- Top of page
- Materials and Methods
CD4+ CD25+ regulatory T cells (Treg) are potent suppressors, and play important roles in autoimmunity and transplantation. Recent reports suggest that CD4+ CD25+ Treg are not a homogeneous cell population, but the differences in phenotype, function, and mechanisms among different subsets are unknown. Here, we demonstrate CD4+ CD25+ Treg cells can be divided into subsets according to cell-surface expression of CD62L. While both subsets express foxp3 and are anergic, the CD62L+ population is more potent on a per cell basis, and proliferates and maintains suppressive function far better than the CD62L– population and unseparated CD4+ CD25+ Treg. The CD62L+ population preferentially migrates to CCL19, MCP-1 and FTY720. Both CD62L+ and CD62L– subsets prevent the development of autoimmune gastritis and colitis induced by CD4+ CD25–CD45RBhigh cells in severe combined immunodeficiency (SCID) mice. Overall, these results suggest CD4+ CD25+ Treg are not a homogenous cell population, but can be divided into at least two subsets according to CD62L expression. The CD62L+ subset is a more potent suppressor than the CD62L– population or unfractionated CD4+ CD25+ Treg cells, can be expanded far more easily in culture, and is more responsive to chemokine-driven migration to secondary lymphoid organs. These properties may have significant implications for the clinical manipulation of the CD4+ CD25+ CD62L+ cells.
Tolerance is a feature of the immune system that is intimately related to discrimination between self and nonself. Clonal deletion of self-reactive T cells in the thymus is a primary tolerance mechanism, while induction of unresponsiveness or anergy in post thymic T cells may be required for the establishment of peripheral tolerance. Recent data show that Treg may play a critical role in the induction and maintenance of immune tolerance (1–5). Various types of Treg cells have been described, including Tr1 cells, Th3 cells, CD4+ CD25+ Treg cells, and others (6). CD4+ CD25+ Treg cells were first described by Sakaguchi (7). Five to 10% of CD4+ T cells constitutively expressed the α chain of IL-2-receptor (CD25) and were crucial for the control of autoreactive T cells in vivo. These cells are generated in the thymus of naïve mice, perhaps via altered negative selection by self-antigen (8). Subsequent in vitro studies showed that CD4+ CD25+ cells are typically anergic, unresponsive to TCR stimulation alone, but proliferate after addition of exogenous IL-2 (9). These cells suppress the proliferation of other CD4+ and CD8+ T cells in an antigen-nonspecific manner via a cell contact-dependent, cytokine-independent mechanism (10,11). A similar population of CD4+ CD25+ Treg cells has been defined in humans, with identical phenotypic and functional properties (12–15).
CD4+ CD25+ Treg are potent suppressors in a number of in vivo models of autoimmunity, including gastritis, thyroiditis, inflammatory bowel disease and insulin-dependent diabetes (16–19). Regulation of disease activity in vivo seems to require both suppressor cytokines (such as IL-4, IL-10 and TGF-β) and cell contact-dependent mechanisms. Recent reports addressing the role of CD4+ CD25+ Treg in allogeneic responses or transplantation show that these cells also suppress the proliferative responses of CD4+ CD25– T cells to alloantigenic stimulation in vitro (1,2), and are required for ex vivo induction of tolerance to alloantigen via costimulatory blockade (3). Alloantigen-specific CD4+ CD25+ Treg can maintain tolerance in vivo, and require IL-10 and CTLA4 for their functional activity (4,5). CD4+ CD25+ Treg can significantly delay or even prevent GVHD (20,21), and the protective effect is partially dependent on IL-10 (22). These diverse findings suggest heterogeneity in the CD4+ CD25+ Treg population or at least heterogeneity in suppressive mechanisms, depending on the experimental models and protocols. Many additional issues require further clarification, such as how many subsets of CD4+ CD25+ Treg exist, what are the functions of these subsets, what are their mechanisms of suppression, what is their relationship to Tr1 and Th3, how to expand these subsets to retain their suppressive function, and how to generate antigen-specific Treg cells for therapeutic purposes.
