Regulatory T Cell Suppression Is Potentiated by Target T Cells in a Cell Contact, IL-35- and IL-10-Dependent Manner¹

Although several studies have assessed the mechanism of regulatory T cell (T_reg)³-mediated suppression, many details remain unknown. It is thought that T_reg can suppress through direct contact with target cells or APCs (1, 2, 3) and by using secreted cytokines (4, 5, 6). The conventional assay for T_reg function in vitro tests the ability of T_reg to suppress conventional T cell (T_conv) proliferation when cultured together in a tissue culture plate. Experiments using Transwell inserts, in which T_reg and T_conv are separated by a permeable membrane, have demonstrated that T_reg are essentially unable to suppress T_conv proliferation when T_conv and T_reg are in separate compartments (1, 3). These data led to the notion that T_reg-mediated suppression is contact dependent. The addition of IL-10- and TGF-β-neutralizing Abs to conventional in vitro T_reg assays does not inhibit suppression by T_reg, suggesting that these cytokines are not required for T_reg-mediated suppression in vitro (1, 3, 7, 8). However, secreted cytokines are an important means of T_reg-mediated suppression in vivo (4, 9, 10, 11, 12, 13, 14, 15, 16, 17), making it difficult to reconcile these differential requirements.

We recently identified a novel inhibitory cytokine, IL-35, which is a member of the IL-12 family and a heterodimer comprised of Ebi3 (IL-27β) and Il12a/p35 (IL-12β) (18). It is secreted by T_reg, but not T_conv, and is required for maximal T_reg function in vitro and in vivo. We also showed that ectopic expression of IL-35 by T_conv is sufficient to confer regulatory activity and that rIL-35 can suppress the proliferation of anti-CD3-stimulated T_conv. Interestingly, preliminary data suggested that T_reg recovered from an in vitro T_reg assay, and thus in the process of active suppression, dramatically up-regulated Ebi3 and Il12a mRNA expression. This indicated that the interaction between T_conv and T_reg might potentiate IL-35 secretion and led to the hypothesis that T_conv:T_reg coculture might also enhance T_reg function. Therefore, it is possible that in the Transwell experiments reported to date, T_reg were unable to suppress T_conv proliferation because optimal suppression, including the secretion of inhibitory cytokines, is potentiated by signals derived from the T_conv that are being suppressed. In this study, we tested this possibility using Transwell culture experiments to determine the requirements for, and necessity of, IL-35 and IL-10 in T_reg-mediated suppression. We also determined the activation requirements necessary for T_conv-induced T_reg suppression.

Ebi3^−/− (C57BL/6: F11, now >98.83% B6 by microsatellite analysis (Charles River Laboratories)) were initially provided by R. Blumberg and T. Kuo (Brigham and Women’s Hospital, Boston, MA), and subsequently obtained from our own breeding colony, which was rederived at Charles River Breeding Laboratories and housed at St. Jude Children’s Research Hospital. Foxp3^gfp mice (C57BL/6: F7, now >95.32% B6 by microsatellite analysis (Charles River Laboratories)) were provided by A. Rudensky (Howard Hughes Medical Institute, University of Washington, Seattle, WA). Il10^−/− mice (N10F33) were provided by T. Geiger (St. Jude Children’s Research Hospital, Memphis, TN). Il12a^−/− (N11F31), C57BL/6, and Thy1.1 C57BL/6 mice were purchased from The Jackson Laboratory. All animal experiments were performed in American Association for the Accreditation of Laboratory Animal Care-accredited, specific pathogen-free, helicobacter-free facilities in the St. Jude Animal Resource Center following national, state, and institutional guidelines. Animal protocols were approved by the St. Jude Animal Care and Use Committee.

