Transforming growth factor- b inhibits human antigen-speciﬁc CD4 + T cell proliferation without modulating the cytokine response

Transforming growth factor (TGF)- b has been demonstrated to play a key role in the regulation of the immune response, mainly by its suppressive function towards cells of the immune system. In humans, the effect of TGF- b on antigen-speciﬁc established memory T cells has not been investigated yet. In this study antigen-speciﬁc CD4 + T cell clones (TCC) were used to determine the effect of TGF- b on antigen-speciﬁc proliferation, the activation status of the T cells and their cytokine production. This study demonstrates that TGF- b is an adequate suppressor of antigen-speciﬁc T cell proliferation, by reducing the cell-cycle rate rather than induction of apoptosis. Addition of TGF- b resulted in increased CD69 expression and decreased CD25 expression on T cells, indicating that TGF- b is able to modulate the activation status of in vivo differentiated T cells. On the contrary, the antigen-speciﬁc cytokine production was not affected by TGF- b . Although TGF- b was suppressive towards the majority of the T cells, insensitivity of a few TCC towards TGF- b was also observed. This could not be correlated to differential expression of TGF- b signaling molecules such as Smad3, Smad7, SARA (Smad anchor for receptor activation) and Hgs (hepatocyte growth factor-regulated tyrosine kinase substrate). In summary, TGF- b has a pronounced inhibitory effect on antigen-speciﬁc T cell proliferation without modulating their cytokine production.


Introduction
A natural state of tolerance is maintained by multiple mechanisms, such as clonal deletion, anergy and the presence of regulatory T cells (1±4). Regulatory T cells, also called T r or T h 3 cells, mediate their suppressive effect mainly by secretion of IL-10 and transforming growth factor (TGF)-b (5,6). Many investigators have demonstrated the importance of the immunosuppressive cytokine TGF-b in the homeostasis of T cell regulation. TGF-b1 knockout mice develop severe multifocal in¯ammatory responses and die at 3±4 weeks of age (7). Evidence for a positive role of TGF-b in tolerance induction is demonstrated in studies where administration of TGF-b resulted in an improved clinical course in mice with experimental allergic encephalomyelitis and in delayed heart allograft reactions in rodents (8,9).
Although TGF-b can exert its effect on many different cell types, CD4 + T cells may be the main target since anti-CD4 antibodies are protective in TGF-b1 knockout mice (7). TGF-b exerts its effect on T cells via binding to the TGF-bRII, which recruits and phosphorylates TGF-bRI. TGF-bRI activates the ligand-speci®c Smad proteins (Smad2 and Smad3). Smad2 and Smad3 are kept in the cytoplasm where they are bound to the proteins SARA (Smad anchor for receptor activation) and Hgs (hepatocyte growth factor-regulated tyrosine kinase substrate) (10). Upon activation, Smad2 and Smad3 are released from SARA and Hgs, and are translocated to the nucleus where they activate TGF-b target genes. Recent studies showed that SARA and Hgs can act as attenuators of TGF-b responsiveness (11). The inhibitory Smad protein, Smad7, prevents activation and/or nuclear translocation of Smad2 and Smad3 (12). Disruption of TGF-b signaling in T cells was observed in mouse models after induction of Smad7 overexpression. This resulted in the development of a hyperreactive T cell response and antigen-induced airway in¯ammation after antigen inhalation (13). Moreover, in humans it was found that, in patients suffering from in¯ammatory bowel disease, Smad7 was up-regulated in mucosal T cells compared to T cells isolated from control subjects (14). These studies show that the effect of TGF-b was limited due to the high Smad7 expression in antigen-speci®c T cells, which resulted in the development of in¯ammation.
Diminished expression of TGF-bRII may also lead to unresponsiveness of T cells towards TGF-b. Studies with mice expressing a dominant-negative form of TGF-bRII showed a severe in¯ammation in the gut and lung, both sites where the immune system is constantly triggered by environmental antigens (15±17). This demonstrates that the expression of TGF-bRII is necessary in maintaining tolerance at mucosal sites. In human T cells it was recently shown that the expression of TGF-bRII could be modulated via IL-10 (18). IL-10 could enhance the expression of TGF-bRII, thereby increasing the suppressive effect of TGF-b. Thus, the responsiveness of T cells towards TGF-b depends on the differential expression of the Smad proteins as well as TGF-bRII at the time of antigen-speci®c stimulation.
The effects of TGF-b on established antigen-speci®c memory T cells have only been studied in murine models. TGF-b was shown to be suppressive towards murine T h 1 memory cells and to stimulate murine T h 2 cells (19). In humans, the effect of TGF-b was shown after mitogenic stimulation of T cells, isolated from healthy donors (20,21). However, the effect of TGF-b on human antigen-speci®c T cell responses has not been investigated yet. Therefore, in this study cow's milk-speci®c T cell clones (TCC) derived from in vivo differentiated peripheral blood T cells were used to investigate the effect of TGF-b on the T cell itself. These TCC were maintained in vitro by repeated stimulation in an antigen-speci®c manner. The proliferative capacity and cytokine pro®les of these TCC in the presence or absence of TGF-b was investigated. To examine the mechanism via which TGF-b could exert its function, analysis of the percentage of apoptotic cells, number of cell divisions and expression of T cell-surface markers was performed. In addition, the modulating effect of IL-10 on TGF-b induced inhibition of T cell proliferation was examined, as well as the correlation between sensitivity of T cells for TGF-b and mRNA expression of TGF-b signaling molecules.
Using an antigen-speci®c clonal CD4 + T cell model this report demonstrates that TGF-b is an effective suppressor of antigen-speci®c proliferation without affecting the cytokine production of the in vivo differentiated T cells. T cells derived from different donors (allergic and non-allergic) as well as T cells with different cytokine pro®les were equally inhibited by TGF-b in their antigen-speci®c proliferation. This model suggests a possible modulating role of TGF-b in T cellmediated disorders to suppress the T cell response.

