Myelin oligodendrocyte glycoprotein
Myelin basic protein
Regulatory T cell
Cord blood mononuclear cells
Glucocorticoid-induced TNF receptor family-related gene
Glutamic acid decarboxylase
- β2 m:
Regulatory T cells expressing CD25 have been shown to protect rodents from organ-specific autoimmune diseases. Similar CD25+ cells with a memory phenotype exerting suppressive function after polyclonal or allogeneic stimulation are also present in adult human blood. We demonstrate that adult human CD25+ cells regulate the response to myelin oligodendrocyte glycoprotein (MOG), as depletion of CD25+ cells increases responses of PBMC and the addition of purified CD25+ cells suppresses MOG-specific proliferation and IFN-γ production of CD4+CD25– T cells. In contrast, cord blood CD25+ cells do not inhibit responses to self antigens, and only a small subpopulation of cord CD25+ cells expresses the typical phenotype of adult regulatory T cells (CD45RA– and GITR+) enabling suppression of polyclonal responses. We conclude that activation of self-reactive T cells in normal healthy individuals is prevented by the presence of self-antigen-specific CD25+ regulatory T cells and that the majority of these cells mature after birth.
Since T cells are equipped with a randomly generated TCR, there must be mechanisms that ensure that T cells are activated in response to foreign but not to self antigens. Most T cells that encounter their specific antigen in the thymus are deleted by negative selection 1. Nevertheless, self-antigen-specific T cells do circulate in the periphery and can be activated in vitro and in vivo, although these autoreactive T cells often ignore their self antigen 2–4. The fact that autoimmunity occurs in many mouse models after depletion of T subpopulations clearly implicates that specific regulatory T cells (Treg) play an important role in preventing autoimmunity via an active "dominant" mechanism 4. Apart from experimentally induced suppressive T cells (e.g. in oral tolerance, in transplantation tolerance or Tr1 cells), it is well documented that mice and humans have so-called "naturally occurring" Treg that are characterized by the expression of CD25, CD152 (CTLA-4), CD122 and glucocorticoid-induced TNF receptor family-related gene (GITR) 4, 5. Some experimental models also point to the presence of CD25– Treg 6–8. CD25+ Treg comprise about 10% of murine CD4 T cells with essentially all CD25-expressing cells being regulatory although with different potencies 7, 9–11. In humans, only the brightest CD25-stained cells (CD25++, approximately 2% of all CD4 T cells) from peripheral blood function as Treg 12.
CD25 Treg are generated in the thymus, possibly through selection at a defined range of affinity, and they prevent various autoimmune diseases such as gastritis, colitis, diabetes and experimental autoimmune encephalomyelitis (EAE) 4, 6, 13–15. The prevention of organ-specific diseases indicates that the Treg must be specific for various autoantigens. However, studies directly demonstrating human Treg regulating responses to organ-specific self antigens are lacking. Most research on the function of Treg in vitro was conducted using polyclonal or allogeneic stimulation or murine TCR transgenic cells 4, 5, 9, 16, 17. These studies showed that Treg need to be activated through their TCR to exhibit suppressive function and that they suppressed CD4 or CD8 T cells in a cell-contact-dependent and possibly secretory cytokine-independent mechanism.
T cells from patients and from healthy volunteers react to antigens of the myelin sheath, e.g. myelin oligodendrocyte glycoprotein (MOG) and myelin basic protein (MBP) 3, 18, 19, and, although rarely, to the diabetes-associated self antigens insulin and glutamic acid decarboxylase (GAD) 65 20–22. The relative infrequent cases of overt autoimmunity, however, imply that these self-reactive T cells are normally kept under control in vivo by dominant tolerance mechanisms. We were thus interested to determine whether Treg specific for these antigens can be detected in healthy volunteers. Furthermore, as umbilical cord blood contains a discrete population of CD25+, CD122+, CTLA-4+ cells, we wanted to investigate whether those cells have suppressive function 23, 24. We, therefore, studied the responsiveness of adult PBMC and cord blood mononuclear cells (CBMC) to selected self antigens that are suspected to play an important role in the induction of multiple sclerosis (MBP, MOG) and diabetes (GAD, insulin) in the presence and absence of CD25+ cells. Furthermore, we employed purified populations of adult CD4+CD25– naive and memory T cells to show their reactivity to MOG and their suppression by adult CD4+CD25+ cells. Finally, the suppressive potential of cord blood CD25+ cells was tested using MOG as well as polyclonal stimulation.