T_conv (CD4⁺CD25⁻CD45RB^high) and T_reg (CD4⁺CD25⁺CD45RB^low) from the spleens and lymph nodes of C57BL/6 or knockout age-matched mice were positively sorted by FACS. Purity of sorted T_conv and T_reg was verified by Foxp3 staining and FACS analysis. Where indicated, T_conv (CD4⁺Foxp3⁻CD45RB^high) and T_reg (CD4⁺ Foxp3⁺CD45RB^low) from the spleens and lymph nodes of Foxp3^gfp mice were positively sorted by FACS. Following RBC lysis with Gey’s solution, cells were stained with Abs against CD4, CD25, and CD45RB (eBioscience) for T cell isolation and sorted on a MoFlo (DakoCytomation) or Reflection (i-Cyt). Flow cytometric analysis was performed using a FACSCalibur (BD Biosciences).

Sulfate latex beads (4 μM; Molecular Probes) were incubated overnight at room temperature with rotation in a 1/4 dilution of anti-CD3 plus anti-CD28 Ab mix (13.3 μg/ml anti-CD3 Ab (eBioscience) and 26.6 μg/ml anti-CD28 (eBioscience)). Beads were washed three times with 5 mM phosphate buffer (pH 6.5) and resuspended at 5 × 10⁷/ml in sterile phosphate buffer with 2 mM BSA.

Murine T cells were purified by MACS separation using biotinylated CD8, CD11b, CD11c, CD25, Ter¹¹⁹, B220, pan-NK, Mac-1, and Gr-1 Abs to deplete non-T cells. MACS-purified T cells were seeded at 2 × 10⁶/ml in a 24-well plate coated with anti-murine CD3ε (4 μg/ml) and soluble anti-CD28 (2 μg/ml). Recombinant cytokines and neutralizing Abs were added as indicated for polarization. Th1 conditions, 35 ng/ml rIL-12 and 10 μg/ml αIL-4; Th2 conditions, 50 ng/ml rIL-4 and 10 μg/ml anti-IFN-γ. Cells were split on day 3 and expanded in medium containing IL-2 (10 U/ml). On day 6, cells were collected, washed, and used in functional assays. Polarization was verified by intracellular cytokine staining, cytokine secretion using Luminex technology, and Tbet (Th1)/Gata3 (Th2) quantitative PCR.

Purified T_conv and T_reg were cultured with anti-CD3/CD28-coated latex beads or in the absence of stimuli. Where indicated, T_reg were activated in the presence of T_conv at a 4:1 (T_conv:T_reg) ratio. T_conv were used at 2 × 10⁶ cells/ml, and T_reg at 5 × 10⁵ cells/ml in 100 μl were used in a total volume of 1 ml, with or without anti-CD3/CD28-coated latex beads. After 48 h, T_conv and T_reg were resorted based on Foxp3 expression or based on the expression of congenic markers Thy1.1 for T_conv and Thy1.2 for T_reg for analysis. T cell RNA was isolated from purified cells using the Qiagen micro RNA extraction kit or by using the TRIzol reagent. RNA was quantitated spectrophotometrically, and cDNA was generated using the Applied Biosystems high capacity cDNA reverse-transcription kit. The cDNA samples were subjected to 40 cycles of amplification in an ABI Prism 7900 Sequence Detection System instrument using TaqMan or SYBR Green PCR master mix (Applied Biosystems) (primer sequences listed in supplemental Table I).⁴ Quantitation of relative mRNA expression was determined by the comparative cycle threshold (CT) method (Applied Biosystems User Bulletin no. 2, pg. 11, http://docs.appliedbiosystems.com/pebiodocs/04303859.pdf), whereby the amount of target mRNA, normalized to endogenous β actin or cyclophillin expression, was determined by the following formula: 2^−ΔΔCT.

Immunoprecipitation and immunoblotting were performed, as previously described (6, 7, 8). Cellular supernatants were diluted in lysis buffer containing 0.1% Tween 20, 50 mM HEPES, 150 mM NaCl, 1 mM EDTA, 2.5 mM EGTA, and 1 complete protease inhibitor tablet (Roche) per 50 ml lysis buffer. Supernatants were immunoprecipitated with anti-mouse IL-12a (p35) Ab (clone C18.2; eBioscience) (precoupled to protein G-Sepharose beads). Immunoprecipitates were resolved by SDS-PAGE (Invitrogen Life Technologies), and blots were probed with a biotinylated anti-mouse Ebi3 mAb (30A1; eBioscience). Blots were developed using ECL (Amersham Biosciences), and bands were quantitated by densitometry using a Bio-Rad Gel Documentation System.