Cow's milk-speci®c TCC
Peripheral blood mononuclear cells (PBMC) were isolated from venous blood of non-allergic children (mean age 4.5 years) and of age-matched children with a diagnosed cow's milk allergy (CMA), after informed consent was obtained. The study was approved by the Medical Ethical Committee of the University Medical Center, Utrecht. The diagnosis for CMA was con®rmed by positive cow's milk-speci®c IgE levels as determined by the CAP system FEIA (Pharmacia Diagnostics, Uppsala, Sweden), positive skin prick test and a positive double-blind placebo-controlled food challenge for cow's milk. The non-allergic children had normal total IgE levels and had a negative family history for atopy. Cow's milk-speci®c TCC were established from blood as described previously (22). The antigen used for T cell stimulation was a mix of cow's milk proteins (CMP), which consisted of equal quantities of total casein, a-lactalbumin and b-lactoglobulin (50 mg/ml) (gift from Dr R. Floris, NIZO Food Research, Ede, The Netherlands). In short, PBMC were cultured using an antigen-speci®c culturing system with irradiated autologous Epstein±Barr virus-transformed B cells (EBV-B cells), pre-incubated overnight with CMP, as antigen-presenting cells (APC). If these cultures indicated a high CMP-speci®c T cell proliferation in the lymphocyte stimulation test (LST), T cells were cloned by limiting dilution. Established TCC were tested in LST to determine speci®city for the various CMP and were restimulated every 14 days. The antigen speci®city and IL-4:IFN-g ratio of the TCC used in this report are presented in Table 1.  (23), and 50 U/ml IL-2 and IL-4 (a kind gift from Novartis Research Institute, Vienna, Austria). For the experiments in the presence and absence of TGF-b, IL-2 and IL-4 were omitted from the medium. All media were supplemented with penicillin (100 IU/ml), streptomycin (100 mg/ml) and glutamine (1 mM) (Gibco).

LST and cytokine production
The LST were performed in triplicate in 96-well U-bottom plates (Greiner, Frickenhausen, Germany H]thymidine, supernatant was collected from the triplicate wells and stored at ±20°C. Cytokine production was measured by ELISA according to the manufacturer's recommendations (for IL-4, IL-10 and IFN-g: Sanquin, Amsterdam, The Netherlands; and for IL-2: Diaclone, Besanc Ë on, France). The detection limit was 0.6 pg/ml for IL-4, 1.2 pg/ml for IL-10, 2 pg/ml for IFN-g and 15 pg/ml for IL-2.

Flow cytometry
To examine the effect of TGF-b on apoptosis,¯ow cytometry was performed with phycoerythrin (PE)-conjugated Annexin-V according to the manufacturer's recommendations (BD Biosciences, San Jose, CA). To examine the effect of TGF-b on T cell activation markers, antibodies to CD4, CD25 and CD69 (all PE-conjugated), and FITC-conjugated CD28 were used for¯ow cytometry (BD Biosciences). Expression levels of CD25, CD28 and CD69 by each TCC were measured as mean uorescence intensity (MFI) in the presence or absence of TGF-b. The effect of TGF-b on the expression of the costimulatory molecules on the B cells was measured with FITCconjugated CD80 and CD86 antibodies (BD Biosciences). The expression levels of CD80 and CD86 were measured as MFI after culturing of the EBV-B cells in the presence or absence of TGF-b. All the analysis was performed on a Becton Dickinson FACScan with WinMDI software.