2.1 Depletion of CD25+ cells increases reactivity of adult but not cord blood T cells
We first tested a variety of self antigens, food antigens and non-self microbial antigens for their capacity to elicit proliferative responses by PBMC and CBMC in the presence and absence of CD25+ cells. Among adult cells we detected responses to the extracellular part of MOG (MOGIgD) in 12/18 and against β-lactoglobulin (β-LG) in 8/18 volunteers. Both responses were significantly increased when CD25 cells were depleted (3.7±2.2-fold increase, p<0.001 for MOG and 3.5±3.2-fold increase, p<0.001 for β-LG), indicating the presence of MOG-reactive T cells in every adult. We did not detect responses to other self antigens, and responses to the recall antigens were not enhanced in the absence of CD25+ cells (Fig. 1). IFN-γ production correlated with proliferation, as depletion of CD25+ cells also increased IFN-γ production (data not shown).
Mononuclear cells from the various cord blood samples proliferated to multiple self antigens [MOG, β2-microglobulin (β2m), MBP, GAD] and some also to β-LG, tetanus toxoid (TT) and PPD. Responses were especially pronounced for MOG (7/8) and β-LG (6/8). However, no increase in the proliferation of cord blood cells was detected in the absence of CD25+ cells. In summary, CBMC reacted to more self antigens than adult blood PBMC, and CD25+ cells in adult blood but not cord blood cells suppressed the proliferation.
2.2 Memory and naive CD4 T cells from adult blood proliferate in response to MOG
To determine which cells within the PBMC population proliferated, CFSE-labeled cultures were analyzed by flow cytometry at day 6. CD4, CD8 and B cells proliferated in response to MOG (data not shown). Since primary responses are usually difficult to detect, we were interested to assess whether the MOG-specific cells had been primed and expanded. Purified CD4 T cells were sorted into naive (CD62L+CD45RA+) and memory (CD62L+CD45RA– and CD62L–CD45RA+/–) T cells as well as into CD25– naive and memory T cells, and cultured with antigen-pulsed dendritic cells (DC). While responses to TT clearly resided in the memory cell population, MOG elicited proliferation in both memory and naive T cells with counts being higher in the memory cells and highest in CD25– memory cells (Fig. 2A). IFN-γ production reflected the proliferation data for memory T cells, while naive T cells did not produce IFN-γ (Fig. 2B). Finally, memory TT responses were not altered by the depletion of CD25 cells.
2.3 MOG activates CD25+ T cells from adults and induces their suppressive function
The above-mentioned depletion data implicated an antigen-specific suppressive function of CD25+ Treg. First, we tested whether CD25+ Treg could recognize MOG. Purified CD25+ cells proliferated vigorously (Fig. 3A) and produced high amounts of IFN-γ (Fig. 3B) when stimulated with MOG and IL-2. In contrast, MOG alone did not induce proliferation of CD25+ cells, and the addition of IL-2 alone induced some proliferation in the absence of IFN-γ production (Fig. 3A, B). Next, we examined the suppressive activity of purified CD25+ cells. The addition of CD25+ cells to CD4+CD25– T cells at a 1:1 ratio led to a strong decrease in MOG-induced proliferation (in four independent experiments we obtained 62%, 83%, 74%, and 29% inhibition, respectively), while a 1:5 ratio resulted in less pronounced suppression (Fig. 3C). The reduced proliferation was not due to crowding in the wells, as doubling the number of CD4+CD25– T cells did not yield reduced proliferation (Fig. 3C). The suppressive effect of CD25+ cells was even more apparent when measuring IFN-γ production, which was almost completely abrogated by addition of CD25+ cells at both 1:1 and 1:5 ratios (Fig. 3D).
