Regulatory T cells (Treg) play critical roles in maintaining tolerance and preventing autoimmunity. It is not fully clear how these cells are generated and maintained. Here, we show that resting B cells are able to expand Treg. This expansion requires TGF-β3 and signaling through the TCR and CD28. Upon activation, B cells express less TGF-β3, which reduces their capacity to expand Treg and which also results in increased Treg death. This may ensure that B cells can function as potent professional antigen presenting cells during infections. However, in the absence of any infection, we find that B-cell-deficient µMT mice have decreased percentages of Treg in the periphery. Our data suggest that resting B cells, which may be presenting self-antigens to T cells, can expand and maintain specific Treg and thus might be involved in the prevention of autoimmunity.
Regulatory T cells (Treg) are a subset of T cells that can actively suppress innate as well as adaptive immunity and have been implicated in self-tolerance, autoimmunity, cancer, transplantation immunology and a myriad of infectious diseases 1–6. There are numerous different subsets of Treg based on their phenotypic markers, amonge which the naturally occurring Treg (nTreg) population is well studied and characterized. nTreg are a subset of CD4+ T cells expressing CD25 and the transcription factor Forkhead box protein 3 (Foxp3). Foxp3 has been shown to be the master regulator for the generation of nTreg. Although Foxp3−/− mice lack CD4+CD25+ T cells, expression of Foxp3 in CD4+CD25- T cells is sufficient for converting them into CD4+CD25+ T cells with suppressive function 7, 8.
CD4+CD25+Foxp3+ Treg arise in the thymus and enter into the periphery where they constitute ∼5–10% of the CD4+ T cells. nTreg generation in the thymus is thought to require high-affinity peptide-MHC class II interactions in contrast to low-affinity peptide-MHC class II interactions required for positive selection of conventional CD4+ T cells 9, 10. There are also strong evidences for the generation of CD4+CD25+Foxp3+ Treg in the periphery. Experiments carried out in thymectomized or RAG-2-deficient mice lacking CD4+CD25+ T cells show that nTreg can be generated in the periphery from CD4+CD25− T cells 11, 12.
In addition to co-stimulation via the B7-CD28 and TCR-MHC class II pathways, various cytokines such as IL-2, TGF-β1 and IL-10 have been associated with the generation of peripheral Treg. IL-2 and IL-2R knockout mice have few, if any, nTreg and develop severe lymphoproliferation and autoimmunity 13, 14. TGF-β1 or IL-10 have been shown to expand Treg and even convert CD4+CD25− T cells into CD4+CD25+ T cells with suppressive functions 13, 15–18.
Antigen presenting cells (APC) express both MHC class II as well as B7 molecules and secrete cytokines that may modulate the differentiation of T cells into either T-effector or T-regulatory cells 19. Most studies have focused on dendritic cells (DC), which are thought to be key players in generating peripheral Treg 20–22. B cells are also APC that present antigen to T cells. Resting B cells provide little help to activate naïve CD4+ T cells and are thought to be important for tolerance induction 23, 24. However, their role in generating peripheral CD4+CD25+Foxp3+ T cells is unknown. Here, we show that resting splenic B cells can significantly expand CD4+CD25+Foxp3+ Treg in a TGF-β3-dependent manner. Our data suggest a role for peripheral resting B cells in maintaining Treg populations during homeostasis. Not surprisingly, as shown earlier 25 and by us, B-cell-deficient µMT mice have a significantly lower percentage of CD4+CD25+Foxp3+ T cells in the spleen than in age-matched controls.