Carboxy-¯uorescein diacetate succinimidyl ester (CFSE) proliferation assay
To analyze the number of cell divisions, CFSE was used at a ®nal concentration of 0.5 mM (Molecular Probes, Eugene, OR). T cells were resuspended at a concentration of 10 7 cells/ml in PBS and labeled with CFSE for 10 min at 37°C. Unbound CFSE was quenched by washing the cells twice in culture medium containing 10% FCS. Before co-culture, the labeled T cells were left in medium for 2 h at 37°C and washed once more to prevent non-speci®c labeling of the EBV-B cells. T cells were stimulated in a 1:1 ratio with the pre-incubated EBV-B cells (10 6 :10 6 ), and harvested after 72, 96 and 120 h of co-culture. For analysis, the cells were washed, stained with PE-conjugated CD4 and analyzed with a FACScan (BD Biosciences). The number of cell divisions was correlated with the decrease in MFI.

Statistical analysis
Statistical analysis of the effect of LAP on TGF-b-induced inhibition was performed with a Mann±Whitney U-test. Nonparametric paired analysis (Wilcoxon test) was performed to examine differences in the expression of cell-surface markers (CD25, CD28 and CD69), distribution over the cell cycles induced by TGF-b and differences in the expression of TGF-bsignaling molecules 96 h after stimulation. Differences associated with P values of less than 0.05 were considered signi®cant.

TGF-b inhibits antigen-speci®c T cell proliferation
To determine the optimal effective dose of TGF-b administered, the effect of several concentrations TGF-b on antigen-speci®c T cell proliferation was tested (Fig. 1A). The optimal dose for suppression appeared to be 1 ng/ml, which was used throughout the following experiments. TGF-b-induced inhibition of antigen-speci®c proliferation was most pronounced at 96 h after stimulation, therefore the results from 96 h are shown unless the results on earlier time points deviate from the result at 96 h (Fig. 1B). The effectiveness of tolerance induction has been demonstrated to be dependent on the dose of the antigen administered. Therefore it was examined whether TGF-b was able to suppress antigen-speci®c T cell proliferation in TCC stimulated with low (2 mg/ml) and high (50 mg/ml) doses of antigen. Since the low dose of antigen (2 mg/ml) results in an average reduction in proliferation of 55% (data not shown), this dose was considered to be a partial stimulation. As is shown in Fig. 1(C), the suppressive effect of TGF-b on antigen-speci®c proliferation was comparable between partial and full antigen stimulation of T cells. Addition of a TGF-b neutralizing peptide (LAP) in the culture system signi®cantly reversed the inhibitory effect of TGF-b (Fig. 1D, Mann±Whitney U-test, *P < 0.01). Inhibition of antigen-speci®c proliferation of TCC by TGF-b derived from both non-allergic as well as allergic donors was not signi®cantly different (mean percentage proliferation in the presence of 1 ng/ml TGF-b; non-allergic 38.5 T 7.0% and allergic 46.1 T 10.6%). Only two out of 17 TCC tested in an antigen-speci®c LST were not inhibited by TGF-b (Fig. 1E).