l-selectin (CD62L) is a member of the selectin adhesion molecule family, and is required for lymphocyte homing to peripheral lymph nodes via binding to sialylated oligosaccharide determinants on high endothelial venules in peripheral lymph nodes (23,24). CD62L is constitutively expressed on most thymocytes and leukocytes, including B and T lymphocytes, neutrophils, monocytes, and eosinophils (25–26). Unlike other leukocyte adhesion proteins, CD62L is rapidly down-regulated from the cell surface upon cell activation (27). The level of CD62L expression, along with other markers, distinguishes naïve T cells from most effector/memory T cells (28). Recent investigations show that memory T cells can be further divided into functionally distinct subsets, memory effector T cells and central memory T cells, according to the expression of CD62L and CCR7 on the cell surface. Central memory T cells express high levels of CD62L and CCR7, home to and proliferate within secondary lymphoid organs, and differentiate into CCR7– effector cells upon secondary stimulation. In contrast, memory effector T cells express low levels of CD62L and CCR7, and display immediate effector function (29–32).
As Treg are CD25+, this argues they are antigen experienced. Conversely, the fact that many Treg are CD62+ and CD45RBlow suggest that they have not undergone full activation, or have reverted back to a partially naïve phenotype. CD62L expression has also served to define certain CD4+ regulatory T-cell subpopulations (33–35). In prediabetic nonobese diabetic (NOD) mice, CD4+ splenocytes prevent transfer of diabetes to immunodeficient NOD recipients; the regulatory CD4+ T cells express high levels of CD62L, in contrast to diabetogenic T cells that reside within the CD62L– population. More recent studies demonstrate that the regulatory T-cell population in prediabetic NOD mice is CD4+ CD25+ CD62+ (36). Given these findings, we hypothesized that CD4+ CD25+ Treg cells could be divided into functionally distinct subsets according to the expression of CD62L. We found that CD4+ CD25+ Treg cells are indeed not a homogeneous cell population; CD62L is highly expressed on the surface of 50–60% of the cells. CD62+ and CD62L– populations are similar in that they are both anergic and suppressive and can be initially expanded in vitro without loss of the suppressive activity. However, there are substantial differences in their receptor expression, suppressive potency, suppressive mechanisms, and cytokine production. Of particular importance is a difference in their proliferative capacity and migration to chemokines. These results are important for defining the mechanisms of cellular activity of distinct Treg subsets.
- Top of page
- Materials and Methods
Several reports show that CD4+ CD25+ Treg cells can be further defined as subpopulations expressing CD62+, CD45RBlow (in mice) or CD45RO+ (in human) (2,10,36). Our data support the conclusion that CD4+ CD25+ Treg are not a homogeneous cell population, but can be divided into at least two subsets according to the cell-surface expression of CD62L. CD62+ and CD62L– populations are similar in that they are both anergic and suppressive, express foxp3, and can be expanded in vitro without initial loss of suppressive function. However, there are substantial differences in their cell-surface phenotype, suppressive capacity, functional mechanisms, proliferative capacity, cytokine production, and migration. The majority of CD4+ CD25+ CD62+ cells also expressed high levels of CD45RB (CD45RBhi), while the majority of CD4+ CD25+ CD62L– cells expressed low levels of CD45RB (CD45RBlo). Additional studies demonstrated that the CD45RBloCD4+ CD25+ CD62+ subset was more suppressive than the CD45RBhiCD4+ CD25+ CD62+ population (data not shown). Thus, these subsets are also likely to be heterogenous and further significant subdivisions may be defined in the future. The CD4+ CD25+ CD62+ cells also express higher levels of IL-10R1, CD80 and CD86 in comparison with CD4+ CD25+ CD62L– cells, suggesting other functional and lineage differences. Thornton and Shevach (11) earlier concluded that the CD62+ and CD62L– subsets were not different, however, in their studies the CD62+ subset was a more potent suppressor and they did not analyze these cells for the same range of variables as in the current study. Kuniyasu et al. (60) reached a similar conclusion, but again did not analyze the range of variables in the current study and only assessed Treg function at a high Treg : responder ratio of 1 : 1.