2.4 Cord blood CD25+ cells do not suppress antigen-specific responses
The data in Fig. 1 suggested that cord blood does not contain CD25+ cells with a potent suppressive function at the natural ratio of one CD25+ cell to nine naive T cells. Thus, we investigated whether antigen-specific suppression could be detected at a 1:1 ratio of cord blood naive T cells and cord blood CD25+ T cells. We were unable to see significant and consistent suppression of MOG-specific proliferation of cord blood CD4+CD25– T cells upon the addition of cord blood CD25+ cells (Fig. 4). The absence of suppression in the antigen-stimulated cord blood cultures could be due to a low frequency of antigen-specific CD25+ Treg or due to an immature population of Treg precursors in the newborn. In preliminary experiments we could not detect MOG-reactive Treg, as the addition of APC, MOG and IL-2 to purified cord blood CD25+ cells did not induce higher proliferation as the addition of APC and IL-2 alone (data not shown).
2.5 Cord blood CD25+ cells suppress polyclonal T cell activation
Since cord blood CD25+ cells are CD122+ and express intracellular CTLA-4 (CD152) 23, we were surprised that we could not detect suppression of MOG-induced proliferation. Thus, we assessed a potential suppressive function after polyclonal stimulation. The addition of cord blood CD25+ cells to cord blood CD25– cells at a 1:1 ratio in the presence of cord APC and PHA or staphylococcal enterotoxin B (SEB) led to suppression of proliferation up to 50%, similar to the suppression exerted by adult CD25+ cells (data not shown).
Reasoning that PHA/SEB stimulates CD4+CD25– T cells too quickly and too strongly in order to record more potent suppression by CD25+ cells, we tested whether cord blood CD25+ cells could suppress after anti-CD3 stimulation. As cord blood CD4+CD25– cells do not proliferate in response to anti-CD3 (data not shown and 25), we used adult CD4+CD25– cells as responder population. First, we determined whether anti-CD3 could activate the CD25+ cord blood T cells that were cultured with allogeneic adult APC. While cord blood CD25+ cells showed only a moderate allogeneic proliferation when cultured with adult APC and IL-2, the addition of anti-CD3 to this culture increased the proliferation fivefold, indicating that CD25+ cord blood cells are activated by anti-CD3 (Fig. 5A). Then, we added cord blood CD25+ cells to the adult CD4+CD25– cells in the presence of adult APC and anti-CD3. This led to a significant suppression at a 1:1 cell ratio (p<0.05), the degree of suppression decreasing with fewer cord CD25+ cells added (Fig. 5B, C). Furthermore, reduction of proliferation was identical to that induced by the addition of adult CD25+ cells (p<0.05). We conclude that cord blood CD25+ cells function as Treg after polyclonal stimulation and that cord blood CD4+CD25– cells can be suppressed.
2.6 Cord blood CD25+ cells are mainly naive and only few cells express GITR
Adult blood CD4+ cells can be divided into CD25– cells, CD25int cells (expressing CD25 at intermediate density) and very few CD25++ cells (expressing CD25 at very high density) (Fig. 6A). Previous work has shown that only the CD25++ cells function as Treg 12. Using anti-CD25-coated magnetic beads for the purification of CD25+ cells, one obtains a mixture of CD25int and CD25++ cells [mean fluorescence intensity (MFI) after purification: 67±13, n=8; before purification MFI CD25++ cells: 150±30, CD25int cells: 28±6, n=8; Fig. 6A], of which probably only the brightest 40% of the cells (that co-express CD122 and CD152) are actually suppressive. In contrast, cord blood lymphocytes contain only one population of CD25+ cells that express CD25 at an intensity between the adult CD25int and CD25++ cells (MFI before purification 70±17, n=10). After purification they are enriched for the brighter CD25+ cells (MFI after purification 89±23, n=10) (Fig. 6A). Since these cells purify as a homogenous CD25 population, we would have expected them to be more suppressive than adult CD25+ cells that are contaminated by CD25int non-suppressive cells. As they were rather equivalent in the degree of suppression they induced, we reasoned that only a subpopulation of cord blood CD25+ cells display Treg properties.