TGF-b inhibits cell cycle progression of antigen-speci®c T cells
To investigate the inhibitory mechanisms of TGF-b on the antigen-speci®c T cell proliferation, the percentage of apoptotic cells was examined. Annexin-V + cells were determined in the presence and absence of TGF-b at 24, 48, 72 and 96 h after stimulation. Addition of TGF-b appeared to have no signi®cant effect on the percentage of Annexin V + cells at all time points (Fig. 2A).
In contrast, the number of cell divisions, measured as a decrease in CFSE¯uorescence, was clearly diminished at 96 h after antigen-speci®c stimulation in the presence of TGF-b. Figure 2(B) shows that the T cells, which were stimulated in the absence of TGF-b, undergo more cell divisions compared to T cells stimulated in the presence of TGF-b (MFI = 120.38 versus 170.45 respectively, P < 0.05). The amount of cells in each cell division demonstrates that cells, cultured in the presence of TGF-b, are more abundant in the ®rst two cycles compared to cells cultured without TGF-b (29 and 37% versus 15 and 29%). In contrast, T cells cultured without TGF-b are more abundant in the later two cell cycles (21 and 13% versus 9 and 5% of the cells cultured with TGF-b).
TGF-b has a differential effect on T cell activation and co-stimulation markers (CD25, CD69 and CD28), but does not in¯uence the expression of CD80 and CD86 on APC In concordance with the demonstrated inhibition of antigen-speci®c proliferation, TGF-b signi®cantly decreased the expression level of the IL-2 receptor, CD25 (P < 0.05). In contrast, the expression level of CD69 was signi®cantly upregulated in the presence of TGF-b ( Fig. 3A and B). The effect of TGF-b on the expression of CD25 and CD69 was present at 24, 48, 72 and 96 h after antigen-speci®c stimulation (data not shown). To examine whether TGF-b directly affected the CD25 expression or indirectly via a down-regulation of IL-2 production, IL-2 production in the culture supernatant was measured using ELISA. Although the amounts of IL-2 production were very low, addition of TGF-b did not affect the production of IL-2 by T cells (mean production T SEM at 24 h: no TGF: 221 T 111 pg/ml, TGF: 212 T 101 pg/ml; at 96 h: no TGF: 22 T 2.6 pg/ml, TGF: 25 T 2.8 pg/ml). To analyze the possibility that the effect of TGF-b was mediated via changes in costimulatory molecules, the expression levels of CD28 on T cells, and CD80 and CD86 on the APC (EBV-B cells), were measured in the presence and absence of TGF-b. TGF-b did not affect the expression levels of CD80 and CD86 on the EBV-B cells. In contrast it did decrease the expression level of CD28 on T cells (P < 0.05) (Fig. 3C).

IL-10 has no enhancing effect on TGF-b induced inhibition of antigen-speci®c T cell proliferation
In Fig. 4(A) the effect of IL-10 and/or TGF-b on T cell proliferation is shown for three representative TCC. An additive inhibitory effect of TGF-b and IL-10 was only observed in TCC that were inhibited by IL-10 itself (TCC 9.6). The more pronounced inhibition found in TCC 9.6 seemed to be due to a cumulative rather than a synergistic inhibitory effect on proliferation of IL-10 and TGF-b. Most TCC were not inhibited in their antigen-speci®c proliferation by IL-10 (15 out of 17 tested). In these TCC, IL-10 did not enhance the inhibitory effect of TGF-b. It was examined whether addition of IL-10 could enhance the expression of TGF-bRII on antigen-speci®c T cells. Figure 4(B) demonstrates that mRNA levels of TGF-bRII of a representative CMP-speci®c TCC before and 96 h after antigen-speci®c stimulation with CMP were not affected in the presence of TGF-b or IL-10.

TGF-b does not affect cytokine production of antigen-speci®c T cells
To analyze the effect of TGF-b on cytokine production, supernatants were analyzed at 24, 48, 72 and 96 h of culturing with and without TGF-b. One representative TCC is shown to demonstrate that the production of cytokines during culture decreased over time (Fig. 5A). However, no effect on cytokine production was observed by TGF-b during the 96 h of stimulation. In Fig. 5(B), the production of IL-4 and IFN-g as well as the ratio of IL-4:IFN-g is shown for all TCC at 96 h after stimulation in the presence and absence of TGF-b. There is no signi®cant effect of TGF-b on the production of either IL-4 or IFN-g. Also, the production of IL-10 measured in the supernatant was not altered in the presence of TGF-b (Fig. 5C). There was no correlation between the inhibitory effect of TGF-b on the antigen-speci®c proliferation and the cytokine production observed.

Suppression of the antigen-speci®c T cell response by TGF-b 1499
mRNA expression of TGF-b signaling molecules is not related to susceptibility for TGF-b Two out of the 17 tested TCC were found to be unresponsive towards TGF-b-induced inhibition of proliferation. It was examined whether the decreased sensitivity of these TCC (TCC 9.7 and 2.2) for TGF-b was a result of an increased Smad7 expression (inhibitory Smad protein) or a decreased Smad3 expression (stimulatory Smad protein). Moreover, since it was recently demonstrated that a high expression of SARA and Hgs reduced the sensitivity of T cells towards  TGF-b (11), the mRNA levels of SARA and Hgs from TGF-bsensitive and -insensitive TCC were also analyzed. The mRNA expression of Smad3, Smad7, SARA and Hgs of several TCC was analyzed prior to stimulation, and showed no correlation with the sensitivity of TCC towards TGF-b (Fig. 6A). Antigen-speci®c stimulation of TCC in the presence of TGF-b resulted in a signi®cantly higher mRNA expression of SARA compared to TCC stimulated in the absence of TGF-b (Fig. 6B). Expression of Hgs also tended to be higher in the presence of TGF-b (not signi®cantly, P = 0.08). The expression levels of Smad3 and Smad7 were not signi®cantly different in the presence of TGF-b.