Expansion of T-regulatory cells in the presence of resting B cells
It has been shown that although activated B cells promote efficient activation of T cells, T cells proliferate minimally in the presence of resting B cells 23. This suggests that resting B cells are tolerogenic. To set up this tolerogenic system, we co-cultured purified splenic B cells with purified splenic T cells in a 1:1 ratio in the presence of suboptimal concentrations of anti-CD3 antibody (resting B-cell–T-cell co-culture). As controls, we added anti-IgM to activate B cells (activated B-cell–T-cell co-culture). As expected, we found that resting B cells were able to promote only partial activation of CD4+ T cells (Fig. 1), whereas activated B cells were able to help T-cell expansion (p<0.05) and differentiation into IFN-γ producing Th1 cells (p<0.001). To determine if tolerogenic resting B cells could affect Treg numbers, we determined CD4+Foxp3+ Treg numbers and percentages in the resting B-cell–T-cell co-culture by flow cytometry. By 72 h, there was a considerable increase in the percentage of CD4+Foxp3+ T cells in the presence of naïve B cells as well as irradiated B cells (Fig. 2A). This suggests that in the absence of antigens, B cells are able to increase Treg ratios, implicating a role for resting B cells in the maintenance of peripheral Treg. However, the presence of Treg might prevent adequate immune responses against infectious agents. To address such a scenario, we performed the above experiment in the presence of external stimuli such as anti-IgM or LPS that would result in the activation of B cells. We found that upon activation, B cells were no longer able to increase Treg populations since both percentages and numbers of Foxp3+ T cells are affected (Fig. 2B and C). In addition to anti-IgM or LPS-treated splenic B cells, CH12 B cells (pre-activated B-cell lymphoma cells) were also unable to increase Treg. To confirm that the increase in Treg numbers resulted in increased suppressive activity, we performed a suppression assay. After stimulation with resting or activated B cells, T cells were sorted and incubated with freshly isolated naïve T cells, irradiated APC and anti-CD3. Figure 2D shows that T cells sorted from resting B-cell–T-cell co-cultures could suppress naïve T-cell proliferation more efficiently than T cells sorted from activated B-cell–T-cell co-cultures. Further, if the ratios of Treg to naïve T cells were increased, increased suppression was observed (p<0.05). Thus, functional Treg are generated in the presence of B cells; however, more Treg are generated in the presence of resting B cells than in the presence of activated B cells.
We next investigated if resting B cells could de novo generate Treg by converting CD4+CD25-Foxp3− T cells into CD4+CD25+Foxp3+ T cells. We incubated CD4+CD25- T cells with resting B cells as before and determined the generation of CD4+Foxp3+ T cells by flow cytometry. Resting B cells were unable to generate any CD4+Foxp3+ T cells in the absence of existing Treg populations (Fig. 3A), implying that B cells could expand and/or help in the survival of Treg, but cannot de novo generate Treg. If resting B cells expand Treg, then Treg should divide and proliferate in their presence. We labeled freshly isolated whole splenic T-cell populations with carboxy-fluoroscein succinimidyl ester (CFSE) and incubated them with anti-CD3 in the presence of resting B cells or activated B cells. As expected, B cells increased CD4+ T-cell proliferation, which was markedly enhanced in the presence of anti-IgM (Fig. 3B). Importantly, most Foxp3+ T cells underwent multiple rounds of cell division in the presence of B cells, indicating that resting B cells result in the expansion of Treg. Since a few Foxp3+ T cells also appeared to proliferate in the presence of activated B cells, we wanted to determine the exact number of Treg that underwent division in the presence of resting or activated B cells. It was also possible that activated B cells could affect the survival of Treg. To address these questions, we followed the survival and proliferation of CFSE-labeled CD4+CD25+ T cells (Treg) in the presence of resting or activated B cells. Figure 3C shows that more number of Treg underwent division in the presence of resting B cells than that in activated B cells. Furthermore, fewer Treg survived in the presence of activated B cells.