Phenotypic studies revealed that only an average of 14% (8.5–28%, n=9) of cord blood CD25+ cells co-express GITR, while an average of 83% (77–86%, n=3) of the adult CD25++ cells were GITR+ (Fig. 6B). Fig. 6C depicts the phenotype of the positively selected CD25+ cells that were used in the suppression assay. While an average of 70% (64–78%, n=3, depending on how much CD25int cells have co-purified) of the CD25+ cells from adult blood were GITR+CD45RA–, only 13% (8.5–18%, n=4) of the cord blood CD25+ cells displayed this phenotype. As only CD45RA–CD25+ cells suppress T cell proliferation 16, 26, we conclude that only a small subpopulation of cord blood CD25+ cells are functional Treg.
The prevention of organ-specific autoimmunity such as gastritis and diabetes via CD4+CD25+ Treg in mice indicated the presence of Treg recognizing organ-specific self antigens. However, Treg recognizing disease-initiating/driving organ-specific self antigens have not yet been directly demonstrated in vitro or in vivo. Another open question is the ontogeny of the maturation of Treg. It has been suggested that autoimmunity after neonatal thymectomy is due to the rarity of CD25+ cells before day 3 of life, although this is controversial 27, 28. An alternative explanation would be that neonatal CD25+ cells are immature. The present study addresses these issues using human peripheral blood and cord blood T cells.
We clearly demonstrate that proliferation of human CD4 T cells to MOG is controlled by MOG-specific Treg, and our data indicate that the same applies to β-LG but not foreign antigens. We could not detect responses to other self antigens in the adult blood, while cord blood cells responded to many self antigens (MOG, β2m, MBP, GAD) in addition to some foreign antigens (β-LG, PPD and TT). Importantly, CD25+ cells did not control the responses of cord blood cells to MOG, indicating that cord blood lacks functional self-antigen-specific Treg.
We confirm that MOG is the most potent inducer of T cell activation among the myelin proteins with 60% of normal healthy adult volunteers responding in our study (see also 3, 18, 19). However, importantly we demonstrate that the magnitude of proliferation increased after removal of CD25+ cells, which revealed MOG-specific proliferation along with increased IFN-γ production in all adult specimens tested. Interestingly, MOG was the only self antigen in our study to which responses were higher in adult blood than in cord blood. This can be explained by the fact that cord blood responses are primarily from naive T cells, while the responses in the adult were from naive and even more pronounced from memory T cells. Thus, APC seem to continually present MOG and expand MOG-reactive T cells. The location of this presentation is at present unknown, but we would speculate that it is in the CNS-draining LN as other self antigens are presented by DC in draining LN in the steady-state condition 29. However, most importantly we further demonstrate that CD25+ Treg are similarly activated against MOG and suppress CD4 effector T cell function. This clearly invokes Treg in the control of self-reactivity to a disease inducing self antigen. MOG-specific Treg may thus be instrumental in preventing multiple sclerosis in healthy individuals, a role that has very recently been demonstrated in the experimental model EAE 15.
Depletion of CD25+ cells increased T cell proliferation to MOG and β-LG, but not TT, PPD or Candida. A possible explanation would be that depletion of CD25+ cells not only depleted Treg but also some (Fig. 1) or all (Fig. 2) CD25int cells, which might contain TT-, PPD- or Candida-specific T cells. As the same argument can be applied to the self antigens, the magnitude of the suppressive function of Treg could have been underestimated in Fig. 1.
