN-γ and IL-10 remained significantly lower in the cultures of CD25+ Treg than in those of CD25– T cells. In summary, cord blood CD25+ Treg are not able to suppress proliferation or cytokine production induced by SEB, while CD25+ Treg from thymus have this capacity. One possible explanation for the discrepancy in suppressive ability could be that the cord blood CD25+ T cell population is contaminated by effector T cells in contrast to the thymic CD25+ T cells. In order to investigate this we decided to study the expression of FOXP3.
CD25+ Treg in cord blood express less FOXP3 than CD25+ Treg from thymus
We analysed the expression of FOXP3 mRNA by quantitative reverse transcription-polymerase chain reaction (RT-PCR) and compared the levels in CD25+ T cells from thymus and cord blood. Figure 5a shows histograms of the CD4+CD25– and CD4+CD25+ T cell fractions from thymus and cord blood after magnetic bead separation and Fig. 5b depicts their expression of FOXP3. We found that CD25+ T cells from cord blood expressed on average half as much FOXP3 mRNA as thymic CD25+ T cells. This may indicate that the CD25+ T cell fraction in cord blood contains a lower proportion CD25+ Treg than do purified thymic CD25+ T cells. We have shown previously that CD25+ T cells from cord blood can be divided further into one population of cells expressing intracellular CTLA-4 and CD122 and one population of cells not expressing these markers.14 Because CD25+CTLA-4+ T cells are mainly CD45RA+ and CD25+CTLA-4– T cells are CD45RA– we investigated whether the CD25+CD45RA+ cells were the major expressors of FOXP3 mRNA. Cord blood CD4+ T cells where sorted by FACS according to the expression of CD25 and CD45RA (Fig. 5c). We found that CD25+CD45RA+ T cells expressed twofold higher levels of FOXP3 than the CD25+CD45RA– T cells (Fig. 5d). Of note, the absolute expression levels of FOXP3 in cord blood in Fig. 5b,d should not be compared, as different purification techniques and individuals were used. Our data indicate that cord blood contains two different populations of CD25+ cells according to cell surface phenotype and FOXP3 expression. CD25+CD45RA+ cells express high levels of FOXP3 mRNA and thus most probably qualify as the dominant Treg population, while the CD25+CD45RA– cell fraction contains lower levels of FOXP3 mRNA.
During the last few years several studies have demonstrated the presence of CD4+CD25+ Treg in humans, which share striking similarities with their murine counterparts. Human CD25+ Treg capable of suppressing proliferation and cytokine production after various polyclonal stimuli have been isolated from peripheral blood, cord blood, tonsils and thymus.3,13,18,23 Despite the fact that CD25+ Treg are activated in an antigen-specific manner, few attempts have been made to investigate responses elicited by specific antigens in humans or normal mice. In this study we demonstrate for the first time that CD25+ Treg directly isolated from human thymus can respond to self-antigen by inhibiting the IFN-γ production of CD25– T cells. In addition, we show that this response is enhanced after migration to the periphery because CD25+ T cells in cord blood suppress not only IFN-γ and IL-13 production but also proliferation.
In previous reports, human thymic CD4+CD25+ Treg have been found to suppress proliferation and cytokine production induced by alloantigens and PHA.18,24 In accordance with these results, we show that CD25+ Treg derived from thymus are anergic and produce low or undetectable levels of cytokines, while they suppressed the proliferation as well as IFN-γ, IL-13 and IL-10 production by CD25– thymocytes after stimulation with SEB. In contrast, stimulation with specific antigens resulted only in suppression of proliferation in a minority of the tested individuals. It is conceivable that this is due to a low frequency of MOG-specific CD25+ Treg in the thymus, which makes suppression of proliferation harder to detect. However, thymic CD25+ Treg from individuals who responded with IFN-γ production suppressed the response after simulation with both MOG and beta-LG, which indicates that functional antigen-specific CD25+ Treg are indeed present in the thymus. In our experience suppression of IFN-γ can be more pronounced than suppression of proliferation, also when using Treg and Teff from adult blood.3 This discrepancy between suppression of proliferation and cytokine production is difficult to explain. Currently little is known about how CD25+ Treg mechanistically inhibit different T cell responses such as proliferation and production of various cytokines. However, it has been observed that CD25+ Treg from birch-allergic patients stimulated with birch allergen lack the ability to suppress Th2 cytokines during season while the ability to suppress proliferation was retained.21,25 This indicates that in a different system the production of a primary cytokine such as IL-2 may be regulated differently from the production of secondary cytokines such as IL-13 and IL-5.