TGF-β3 production by resting B cells is essential for Treg expansion/survival
Numerous cytokines have been shown to induce Treg. TGF-β1 and IL-10 are amonge the most widely studied 18, 26, 27. Both can inhibit proliferation of T cells and are also important in the differentiation of T cells into Treg. Recently, IL-4 and IL-13 have also been shown to induce Foxp3+ T cells from CD4+CD25− T cells 28. We reasoned that resting B cells must express a cytokine that is not expressed by activated B cells, preferentially enabling resting B cells to increase Treg numbers. To identify candidate factors, we performed a preliminary microarray analysis of RNA from B cells sorted from resting B-cell–T-cell co-cultures and from anti-IgM activated B-cell–T-cell co-cultures. We found either increases or no significant changes in gene expression of the known cytokines TGF-β1, IL-10 and IL-4 or IL-13. Instead, TGF-β3 was decreased in activated B cells when compared with resting B cells. We then confirmed the microarray analysis with semi-quantitative RT-PCR to determine expression levels of TGF-β1, IL-10 and TGF-β3 in B cells sorted from resting B-cell–T-cell or activated B-cell–T-cell co-cultures. Figure 4A shows that expression levels of TGF-β3 were decreased, but TGF-β1 and IL-10 expression levels increased in activated B cells. Quantitative real-time PCR analysis further confirmed a 10-fold reduction in TGF-β3 expression and an increase in TGF-β1 expression in activated B cells (Fig. 4B). TGF-β3 belongs to the TGF-β family of cytokines and has been widely studied for its role during development of various organs but has no known function in expansion of Treg 29.
Since TGF-β3 levels are higher in resting B cells, we wanted to determine whether it is essential for the expansion and/or survival of Treg by resting B cells. We added neutralizing anti-TGF-β3 to resting B-cell–T-cell co-cultures and determined the percentages of Foxp3+ T cells by flow cytometry. We found that in the presence of neutralizing concentrations of anti-TGF-β3, resting B cells failed to increase %Treg (Fig. 4C and D). Further, Fig. 4E shows that anti-TGF-β3 affects both expansion and survival of Treg. In order to ensure specificity of our neutralizing anti-TGF-β3, we determined its ability to neutralize the suppressive effect of recombinant TGF-β1 and TGF-β3. Since both TGF-β1 and TGF-β3 can suppress anti-CD3 dependent T-cell proliferation, we cultured naïve T cells with anti-CD3 (2.5 µg/mL), in the presence or absence of recombinant TGF-β1 or TGF-β3 and/or with anti-TGF-β3. As seen in Supporting Information Fig. 1, anti-TGF-β3 is able to completely neutralize the suppressive effect of TGF-β3, but not of TGF-β1.
Treg expansion requires TGF-β3 as well as signaling through TCR and CD28
To determine if TGF-β3 is sufficient for Treg expansion, we incubated T cells with varying concentrations of recombinant TGF-β3 and anti-CD3. Fig. 5A and Supporting Information Fig. 2A show that even the highest concentration of TGF-β3 used in the culture was unable to expand Treg. TGF-β1, the known suppressive cytokine, not only signals through the same TGF-βRII receptor as TGF-β3 but also requires signaling through both TCR as well as CD28 26, 30. Since TGF-β3 alone was insufficient to expand Treg, we provided the additional signals of anti-CD28. As in the case of TGF-β1, the presence of anti-CD28 along with anti-CD3 and recombinant TGF-β3 was sufficient to expand Treg populations. Interestingly, the addition of TGF-β3 to the resting B-cell–T-cell co-culture did not further increase the percentage of Treg, although activated B cells could now expand Treg in the presence of TGF-β3 (Fig. 5B and Supporting Information Fig. 2B).
Absence of B cells in vivo affects Treg numbers in the periphery
We reasoned that since resting B cells provided both CD28 as well as TGF-β3 signals required for Treg expansion, they would contribute to the percentage of Treg in vivo. Suto et al.25 have shown that CD4+CD25+ T cells are decreased in the spleen of B-cell-deficient µMT mice. We compared the percentage of CD4+CD25+Foxp3+ Treg in the spleen of B-cell-deficient µMT mice with wild-type C57B6/L mice. Confirming the previous data, B-cell-deficient mice had significantly lower percentages of Treg in the spleen than control wild-type mice (Fig. 6A and B). Since DC are also thought to be able to expand Treg, we reasoned that perhaps DC could compensate for B-cell deficiency. We therefore determined DC percentages in the spleens of the same mice and found CD11c+ DC percentages to be markedly enhanced in µMT mice than in wild-type mice (Fig. 6C and D). Interestingly, only the CD11clo populations were increased in the µMT mice without affecting the CD11chi populations.