Taams et al. have recently reported that CD25+ cells suppress responses to self (human heat shock protein) and food antigens (cow milk antigens) as well as foreign antigens (TT and PPD) 26. While our data support their findings with regard to self antigen and food antigen, we disagree on the suppression of microbial foreign antigen. When we added FACS-sorted CD25++ cells to CD25– memory cells (1:1) and stimulated them with TT-pulsed DC, we did not detect reduction of proliferation (data not shown). Possibly, it is more difficult to suppress cultures stimulated by DC rather than by PBMC 12. Since in our hands depletion of CD25+ cells increased proliferation of PBMC in response to TT and PPD in some donors (and only a few donors were used in the Taams study), we would argue that foreign antigen-specific Treg are rare and might only be visible after recent contact with the antigen or if antigen is persistently present. In contrast, self-antigen-reactive Treg are continually restimulated and constantly suppress activation of self-reactive effector T cells. Similarly, as food antigens are constantly delivered to the immune system, Treg should be active against those antigens to prevent food intolerance.
Our second important observation is that cord blood T cells react towards many more self antigens than T cells from adults. In addition to the published IFN-γ production of CBMC in response to MBP, PLP, myelin-associated glycoprotein (MAG) and acetylcholine receptor 30, we show proliferation of PBMC to MOG, β2m, and GAD as well as to the food antigen β-LG. Previous studies revealed that 2% of cord blood samples contain anti-GAD Ab, and they were thought to be transferred from the (non-diabetic) mother as they disappear in the first year of life 31. According to our data, anti-GAD Ab could be produced in the infant itself and then disappear when the anti-GAD T cell responses wane due to the establishment of antigen-specific Treg. Indeed, we did not detect responses of adult PBMC against GAD. This is consistent with other studies reporting no GAD-reactive T cells in adults or only in a very small percentage of healthy individuals 20, 22, but contrasts with a report of vigorous responses to GAD65 in a longer culture period using exogenous IL-2 (which at the same time alleviates suppression)21. It is thus possible that responses to other self antigens can be uncovered after depletion of Treg combined with the use of more sensitive culture and readout methods.
Apart from a few exceptions, none of the responses in neonatal blood were enhanced after the depletion of CD25+ cells. In addition, we could not detect MOG-specific suppression using purified cord blood CD25+ cells. We believe that this can be explained by limited numbers of Treg that are specific for the antigen and that are mature enough to exert suppressive function.First, in preliminary experiments aimed at detecting proliferation of MOG-reactive cord blood CD25+ cells, we did not detect a significant higher proliferation in cells stimulated with MOGand IL-2 vs. IL-2 alone. This indicates that MOG-reactive CD25+ cells are very rare. Secondly, cord blood CD25+ putative Treg (as judged by the expression of CD122 and CTLA-4) express lower levels of CD25, are mainly CD45RA+, CD62Lhigh and CD38high23, 24, indicating a naive phenotype, and only few express GITR. Since among adult CD25+ cells only the CD45RA–CD45RO+ subpopulation suppresses 16, 26, we conclude that only a subpopulation of cord blood T cells is mature enough to suppress.
Polyclonal stimulation, in contrast, might activate sufficient Treg among cord blood cells to record suppression (similar as in adult cells that are a mixture of non/less-suppressive CD25int and suppressive CD25++ cells). This is consistent with the observation that allo-activation of CD4+ cord blood T cells yielded double-hit kinetics in limiting-dilution analysis, while CD4+CD25– cells had a single-hit kinetics which was suggestive (but did not prove) suppressive function of cord blood CD25+ cells 24. Alternatively, the polyclonal stimulation could mature cord blood CD25+ cells and thus induce a more potent activity than self antigen. Further experiments sorting the various subpopulations within cord blood cells should shed further light on the open issues. Also it will be very interesting to determine when in the postnatal life antigen-specific Treg can be detected and which factors areneeded for their maturation. This may also teach us how to activate Treg in patients that suffer from autoimmune disease and have reduced numbers of CD25+ cells in the blood 32.