Although CD4+ CD25+ cells secrete more IL-10 than CD4+ CD25– cells (61), the role of IL-10 in the suppressive function of CD4+ CD25+ Treg cells remains controversial. Most in vitro studies show that the suppressive function of CD4+ CD25+ cells is independent of IL-10. CD4+ CD25+ cells from IL-10-/- mice retain suppressive capacity in vitro (9), and neutralization of IL-10 or IL-10R1 has no effect on the suppressive function (2). However, several in vivo studies suggest that IL-10 is required for the regulatory function of CD4+ CD25+ cells. CD4+ CD25+ CD45RBlow cells prevent inflammatory bowel disease via an IL-10-dependent mechanism, and CD4+ CD25+ CD45RBlow cells from IL-10-/- mice fail to protect from colitis (45,62). Alloantigen-specific CD4+ CD25+ Treg cells responsible for the maintenance of tolerance to donor alloantigens in vivo require IL-10 and CTLA4 for their functional activity (4,5). CD4+ CD25+ T cells that delay or prevent GVHD are partially dependent on IL-10 (20,22). In our investigations, we showed that neither IL-10, IL-10R1, nor IL-10R2 were important for Treg development or function or for the ability of responder T cells to be subjected to the suppressive function of Treg. By implication, the cytokines IL-22, IL-28, and IL-29 are also not important for Treg, as they rely on the IL-10R2 chain (47,48). As these in vitro assays are APC-dependent, IL-10 and IL-10R are not required at the level of either T cells or APCs. The requirement for IL-10 in vivo suggests that other cell types may be subject to the influences of Treg and IL-10 (63).
It has been shown repeatedly that the suppressive function of CD4+ CD25+ Treg is cell contact-dependent in vitro. When the suppressors were separated from responders by a semipermeable membrane, no suppression was observed. The molecular nature of the contact-dependent interaction is not fully established. Nakamura and colleagues showed that stimulated CD4+ CD25+ cells produce TGF-β1 and express TGF-β1 on the cell surface, suggesting that CD4+ CD25+ cells may exert immunosuppression by cell–cell interaction involving cell-surface TGF-β1 (43), although this association has been contradicted by others (64). CTLA4 is a negative regulator of T-cell activation, and is expressed constitutively by CD4+ CD25+ cells (65), although most investigators have failed to demonstrate inhibitory effects of CTLA4 blockade on Treg function (12–15). Nonetheless, some reports show that the addition of anti-CTLA-4 antibody abrogates suppression in vitro and in vivo in colitis and transplantation models (5,65,66), suggesting that in some situations Treg cell-surface CTLA4 is functional. Our experiments show that suppression by the CD62L– subset is partially CTLA4-dependent, different from unseparated CD4+ CD25+ Treg or the CD4+ CD25+ CD62+ subset. Further, the CD62L– subset expresses lower levels of the CD80 and CD86 ligands. These results suggest that it is the CD62L– subset that depends on CTLA4 for some of its suppressive function. The result might suggest that the cell-surface CTLA4 does not necessarily regulate intrinsic Treg activation, but rather focus these cells on CD80+ or CD86+ targets.
CD4+ CD25+ Treg cells account for only 5–10% of the total CD4+ population in both mice and humans. Therefore, the administration of sufficient numbers of freshly isolated CD4+ CD25+ cells is not therapeutically practical, and the expansion of CD4+ CD25+ Treg is critically important for clinical cellular therapy. Several groups have tried to expand CD4+ CD25+ Treg in vitro with limited success (1,2,11,20,21,49). In most situations, CD4+ CD25+ Treg could be expanded 10–30-fold after 7–14 days in culture, remained anergic and retained their suppressive properties. However, after several more rounds of restimulation, CD4+ CD25+ T cells failed to suppress CD4+ CD25– T cells and were responsive to anti-CD3 in the absence of exogenous IL-2, suggesting that CD4+ CD25+ Treg cells might lose their anergic and suppressive properties after long-term culture (67). In our system, we stimulated the CD62+ and CD62L– subsets with anti-CD3 mAbs, syngeneic T-depleted APCs and IL-2. The CD62+ subset proliferated well in vitro, could be expanded 20–30-fold, and retained suppressive function for at least 30 days in culture; far better than the CD62L– subset, which proliferated less well and for only a short period of time. Interestingly, unseparated CD4+ CD25+ Treg also lost suppressive function after 30 days in culture. This suggests that the CD62+ subset can be expanded for longer periods while retaining suppressive properties, and that stimulation of this subset within the unseparated CD4+ CD25+ population may even impede their development. Thus, previous failed attempts to sustain Treg in culture may have resulted from improper selection of the correct subset at culture initiation. These findings are relevant for the use of these cells as cellular therapy in immune diseases. Future studies must define additional culture conditions that allow further cellular expansion with retention of suppressive function.