For many years it has been debated whether B cells can induce tolerance. More recently, the role of B cells as APC has been re-visited. Antigen presentation by B cells has been shown to induce tolerance 24, 31, and resting B cells are known to play essential roles in transplantation tolerance and tolerance against tumors in a MHC-class II dependent manner 32–34. Unlike DC, B cells are APC that do not constitutively express high levels of co-stimulatory molecules like MHC class II and B7 35. B cells are therefore thought to induce anergy in T-cells by providing only the first signal for T-cell activation 23, 24. We have identified another mechanism by which resting B cells could be involved in maintaining tolerance by expanding CD4+CD25+Foxp3+ Treg populations. Recently, Reichardt et al. have shown that B cells generate Treg in the presence of specific antigen 36. Unlike our results, they show that B cell-stimulated T cells do not have enhanced Foxp3 expression but are functionally suppressive. There are many different subsets of Treg including those that are Foxp3 independent 37 and perhaps different Treg subsets are generated in response to different stimulating conditions.
If B cells are indeed required to expand CD4+CD25+Foxp3+ T cells, their absence should affect nTreg numbers in vivo. CD4+CD25+ T cells have been shown to be decreased in the spleen of B-cell-deficient mice 25. We looked at percentages of CD4+CD25+Foxp3+ T cells in B-cell-deficient µMT mice and found that they were significantly lower than age-matched controls, although not altogether absent. This suggests that while resting B cells are important for the maintenance of peripheral Treg, other cell types might be able to perform similar functions. Other APC such as immature DC, epithelial cells and macrophages have been shown to perform similar functions of generating Treg 38–41. Thus, in the µMT mice, these other APC might be able to compensate for the absence of B cells and contribute to the maintenance of Treg in the periphery. While looking at DC populations in the same µMT mice, we found a greater percentage of splenocytes expressing CD11c, a DC marker. Interestingly only the CD11clo DC, but not the CD11chi DC, were affected in µMT mice. We also determined Treg and DC percentages in the spleens of JHD mice, a B-cell-deficient mouse strain on the Balb/c background. In this case, we found no significant differences in Treg percentages between JHD mice and their wild-type controls (data not shown). However, unlike µMT mice, JHD mice had increased CD11clo as well as CD11chi DC percentages in the spleen (data not shown). These increased DC populations might be able to compensate for the absence of B cells and maintain peripheral Treg populations during homeostasis in the B-cell-deficient mice.
What happens when the immune system is challenged and when T-effector cells should be expanding rather than Treg? We show that in comparison to resting B cells, fewer Treg expand and more Treg undergo cell death in the presence of activated B cells. Instead of the known Treg inducing cytokines, TGF-β1 and IL-10, we found that TGF-β3 was the cytokine necessary for the expansion and survival of Treg. Thus, in homeostasis TGF-β3 expressed by resting B cells could provide both survival and expansion signals for Treg; however, in the presence of agents capable of activating B cells, TGF-β3 expression is decreased, resulting in antigen-specific T-effector cell expansion rather than Treg expansion. Further, if TGF-β3 is exogenously provided, activated B cells are also able to increase Treg and are perhaps even more efficient than resting B cells in increasing Treg (however, this is not statistically significant). Perhaps the addition of exogenous TGF-β3 provides survival signals necessary for maintaining Treg numbers.
Interestingly, both TGF-β1 and TGF-β3 belong to the same family of multifunctional cytokines involved in cell proliferation and differentiation. They are highly homologous and signal through the same receptor, TGF-βIIR. Both TGF-β1 and TGF-β3 are dependent on B7-CD28 and MHC-class II-TCR signaling for expansion of Foxp3+ T cells (Fig. 5a) 26, 27. All isoforms of TGF-β are secreted in a latent form and must be cleaved before becoming activated and being able to bind to the receptor 42. Thus, even though TGF-β1 expression is higher in activated B cells, it is likely that TGF-β1 might not be cleaved to its mature form and thus may not be functional.