Cytokine secretion in cell culture supernatants was measured by Luminex technology. T_conv and T_reg were used at 2 × 10⁶ cells/ml and 5 × 10⁵ cells/ml, respectively, in a 100 μl vol in a 96-well round-bottom plate with or without anti-CD3/CD28-coated latex beads. After 72 h, the supernatants from these in vitro coculture assays and Transwell experiments were collected and analyzed for IL-10, IL-2, and TGF-β using Millipore multiplex kits from Linco Research. Cytokine concentrations were determined, in duplicate, using standard curves of reference proteins supplied by the manufacturer.

For standard proliferation assays, 5 × 10⁴ T_conv were activated with anti-CD3/anti-CD28- or anti-Vβ8-coated latex beads. Cultures were pulsed with 1 μCi of [³H]thymidine for the final 8 h of the 72-h assay and harvested with a Packard harvester. The cpm were determined using a Packard Matrix 96 direct counter (Packard Biosciences). In vitro T_reg function was measured by culturing 5 × 10⁴ T_conv with anti-CD3/anti-CD28-coated latex beads, and titrations of T_reg. T_conv were activated with anti-CD3/anti-CD28-coated latex beads for 72 h. For direct comparison of T_reg suppressive capacity with and without T_conv contact, T_reg were also cultured in direct contact with responder T_conv in the bottom chamber of the Transwell plate. In indicated assays, rIL-10 (eBioscience) or neutralizing anti-IL-10 Ab (clone JES5-2A5; BD Biosciences) were added to standard T_reg assays and Transwell experiments. Cultures were pulsed and harvested, as described for proliferation assays.

Where indicated, T_conv were fixed at a 1/5 dilution of 20% formaldehyde in culture medium, incubated at room temperature for 20 min, and washed three times with medium before culture. “Not preactivated” cells were cultured immediately following purification. Where indicated, some cells were preactivated before fixation and culture. For these experiments, T_reg were activated for 24 h at 5 × 10⁵/ml in 96-well round-bottom plates containing anti-CD3 (1 μg/ml) and anti-CD28 (2 μg/ml). Following activation, T_reg were washed thoroughly, counted, and fixed, where appropriate. Transwell experiments were performed in 96-well plates with pore size 0.4 μM (Millipore). Freshly purified responder T_conv (5 × 10⁴) were cultured in the bottom chamber of the 96-well plates in medium containing anti-CD3/anti-CD28- or anti-Vβ8-coated latex beads, where indicated. Cells assayed for regulatory capacity, in medium with or without anti-CD3/anti-CD28- or anti-Vβ8-coated latex beads, were cultured in the top chamber. T_conv and T_reg were either cultured alone at 1.25 × 10⁴/well or cocultured at a ratio of 4:1 with a total of 2.5 × 10⁴ cells in the top chamber. After 64 h in culture, top chambers were removed and [³H]thymidine was added directly to the responder T_conv in the bottom chambers of the original Transwell plate for the final 8 h of the 72-h assay. Cultures were harvested with a Packard harvester. The cpm were determined using a Packard Matrix 96 direct counter (Packard Biosciences).