How can we reconcile our cord blood data with reports that thymic CD25+ cells are fully functional suppressor cells 13, 17, 33, 34? First, as the rodent thymi are taken from adult animals and human thymic samples are usually from children of a few months of age, they may be more mature than the cord blood cells. Second, in vitro experiments have tested polyclonal or allogeneic stimulation but not natural autoantigen-specific suppression, and there is as yet no comparison of the relative potency between pure peripheral blood CD25++ Treg and thymic Treg. Finally, in the in vivo situation transferred CD25+ thymocytes can be further educated in the host 13.
We hypothesize the following steps for the maturation of functional Treg: First, autoantigen-specific CD25+ Treg are selected in the thymus by the self antigens that are presented on thymic epithelial cells 35, 36. Indeed, MOG is expressed in medullary thymic epithelial cells in the mouse 35 and in humans (personal communication B. Kyewski), and research with neoexpression of antigen in the thymus demonstrated selection of Treg in the thymus 14. Second, the autoantigen-specific Treg leave the thymus in a naive state 23 and are then restimulated (and possibly expanded) in the periphery by self-antigen-bearing DC in the draining LN. Upon this restimulation they convert to the memory phenotype and assume their suppressive function. The need of the continuous presence of self antigen has first been suggested by in vivo studies employing gender-specific autoantigens (reviewed in 2) and was then very elegantly proven by Mason's laboratory showing that thyroids need to be present for the survival/presence of thyroid-specific Treg 37. This mechanism of constant "confirmation" of Treg via recognition of self antigen in the periphery will ensure that the repertoire of Treg actually fits the need for the prevention of autoimmunity.
4 Materials and methods
4.1 Collection of blood
Peripheral blood from healthy adult donors was obtained by venous puncture and collected into preservative-free heparin. In addition, we received buffy coats from the local blood donor center.Cord blood from healthy full-term neonates was acquired immediately after delivery from the placenta and umbilical cord and collected in heparin (Leo Pharma, Sweden). The study was approved by the Human Research Ethics Committee of the Medical Faculty, Gothenburg University, Sweden.
4.2 Isolation of cells and flow cytometry
Lymphocytes were isolated as described previously 23. PBMC or CBMC were incubated with anti-CD25 magnetic beads for 15 min at 4°C (10 μl per 107 total cells; Miltenyi Biotech, Germany), and CD25+ cells were depleted according to the manufacturer's instructions. CD4+ cells were purified with Dynabead® CD4 Positive Isolation Kit (Dynal Biotech, Oslo, Norway). CD4+ cells were incubated with anti-CD25 beads (3 μl per 107 adult cells and 10 μl per 107 cord blood cells), and CD25– and CD25+ fractions were recovered. PBMC/CBMC were depleted of CD3+ cells (Dynabeads®), γ-irradiated (25 Gy) and used as APC. Naive (CD45RA+CD62L+) and memory T cells (CD45RA–CD62L– and CD45RA+CD62L–) including or excluding CD25+ cells were obtained from CD4+ cells by cell sorting using a FACSVantage SE (BD, San José, CA). To generate DC from the same volunteer, CD14+ cells were isolated 7 days earlier from PBMC with MACS microbeads and cultivated in medium supplemented with 800 U/ml GM-CSF (Leucomax, Molgramostim, Shering-Plough) and 500 U/ml IL-4 (R&D Systems Europe, Oxon, GB). After 5 days, fresh medium with 100 U/ml TNF-α and antigens were added, and the cellswere allowed to differentiate into mature DC during 2 more days of cultivation.
For flow cytometry we used mAb as described 23 plus purified polyclonal goat anti-hGITR, normal goat IgG (control Ab) and monoclonal PE-conjugated anti-hGITR (R&D Systems, Minneapolis, MN). To determine purity of sorted CD25+ cells, we employed PE-anti-CD25 (clone M-A251, PharMingen, San Diego, CA). PBMC were labeled with 5 μM CFSE (Molecular Probes, Leiden, The Netherlands) at 37°C.