Although TGF-β1 has prominent roles in Treg development, TGF-β3 has been predominantly studied for its role in the development of heart, lung, testis and cleft and in wound healing and chemoprotection 29. We found that TGF-β3 is widely expressed in various tissues such as spleen, liver, heart, brain, intestine, etc. (data not shown). This ubiquitous expression might contribute to expanding Treg populations that recognize self-antigens throughout the body. TGF-β3 expression by resting B cells and epithelial cells of numerous tissues may thus aid in maintaining a suppressive T-cell population in vivo and might contribute to keeping autoimmunity in check. Others have proposed a regulatory role for B cells, whereby B cells can dampen effector T-cell function by mechanisms such as TGF-β1 or IL-10 secretion, antigen presentation and even direct contact with other immune cells 43. A recent study shows that B cells regulate CD4+CD25+Foxp3+ T cells in a B7-dependent manner, resulting in increased production of IL-10 by the Treg and subsequent recovery from experimental autoimmune encephalomyelitis 44. Our data might explain how those B cells regulate Treg numbers.
In vivo, B cells are constantly in contact with circulating T-cell populations, including Treg. Self-antigen specific B cells expressing TGF-β3 would encounter Treg in an antigen specific manner, resulting in Treg expansion. As shown in Fig. 7, we speculate that the TGF-β3 produced by resting B cells might directly act on Treg, signaling their expansion in a MHC-class II and B7-dependent manner. However, it is also possible that TGF-β3 might effectively prevent the expansion of T-effector cells consequently increasing Treg expansion. However, once the B cell becomes activated, it now expands antigen specific T-effector cells. A similar scenario would also occur within tissues expressing TGF-β3, where Treg could also be expanded.
In conclusion, we show that B cells can expand Treg populations in vitro, through the expression of the cytokine TGF-β3.
Materials and Methods
C57BL/6 and B-cell-deficient µMT mice (6-week-old females) were obtained from Jackson Laboratory and housed under specific pathogen-free conditions in the Loyola University Comparative Medicine Facility. All experimental procedures were conducted according to the protocols approved by the Institutional Animal Care and Use Committee.
Cell isolation and culture
Splenic B cells and T cells were isolated by negative selection using the B-cell isolation and pan T-cell isolation kits (Miltenyi Biotec), respectively. Cell purity was always observed to be ≥95%. Cells were cultured in vitro in RPMI 1640 medium supplemented with 10% v/v heat-inactivated FBS, 2 mM glutamine, 5×10–5 M 2-ME, and 100 U/mL penicillin and 100 µg/mL streptomycin (all were obtained from Invitrogen Life Technologies). B-cells (0.1×106) and 0.1×106 T cells were cultured for 72 h either alone or together in a 1:1 ratio in the presence of 0.5 µg/ml anti-CD3 (eBioscience) and/or 10 µg/mL LPS or 5 µg/mL anti-mouse IgM (Jackson ImmunoResearch). In some experiments, irradiated splenic B cells or pre-activated CH12 B cells were used instead of naïve splenic B cells. In other experiments, neutralizing anti-TGF-β3 (MBL international) or recombinant TGF-β3 (Fitzgerald Industries International) was added. To test the specificity of the anti-TGF-β3, we incubated 0.1×106 naïve T cells with 2.5 µg/mL anti-CD3 in the presence of 0.5 ng/mL recombinant TGF-β1 or 0.5 ng/mL recombinant TGF-β3 and/or 2 µg/mL anti-TGF-β3. All cultures were incubated in 96-well round-bottomed plates. In certain cases, CD4+CD25− T cells were sorted by FACS Aria (BD Biosciences) and cultured with B cells in the same 1:1 ratio for 3 days. In others, 0.5×105 FACS-sorted CD4+CD25+ Treg were labeled with CFSE and cultured with 0.5×105 FACS-sorted CD4+CD25− T cells and 0.1×106 B cells for either 1 day or 3 days.