We have previously shown that among CD4⁺ T cells, IL-35 expression and secretion are restricted to Foxp3⁺ T_reg (18). Reports indicate that T_reg require stimulation to mediate suppression, suggesting that T_reg activation may induce the production of T_reg-specific cytokines. To test this hypothesis, T_reg from Foxp3^gfp mice were assayed by real-time quantitative PCR for expression of IL-10 and IL-35 (Ebi3 and Il12a), inhibitory cytokines made by natural T_reg, either in the presence or absence of T_conv. Upon activation of either T_conv or T_reg, IL-10 mRNA levels increased when compared with expression in resting cells (Fig. 1,A). When T_reg from Thy1.1 mice were cultured in the presence of T_conv from Thy1.2 mice, either in the absence or presence of stimulation, and subsequently resorted using Thy1 congenic markers, IL-10 mRNA was reduced in T_reg and absent in T_conv. It is noteworthy that the 3′ untranslated region of Il10 mRNA contains AU-rich regions that can affect its stability, raising the possibility that contact with T_conv may induce this Il10 mRNA reduction (19). We also measured IL-10 secretion from individual T_reg or T_conv populations or T_reg:T_conv coculture to determine the relative amount of IL-10 derived from each cell population. Consistent with mRNA expression, IL-10 secretion by both T_conv and T_reg was enhanced following activation (Fig. 1,B). In contrast, there was no IL-10 secreted by either T_reg or T_conv in the absence of stimulation. Total IL-10 secretion by the T_reg:T_conv coculture was greater than the sum of that generated by the individual cell populations, indicating that cumulative IL-10 secretion was enhanced by T_reg:T_conv coculture. Our results suggest that in the coculture, IL-10 was derived primarily from T_reg because using Il10^−/− T_conv had no impact on mRNA expression in T_reg (Fig. 1,A) nor on total IL-10 secretion (Fig. 1 B). However, using Il10^−/− T_reg substantially reduced the amount of IL-10 secreted. Whether this is reflective of the small amount of IL-10 secreted by T_conv or a role for T_reg-derived IL-10 in inducing production of IL-10 by T_conv remains to be determined. The mechanism that underlies the apparent discordance between Il10 mRNA and protein expression remains unclear and appears complex (20). However, consistent with our observations, several studies in monocytes have suggested that autocrine and perhaps paracrine IL-10 production and STAT3-dependent IL-10R signaling can enhance Il10 mRNA instability and degradation (21).

We have previously shown that in contrast to Il10, Ebi3 and Il12a mRNA levels are essentially unaltered upon TCR ligation (18). However, Ebi3 and Il12a mRNA expression was dramatically up-regulated in T_reg purified from a coculture with T_conv (Fig. 1,C). We now show that this increase in Ebi3 and Il12a mRNA translates into a substantial increase in the total amount of IL-35 secreted (Fig. 1,D). Interestingly, expression of both Ebi3 and Il12a mRNA was up-regulated in the suppressed T_conv purified from this coculture (Fig. 1 C).

Previous studies using Transwell culture plates suggested that T_reg were unable to suppress responder T_conv proliferation when the two cell populations were separated by a permeable membrane (1, 3, 7, 8). It is important to note that in these studies, resting or activated T_reg (in the absence of T_conv) were cultured in the top chamber of a Transwell plate, and their ability to suppress activated T_conv in the bottom chamber was determined by [³H]thymidine incorporation. However, our data indicate that T_reg required direct contact with T_conv for maximal secretion of IL-35, suggesting that cytokine dependence or independence of in vitro T_reg function should be re-examined under these conditions. Thus, different cell populations in the upper chamber of a Transwell plate were assayed for their ability to suppress proliferation of purified responder T_conv that were cultured with anti-CD3 (clone 145-2C11)/ anti-CD28 (clone 37.51)-coated latex beads in the bottom chamber. As other groups have shown, minimal suppression of responder T_conv proliferation in the lower chamber was observed with resting or activated T_conv or T_reg in the upper chamber (Fig. 2^,A). Remarkably, wild-type T_reg in the presence of T_conv (at a 4:1 T_conv:T_reg ratio in the upper chamber) suppressed responder T_conv proliferation (in the lower chamber) by 40% in the absence of cell contact (Fig. 2^,A). Control experiments, in which the total number of cells in the top chamber of the Transwell plate was kept the same, demonstrate that this was not merely a result of combined suppression of individual T_conv and T_reg, but rather T_conv potentiation of T_reg suppression (supplemental Fig. 1).⁴ Because this suppression occurred across a permeable membrane, this inferred a role for soluble factors. Thus, we next assessed whether the inhibitory cytokines IL-35 and IL-10 contributed to the T_conv-potentiated T_reg suppression. Strikingly, IL-35-deficient T_reg from either Ebi3^−/− or Il12a^−/− mice fail to increase suppression across the permeable membrane in the presence of T_conv in the upper chamber (Fig. 2,A). This suggests that IL-35 is required for this cell contact-independent suppression by T_reg in vitro, which correlates with the 40-fold increase in IL-35 secretion observed upon contact with T_conv (Fig. 1,C). Previous studies have suggested that IL-10 does not contribute to T_reg function in vitro, even though there are substantial data to support its importance in vivo. To our surprise, in the Transwell system, Il10^−/− T_reg in the presence of T_conv have reduced suppressive capacity when compared with wild-type T_reg; however, this was greater than in the absence of T_conv in the upper chamber (Fig. 2 A). This raises the possibility that IL-10 and IL-35 act together or synergize for maximal suppression mediated by T_reg and that their function is potentiated by T_conv.