Discussion
TGF-b is an immunosuppressive cytokine, which has an important role in the maintenance of tolerance to self-antigens and environmental antigens (25). Disruption of TGF-b signaling in T cells or a diminished TGF-b production is involved in the development of several immune disorders (13±15,26). In murine, as well as human, disease models it was shown that induction of peripheral tolerance can be achieved via an enhancement of TGF-b production by CD4 + T cells (27±32). To characterize TGF-b as a potential therapeutic agent in human disease models it is necessary to evaluate the effect of TGF-b on differentiated human T cells.
To study the effect of TGF-b on individual CD4 + T cells, an antigen-speci®c model of cow's milk-speci®c TCC was used. This model provides the opportunity to analyze the effect of TGF-b in homogeneous T cell populations that are fully antigen-speci®c, without the disturbance of non-speci®c T cell reactivity. It was demonstrated that the suppressive effect of TGF-b is independent of the amount of antigen used and does not discriminate between T cells derived from nonallergic or allergic donors. The inhibition of the antigen-  speci®c proliferation by TGF-b was not correlated with an increase in apoptosis (Fig 2). In a study with mitogenstimulated human PBMC it was demonstrated that TGF-b could have a protective effect on apoptosis (33). This discrepancy could be explained by the use of mitogenic stimulation versus antigen-speci®c stimulation. However, another study with antigen-speci®c murine effector CD4 + T cells also showed the protective effect of TGF-b on apoptosis (34). In this study, antigen-speci®c T h 2 effector cells were generated from naive CD4 + T cells by polarization in vitro with the usage of IL-4 and anti-IFN-g. These effector T cells were generated in 4 days and cultured for short periods (7 days). The TCC used in our study were polarized and differentiated in vivo, and maintained in culture for longer periods (>30 days). The variation in generation and polarization may result in different stages of T cell differentiation. This may suggest that the effect of TGF-b on T cell apoptosis is associated with the differentiation status of the T cells. The present study is the ®rst to describe that, in an antigen-speci®c system where TGF-b clearly inhibits T cell proliferation, a pronounced difference is observed in the distribution of cells over the different cell divisions between T cells cultured with and without TGF-b (Fig 2). TGF-b has been shown in other cell types to inhibit proliferation via an up-regulation of cell-cycle inhibitors such as p21 or p27 (35,36). The same mechanism may play a role in the TGF-b induced suppression of an antigen-speci®c T cell response. Together, our results suggest that TGF-b mainly exerts its suppressive function not by induction of apoptosis, but by reducing the cell-cycle rate of antigen-speci®c T cells.
In addition to a remarkable inhibitory effect on antigen-speci®c T cell proliferation, TGF-b also modulated T cell activation markers. After 96 h, a signi®cant decrease in the expression of CD25 (IL-2 receptor) and an increase in the expression of CD69 (early activation marker) could be demonstrated (Fig. 3). Since it was previously suggested that a decrease in CD25 expression might be the direct result of a decrease in IL-2 production (37), the levels of IL-2 in the supernatant were also measured. Although the levels of IL-2 were low, the production of IL-2 was not diminished in the presence of TGF-b, indicating that the decrease in CD25 expression was a direct effect of TGF-b on the T cell and not indirectly via in¯uence of IL-2 levels. This is in agreement with a previous study by Kehrl et al., who described that CD25 expression was decreased in the presence of TGF-b (20). In contrast, others showed that the expression levels of CD25 and CD69 were not affected by TGF-b in T cells after mitogenic stimulation (38). However, that study also showed a promoting effect of TGF-b on mitogen-induced T cell proliferation. A possible explanation for the discrepancy between their study and the presented data in our report might be the use of mitogenic stimulation instead of antigen-speci®c stimulation and the use of PBMCs instead of established TCC.
A recent study showed that induction of tolerance in mice, marked by less responsiveness of antigen-speci®c T cells at 96 h, was preceded by the activation of antigen-speci®c T cells, which was determined by an enhanced CD69 expression at 6 h (39). Surprisingly, the expression of CD69 in our study was slightly enhanced by TGF-b even at 96 h. This may indicate that the enhancement in CD69 expression demonstrated by Sun et al. (39) was also induced by TGF-b, suggesting a role for TGF-b in tolerizing antigen-speci®c T cells.