In contrast to thymus, cord blood CD25+ Treg suppressed proliferation and production of IFN-γ and IL-13 by CD25– T cells stimulated with MOG but not SEB. Thus, autoantigen-specific Treg are present in both the thymus and cord blood. In a previous study cord blood CD25+ Treg were not found to suppress MOG induced proliferation.3 However, these experiments were aimed to compare the suppressive ability of cord blood CD25+ T cells and adult Treg and therefore only 50 000 T cells were used in the study. Thus, the lack of suppression of proliferation was due conceivably to the use of fourfold lower T cell numbers than employed in the current study. The increase in cell numbers allowed for a higher proliferation of the cord blood CD25– T cells as well as the detection of MOG-induced suppression of proliferation and cytokine production by the CD25+ T cells. Despite individual variations in responses, cord blood T cells generally responded with higher proliferation and cytokine production than thymic T cells. Together this suggests that the number of MOG-specific effector and regulatory T cells expands in the periphery of the child after being exported from the thymus.
Because cord blood CD25– T cells cannot be stimulated to proliferate by anti-CD3 mAb we used SEB for polyclonal stimulation.26 Interestingly, while thymocytes were suppressed easily by thymic Treg after SEB stimulation, cord blood CD25+ Treg did not suppress CD25– T cells after SEB stimulation and the CD25+ Treg even proliferated themselves and also produced IL-13 when cultured alone. In this regard cord blood CD25+ Treg resemble adult CD25+ T cells, which have also been reported to have diminished capacity to suppress proliferation induced by SEB.27 One could argue that SEB provides a very strong TCR signal that makes CD25– T cells refractory to inhibition by CD25+ Treg even if a low dose of SEB is used.27 In addition, neither adult nor cord blood CD25+ T cells were anergic when stimulated with SEB. In both studies CD25+ T cells were purified by magnetic bead separation and therefore non-regulatory CD25+ T cells might contaminate the CD25+ Treg fraction. Interestingly, the cord blood CD25+ fraction produced the Th2 cytokine IL-13 but not the Th1 cytokine IFN-γ after SEB stimulation, which would argue against a general contamination by effector T cells. It is, however, possible that CD25+ Treg from both cord and adult blood are contaminated selectively by Th2 cells as bead-separated CD25+ Treg from adult peripheral blood produce IL-4 when stimulated with PHA as opposed to thymocyte cultures.18
Recently FOXP3, which is a negative regulator of T cell activation, has been established as a marker for naturally occurring CD25+ Treg. Mice carrying a mutation in the Foxp3 gene display multi-organ autoimmune diseases and lack CD25+ Treg.2 Similarly, children born with a dysfunctional FOXP3 gene also succumb to organ-specific autoimmune diseases.28Foxp3 mRNA has been shown to be expressed by CD4+CD25+ T cells and thymocytes in both mice and human adults and newborns.2,4,5,15,16 We detected lower levels of FOXP3 mRNA in CD25+ T cells from cord blood than in CD25+ T cells from thymus after magnetic bead-purification. This observation made us ask whether only a subpopulation of the cord blood CD25+ cells might be expressing FOXP3. After sorting cord blood CD25+ T cells into either CD45RA+ or CD45RA– T cells by FACS we found that FOXP3 was predominantly expressed by CD25+CD45RA+ T cells. Cord blood CD25+CD45RA+ T cells constitute about two-thirds of the CD25+ T cells and correspond to cells with a regulatory and naive phenotype that express intracellular CTLA-4 and CD122. The remaining third of the CD25+ T cells are CD45RA–, CD62L,low, CD38low, CD122low, CTLA-4– and might constitute effector T cells.14 However, they still express FOXP3 even though at lower levels than the T cells with regulatory phenotype. Therefore, further studies sorting the CD25+ cells into other subpopulations may yield additional information as to where the regulatory cells concentrate. It is also possible that the level of FOXP3 mRNA does not directly correlate with the suppressive function of the cells.
In summary, these data suggest that autoantigen-specific CD25+ Treg are exported from the thymus in low numbers and in a naive state. In the periphery before birth autoreactive CD25+ Treg encounter their specific antigens in the draining lymph nodes and divide, as has been shown by murine studies of non-lymphopenic animals.29–31 Conceivably, after birth CD25+ Treg may need to become activated by antigen possibly in combination with other foreign agents for example bacterial antigens32,33 in order to be expanded further, become fully functional and be converted to a memory phenotype in adults. Further research is needed to provide information on how CD25+ Treg mature and expand in humans and what stimuli might enhance the function of CD25+ Treg in order to decrease the risk of developing autoimmune and allergic disease.
We thank the staff of The Queen Silvia Children's Hospital, Göteborg, for collection of thymus samples and the staff of Mölndal Hospital Delivery Unit for collection of cord blood. We are grateful to Christine Falk, GSF Munich, for her analysis of cytokines in supernatants using the Biocore multiplex system and to Bob Harris, Center for Molecular Medicin, Karolinska Institute, for providing human recombinant MOG. This work was supported by the Medical Faculty, Göteborg University, by the Swedish research council (grant K2002–745X-14455–01 A), by the Vårdal Foundation, by Clas Groschinsky's Memory fund and by the Göteborg Medical Society.