The ability of T_conv to potentiate T_reg-mediated suppression across the Transwell was not dependent upon the strength of proliferation of the responder T_conv in the bottom chamber of the Transwell. Transwell T_reg suppression was seen when the proliferation of responder T_conv alone ranged from 5,000 to 75,000 cpm (data not shown). It should be noted that all cell populations were sorted based on cell surface marker expression, T_conv (CD4⁺CD45RB^highCD25⁻) and T_reg (CD4⁺CD45RB^lowCD25⁺), and were both pure and expressed comparable levels of Foxp3, as determined by intracellular staining (supplemental Fig. 2).⁴ Thus, any differences in their suppressive capacities were not due to differences in purity or Foxp3 expression. To illustrate the robustness of the suppression seen across the Transwell, parallel suppression assays were performed in which T_reg were placed in contact with T_conv in the lower chamber. Our results show that Transwell suppression mediated by T_conv:T_reg coculture was approximately half what was seen in conventional T_reg assays in which T_conv and T_reg are in direct contact (2:1 T_conv:T_reg ratio, 91 vs 59%; 4:1 T_conv:T_reg ratio, 78 vs 39%) (Fig. 2 B). This is surprisingly robust given that the IL-35 and IL-10 secreted by the T_reg in the upper chamber have to diffuse across the entire well to achieve suppressive concentrations. This contrasts with the substantially higher concentration of cytokine that would be present in close proximity to targets with which T_reg are in contact (22).

Recent data suggest that the suppressive effects of T_reg are mediated in part by consumption of IL-2 (23). To determine whether differences seen in the ability of wild-type and cytokine-deficient T_reg to suppress across the Transwell were due to differential consumption of IL-2, we measured IL-2 concentration in the Transwell T_reg assay. Total IL-2 secretion was not reduced in the Transwell culture when responder cell proliferation was suppressed (supplemental Fig. 3a).⁴ These data suggest that the Transwell suppression observed was not due to IL-2 deprivation-mediated apoptosis.

Because there are dual sources of IL-2 in this assay (from T_conv in the top and bottom chambers of the Transwell), IL-2 secretion may in fact be reduced on a per cell basis. Indeed, if one adjusts for T_conv number, IL-2 secretion per 10⁴ cells in the assay was reduced (supplemental Fig. 3b).⁴ To directly assess whether IL-2 expression was reduced by Transwell suppression, we measured IL-2 mRNA expression in the target responder T_conv in the bottom chamber following suppression by T_conv:T_reg cocultured in the top chamber. We found that the responder T_conv IL-2 expression was reduced by 30–50%, depending upon the genotype of the T_reg in the top chamber (Fig. 2 C). This suggests that IL-2 expression is reduced by Transwell suppression in the absence of direct contact with target T_conv in the bottom chamber.