Purified T cells or B cells were labeled with 2.5 µM CFSE for 10 min at 37°C in PBS. CFSE was then quenched with an equal volume of cold media and then washed twice with PBS. Cells were stained with fluorochrome-conjugated anti-CD4 and intracellular anti-Foxp3 (eBioscience) and analyzed by flow cytometry (BD FACSDiva).
At the end of the 3 day co-culture, total viable cells in each sample were counted by trypan blue staining. Cells were then stained and analyzed for Foxp3+ cells by FACS. Numbers of Treg were calculated based on viable cells counted by trypan blue staining. (Total viable cells determined by trypan blue X number of live Treg/total cells analyzed by FACS.) For a number of 7-AAD+ Treg, total cells harvested X number of 7-AAD+ Treg/total cells analyzed by FACS.
Cell culture supernatants were collected after 3 days of co-culture and stored at −20°C until use. Cytokine secretion was analyzed by ELISA using antibody pairs for IFN-γ and IL-4(BD pharmingen and eBioscience).
T cells (1×106) sorted by negative selection (MACS) from resting B-cell–T-cell and activated B-cell–T-cell co-cultures at day 3 were cultured for an additional 3 days with 0.5 µg/mL anti-CD3, 1×106 irradiated splenocytes as APC and freshly isolated naïve splenic T cells (1×106). A total of 1 µCi 3[H]thymidine (Amersham) was added to each well during the last 18 h of culture. Cells were harvested on a PHD cell harvester and incorporation was measured on a scintillation counter (Beckman Coulter).
Semi-quantitative RT-PCR and quantitative real-time PCR
RNA was isolated from sorted B cells after 3 days of culture (resting or activated) using the RNeasy kit (Qiagen) and reverse transcribed into cDNA using random primers. For semi-quantitative RT-PCR, cDNA was serially diluted 1:5 times and PCR was carried out for TGF-β1, TGF-β3, IL-10 and GAPDH (internal control). The primers used were IL-10: Fwd: 5′-TCA AAC AAA GGA CCA GCT GGA CAA CAT ACT GC-3′, Rev: 5′-CTG TCT AGG TCC TGG AGT CCA GCA GAC TCA A-3′; TGF-β1: Fwd: 5′-CGG CAG CTG TAC ATT GAC TT-3′, Rev:5′-CAG TTG CAG GAG CGC ACA ATC ATG TTG-3′; TGF-β3: Fwd: 5′-GTT TCT GGG CCA GCA ACT AG-3′, Rev: 5′-GCA TCT GGG GCC GAG TCA TC-3′; GAPDH: Fwd: 5′-GCT GAG TAT GTC GTC GAG TCT-3′, Rev: 5′-ATC ACG CCA CAG CTT TCC AGA-3′. PCR products were run on a 1% agarose gel containing ethidium bromide and visualized using UV light. Quantitative real-time PCR was performed using the 7500 Fast Real-Time PCR System (Applied Biosystems) and Taqman Gene Expression Assays for TGF-β1, TGF-β3 and GAPDH (Applied Biosystems). Data were analyzed with SDS software 1.4 using the ΔΔCt relative quantitation method.
Statistical analysis was performed with the GraphPadPrism software using one-way ANOVA for the suppression assay or unpaired t-test for all other data. A p value of <0.05 was considered statistically significant.
We would like to thank Dr. John Callaci and Ryan Himes for their help in performing the quantitative real-time PCR analysis, Dr. James Sinacore for his guidance in statistical analysis and Dr. Katherine Knight and Dr. Sahadev Shankarappa for their insightful comments on the article. We would like to thank Loyola University for funding support.
Conflict of interest: The authors declare no financial or commercial conflict of interest.