Lymphocyte trafficking through the endothelium involves a sequence of events relying on adhesion molecules (such as CD62L), chemokine receptors, and integrins. Lymphoid chemokines are important for trafficking into and within lymphoid compartments, whereas inflammatory chemokines attract lymphocytes into peripheral tissue (68). There is contradictory information about the chemotactic activity of CD4+ CD25+ Treg. Some reports indicate that Treg express CCR5; and that its ligand, CCL4, is the most potent chemoattractant for these cells (69). CD4+ CD25+ Treg also express CCR4 and CCR8, and are chemoattracted by their respective ligands, CCL17/TARC, CCL22/MDC, and CCL1/I-309 (55,56). However, Rudensky et al. showed that CD4+ CD25+ Treg are refractory to the lymphoid chemokines CCL19/ELC, CCL21/SLC, CCL2, CCL4, CCL22 and CXCL12 (54). A recent report shows that CD4+ CD25+ CD62+ splenocytes express CCR7 at high levels and migrate toward the lymphoid chemokine ligands CCL19 and CCL21 (ligands for CCR7); whereas CD4+ CD25+ CD62L– splenocytes preferentially express CCR2, CCR4, and CXCR3 and migrate toward the corresponding inflammatory chemokine ligands CXCL9/MIG and CXCL10/IP-10 (ligands for CXCR3), and MDC and TARC (ligands for CCR4; 36). In our experiments, the CD62+ population preferentially migrated to both CCL19 and MCP-1 (ligands for CCR7 and CCR2, respectively), but not to CCL1, CCL5/RANTES, CCL22 and CXCL9 (agonists for CCR8, CCR5, CCR4 and CXCR3). Further, FTY720, a sphingosine-derived immune modulator that causes increased T-cell homing to peripheral lymph nodes from peripheral blood and spleen (38) enhanced CD62+ Treg subset migration to CCL19 and MCP-1 (57–59). The discrepancy among the various reports may be because of small variations in the precise experimental details or mouse strains used, but a more important reason may be that distinct Treg subsets have unique chemokine responses and most reports did not examine discrete subsets. Because the CD62+ subset by definition expresses l-selectin and because it preferentially responds to CCL19, it may actively home in vivo to secondary lymphoid organs and exert suppressive functions at these sites. In this regard, FTY720 could act as an immunosuppressant by promoting Treg migration to secondary lymphoid organs. The CD62L– subset did not respond to many different chemokines and thus may be sessile in the spleen (from which they were isolated) or respond to other chemokines not yet tested. As both subsets inhibited the development of gastritis and colitis, this might mean that they can both migrate to the stomach and colon. Alternatively, they could migrate to different locations in vivo, yet interrupt distinct loci in the progression of autoimmunity, resulting in similar clinical outcomes.
CD4+ CD25+ Treg are not a homogeneous cell population, but can be divided into at least two subsets according to CD62L expression. To our knowledge, this is the first report to demonstrate systematically differences in phenotype, function, and mechanism between the two subsets of CD4+ CD25+ Treg. As the CD62+ population is a more potent suppressor than the CD62L– population or unfractionated CD4+ CD25+ Treg cells, can be expanded far more easily in culture, and is more responsive to chemokine driven migration to secondary lymphoid organs, these cells may be important for clinical use.