Recent studies using preactivated and fixed T_reg have contributed to our understanding of the characteristics and conditions required for T_reg to suppress T_conv proliferation (1, 24). Reports indicate that previously activated T_reg do not require restimulation through their TCR to suppress T_conv proliferation. Moreover, once preactivated, T_reg can be fixed and still retain their suppressive capacity. To determine whether Il10^−/− and Ebi3^−/− T_reg were able to suppress under these conditions, wild-type, Il10^−/−, and Ebi3^−/− T_reg were assayed for proliferative capacity following preactivation, with and without fixation, in both conventional and Transwell T_reg assays. Like wild-type T_reg, freshly isolated, fixed Il10^−/− and Ebi3^−/− T_reg were unable to suppress T_conv proliferation in a conventional T_reg assay. Not surprisingly, if T_reg were preactivated before fixation, all T_reg regardless of genotype were able to mediate suppression, albeit to a slightly lesser degree than freshly isolated T_reg (Fig. 3,A). In a Transwell assay, preactivation did not enhance the suppressive capacity of the T_reg, and preactivated and subsequently fixed T_reg, regardless of genotype, were unable to suppress responder T_conv proliferation in the bottom well (Fig. 3 B).

It is well known that IL-10 is essential for T_reg suppression in vivo (4, 9, 10, 11, 12, 13, 14, 15, 16, 17). However, previous studies using IL-10-neutralizing Abs suggested that IL-10 is not required for T_reg-mediated suppression in vitro (1, 3, 8). We sought to expand upon these results by using an IL-10-neutralizing Ab in both conventional and Transwell T_reg assays. We show that IL-10 neutralization has no effect on suppression in a conventional T_reg assay, as previously reported; however, in the Transwell system, neutralizing IL-10 reduces the suppression mediated by wild-type T_reg from 40 to 15% and further reduces the suppression by Il10^−/− T_reg from 25 to 5% (Fig. 4). Hence, IL-10 secretion from both the T_reg and suppressed T_conv populations appears important to maintaining the fidelity of T_reg suppression across the Transwell. This is surprising given that the amount of IL-10 that the suppressed target T cells (T_sup) produce appears small when Il10^−/− T_reg were cultured with T_conv (Fig. 1, A and B). However, it is possible that the actual IL-10 contribution is much higher because T_reg-derived IL-10 may potentiate IL-10 production by the T_sup. Importantly, Transwell suppression by Il10^−/− T_reg was restored by addition of rIL-10 (Fig. 4). Thus, the differential and potentially synergistic use of IL-10 and IL-35 by T_reg may help explain and unify in vitro and in vivo results regarding the necessity of IL-10 for T_reg suppression. In a Transwell T_reg assay, ∼300 pg/ml TGF-β was detected in the culture supernatant (data not shown). Given the modest amount of TGF-β present in the Transwell T_reg assay, it is unlikely that suppression of T_conv proliferation is influenced by TGF-β; however, this will need to be formally tested.

We have shown that T_reg must be activated in the presence of T_conv to mediate suppression in the absence of cell contact (i.e., across the permeable membrane). Next, we wanted to determine the characteristics required by T_conv to facilitate this suppression. We first showed that polarized Th1 or Th2 CD4⁺ effector T cells in the upper chamber could potentiate T_reg-mediated suppression across the membrane (supplemental Fig. 4).⁴ We then asked whether the T_conv that were cocultured with T_reg required activation to potentiate suppression across the permeable membrane. To determine this, we used two experimental approaches that gave comparable results. First, we cultured T_reg in the presence of T_conv that were fixed with formaldehyde and fail to proliferate (Fig. 5,A). We show that they have a significantly reduced capacity to potentiate T_reg-mediated suppression across the membrane (reduced by 55%) (Fig. 5,B). Second, we used an anti-TCR Vβ8 mAb (F23.1) to stimulate, rather than anti-CD3ε, so that the activation of T_conv in the upper chamber could be controlled by the Vβ expression of the T_conv used (Vβ8⁺ T_reg in the presence of Vβ8⁺ or Vβ8⁻ T_conv). As seen with the fixed T_conv, Vβ8⁻ T_conv fail to proliferate (Fig. 5,C), and the suppressive capacity of Vβ8⁺ T_reg stimulated with anti-Vβ8 mAb-coated beads could be potentiated by Vβ8⁺ T_conv, but only partially by Vβ8⁻ T_conv (45% reduction) (Fig. 5 D). Thus, in both systems, the ability of T_conv to augment T_reg activity was reduced by ∼50%. Thus, potentiation of T_reg suppression across a permeable membrane can be mediated by different CD4⁺ T cell subsets, with which T_reg require direct contact, and was maximal when the T_conv were activated.