The following antigens were used: recombinant human (rh) MOGIgD and β2m expressed in Escherichia coli (originally kind gifts from C. Linington, Max-Planck-Institute, Martinsried, Germany, and P. Travers, Birbeck College, London, GB, respectively) at 10 and 2 μg/ml; purified human MBP (Alexis Pharmaceutical, San Diego, CA) at 10 μg/ml; rhGAD65 (Diamyd Diagnostics, Stockholm, Sweden) at 10 μg/ml; rh-insulin expressed in E. coli at 50 μg/ml (Sigma-Aldrich, St. Louis, MO); purified bovine β–LG and purified chicken egg ovalbumin (OVA) at 50 μg/ml (both Sigma-Aldrich); TT at 10 μg/ml and PPD at 50 μg/ml (both Seruminstitute Copenhagen, Denmark); Candida albicans at 20 μg/ml (Greer Laboratories, Lenoir, NC).Endotoxin levels were measured for all antigens used and were below the concentration of LPS that gave proliferation of PBMC in control experiments. Further, addition of polymyxin B did not block MOG-(2.5 pg endotoxin/ml) or β-LG- (2.3 ng endotoxin/ml) induced proliferations.
4.4 Cell cultures
Total PBMC/CBMC or CD25-depleted PBMC/CBMC (2×105) were cultured for 6 days in AIM-V serum-free medium (Invitrogen, Life Technologies, Belgium) in the presence of the various self antigens in round-bottomed 96-well culture plates. Sorted naive or memory CD4+ T cells (5×105) were stimulated with antigen-pulsed autologous DC (1×104) in round-bottomed 96-well plates with 5% human AB+ serum and cultured for 5 days. For suppression of the MOG responses by CD25+ cells, 5×104 CD4+CD25– T cells were cultured with 5×104 CD4+CD25+ cells and 2×105 T-depleted irradiated PBMC in the presence of 10 μg/ml MOG for 6 days. To overcome anergy ofCD25+ cells, they were cultured with 10 μg/ml MOG in the presence of 2×105 T-depleted PBMC and 50 U rhIL-2 (PeproTech EC, London, GB). CD4+CD25– cord oradult blood-derived responder cells (10×103) were cultured with 10×103, 5×103, 2×103, or 1×103 autologous or allogeneic CD4+CD25+ cells in the presence of 1×105 irradiated T-depleted cells in flat-bottom 96 well plates in X-Vivo15 medium (Bio Whittaker, Walkersville, MD) and stimulated with 1 μg/ml anti-CD3 mAb (OKT-3, Ortho-McNeil Pharmaceutical, Raritan, NJ) for 4 days. At the end of the culture periods cells were pulsed overnight with 1 μCi [3H]thymidine (Amersham PharmaciaBiotech, Little Chalfont, GB). Statistical analysis was done with Wilcoxon signed rank sum test.
4.5 Cytokine quantification
Supernatants were collected after 48 h or at day 5 of culture and stored at –20°C until analysis. Purified and biotinylated anti-human IFN-γ Ab and rhIFN-γ standard were obtained from PharMingen, and the ELISA was performed according to manufacturer's instructions using SA-HRP (CLB, Amsterdam, Holland) for detection and TMB (Sigma-Aldrich) as substrate.
We thank the staff of Mölndal Hospital delivery unit for cord blood, the adult blood donors, and A. Tarkowski, E. Telemo (Göteborg, Sweden), B. Kyewski and B. Arnold (German Cancer Research Center, Heidelberg, Germany) for discussions and critical reading of the manuscript. This work was supported by grants from the Swedish Medical Research Council (71x-13487–02B), FRF Stiftelse and Konung Gustav V:s 80 Årsfond.