The effect of TGF-b on antigen-speci®c T cell proliferation and activation might be mediated via the APC, e.g. by affecting the expression of the co-stimulatory molecules (40). However, the expression level of CD80 and CD86 on the EBV-B cells was not decreased by TGF-b, indicating that TGF-b exerts its function on the T cell itself during the antigen presentation by the APC (Fig 3B). Moreover, TCC stimulated with mitogen in the absence of APC were equally suppressed by TGF-b (unpublished observation). A direct effect on the T cell was further con®rmed by the observation that the expression of CD28 on the T cells was decreased by TGF-b. This suggests that the antigen-speci®c T cells are rendered less responsive towards stimulation by APC, via a direct effect of TGF-b on the T cell. IL-10 and TGF-b both have been demonstrated to be major cytokines in controlling the process of peripheral tolerance. Both cytokines are well known for their suppressive effect on T cells. However, in this study IL-10 only slightly decreased the antigen-speci®c T cell response in a few TCC. It was suggested before by Cottrez et al. (18) that IL-10 might even enhance the suppressive effect of TGF-b via up-regulation of the TGFbRII. In this report, IL-10 did not enhance the suppression of proliferation by TGF-b and the mRNA expression of TGFbRII seemed not elevated in the presence of IL-10.
In mice and humans, sensitivity of antigen-speci®c T cells towards TGF-b was recently shown to depend on the expression of the inhibitory Smad7 protein (13,14). Both studies demonstrate that diminished TGF-b signaling in cells with a high Smad7 expression results in in¯ammation. The present study showed that, prior to and after stimulation, the TCC that are insensitive towards TGF-b do not contain signi®cantly higher levels of Smad7 mRNA compared to TGF-b-sensitive TCC. In addition, for other TGF-b signaling molecules (Smad3, SARA and Hgs), no differences were also observed between sensitive and insensitive TCC prior to or after stimulation (Fig 6A and B). However, this does not exclude differences in protein phosphorylation and/or translocation or differences in protein levels of TGF-b signaling molecules between sensitive and insensitive TCC. Addition of TGF-b during stimulation resulted in a signi®cantly higher mRNA expression of SARA compared to T cells stimulated without TGF-b. A trend towards higher expression of Hgs was observed in the presence of TGF-b. It was recently demonstrated that high levels of SARA and Hgs are correlated with a reduced sensitivity of T cells towards TGF-b (11). This may indicate that the high expression of SARA and Hgs after stimulation in the presence of TGF-b leads to decreased sensitivity of TCC towards TGF-b, suggesting a possible negative feedback mechanism in TGF-b signaling.
It was demonstrated in the present paper that TGF-b is a good candidate to suppress antigen-speci®c CD4 + T cell proliferation without altering the cytokine production of these T cells. Although a decrease in cytokine production was observed over time, which is most likely due to consumption by the T cells, TGF-b was not able to induce a decrease or increase in the cytokines measured. This indicates that a shift from a T h 1 to a T h 2 pro®le or vice versa is not accomplished in vitro. Also, an increase in IL-10 production could not be achieved via the addition of TGF-b, suggesting that TGF-b will not induce differentiated T h 1 or T h 2 cells to adapt a more suppressive phenotype. Several investigators have indicated a possible bene®cial role for TGF-b in treatment of patients with autoimmune diseases, such as systemic lupus erythematosus, reduction in acute graft versus host disease or other chronic in¯ammatory diseases (27±29). In allergic diseases, such as asthma, bene®cial effects of TGF-b were demonstrated in several mouse models (30,31,41). The results of the present study demonstrate that TGF-b is a potent suppressor of the antigen-speci®c T cell response and does not modulate the cytokine production of an in vivo differentiated T cell in an in vitro system. T cells derived from both non-allergic as well as allergic donors are equally suppressed in their antigen-speci®c proliferation by TGF-b, irrespective of the cytokine pro®le of the T cells. This suggests that TGF-b might not only be bene®cial in allergy, but may be able to suppress T cell responses in both T h 1-and T h 2-skewed chronic in¯ammatory diseases. An enhancement of the TGF-b production in vivo at sites where the disease-related antigens are encountered might be of importance in the suppression of the local immune response, inducing a more tolerogenic environment to prevent a detrimental response against innocuous antigens.