Although IL-10 production was primarily derived from T_reg, T_conv also secrete a small amount of IL-10 (Fig. 1, A and B). Therefore, IL-10 production by these suppressed T cells (T_sup) may contribute to the suppression observed in Transwell experiments. To further elucidate the importance of T_sup-derived cytokines in Transwell suppression, we cultured wild-type, Ebi3^−/−, and Il10^−/− T_conv with wild-type T_reg in the top chamber of a Transwell experiment in which wild type T_conv were in the bottom chamber. Cocultures with wild-type and Ebi3^−/− T_conv mediate Transwell suppression equally well; however, cocultures with Il10^−/− T_conv exhibit reduced suppression, suggesting that in addition to the T_reg-derived IL-10, T_sup-derived IL-10 may also contribute to the regulatory milieu (Fig. 6).

These observations present a model for the relative contribution of IL-10 and IL-35 to the suppressive activity of T_reg. There is the general belief that T_reg mediate suppression in a largely cytokine-independent, cell contact-dependent manner, primarily based on studies using Transwell culture plates. However, our results indicate that whereas T_reg activation alone does not maximally induce IL-35 expression or mediate long-distance/distal suppression, as examined across a permeable membrane, both activities are mediated and potentiated by contact with T_conv.

Many interactions between cells of the immune system are mediated by bidirectional signals. Although the mechanism whereby T_conv boost T_reg function is currently unknown, we hypothesize that receptor-ligand interactions between cocultured T_conv and T_reg occur to initiate distinct signaling pathways within each cell type. In the T_reg, a signaling pathway that leads to enhanced IL-35 secretion is most likely coupled with enhanced expression of many other regulatory proteins. Although the identity of this receptor:ligand interaction is unknown, our data suggest that it is most likely mediated by a cell surface receptor on T_conv because fixed cells could mediate T_reg potentiation. Furthermore, although resting cells could mediate potentiation, this was clearly maximal following T_conv stimulation, suggesting that this is an activation-induced ligand.

The observation that T_reg function is potentiated by contact with T_conv suggests an important shift in the concept of cell contact dependency of suppression. Our data suggest that it is not the function of T_reg that is solely contact dependent, but rather the induction of suppression that is mediated by T_conv contact. This does not preclude a role for contact-dependent suppression, but does suggest that this contact may also play a key role in potentiating T_reg function. These data also suggest that the bystander suppression that characterizes T_reg function could be facilitated by cytokine-mediated distal suppression. Given that T_reg represent a minor population, this may be particularly important in allowing T_reg to create a local suppressive milieu within defined/confined anatomical locations, such as lymph nodes. The data presented in this study also help to reconcile conflicting assessments on the role of IL-10 in T_reg-mediated suppression from in vitro and in vivo studies. Given the multitude of regulatory mechanisms available to T_reg (25), loss of a single mechanism may have little effect in a conventional in vitro T_reg assay. It is likely that this does not adequately reflect the challenges faced by T_reg in vivo. Our study shows that IL-10 is indeed a significant contributor to T_reg function in vitro when the T_reg arsenal is restricted. As such, the T_reg:T_conv Transwell assay we have described in this study may be a useful in vitro correlate for in vivo function that may be of particular value in assessing the regulatory capacity of different human T_reg populations.

We thank Karen Forbes for colony management; Richard Cross for FACS; Jennifer Smith for cytokine analysis; staff of the St. Jude Agricultural Research Council Histology Laboratory and Animal Husbandry Unit; and Peggy Just, Gangzhou Li, and Diana Mitchell of eBioscience for Ebi3 Western blotting reagents. We also thank Richard Blumberg and Tim Kuo for the Ebi3^−/− mice, Alexander Rudensky for the Foxp3^gfp knockin mice, and Terrence Geiger for the Il10^−/− mice.

The authors have no financial conflict of interest.