glucocortocoid-induced TNF-like receptor
Among the several mechanisms known to be involved in the establishment and maintenance of immunological tolerance, the activity of CD4+CD25+ regulatory T lymphocytes has recently incited most interest because of its critical role in inhibition of autoimmunity and anti-tumor immunity. Surprisingly, very little is known about potential genetic modulation of intrathymic regulatory T lymphocyte development. We show that distinct proportions of CD4+CD25+FoxP3+ regulatory T cells are found in thymi of common laboratory mouse strains. We demonstrate that distinct levels of phenotypically identical regulatory T cells develop with similar kinetics in the mice studied. Our experimental data on congenic mouse strains indicate that differences are not caused by the distinct MHC haplotypes of the inbred mouse strains. Moreover, the responsible loci act in a thymocyte-intrinsic manner, confirming the latter conclusion. We have not found any correlation between thymic and peripheral levels of regulatory T cells, consistent with known homeostatic expansion and/or retraction of the peripheral regulatory T cell pool. Our data indicate that polymorphic genes modulate differentiation of regulatory T cells. Identification of responsible genes may reveal novel clinical targets and still elusive regulatory T cell-specific markers. Importantly, these genes may also modulate susceptibility to autoimmune disease.
Immune tolerance is established by a variety of mechanisms acting in primary lymphoid organs during lymphopoiesis and in so-called “peripheral” lymphoid organs during the activation or differentiation phase of mature lymphocytes 1. Among the several mechanisms known to be involved, dominant tolerance (i.e. mediated by regulatory or suppressor lymphocytes) has incited great interest because of its potential in treatment of diseases as varied as autoimmune disorders and cancer, as well as in transplantation 2–4. The best studied regulatory T cell subset is of CD4+CD25+ phenotype. These cells were discovered because of their crucial role in the inhibition of multi-organ autoimmune disorder induced by thymectomy of mice at day 3 of life 5. Later, these cells were shown to inhibit inflammatory bowel disease, to fine-regulate immunity to pathogens, to inhibit anti-tumor immunity, and to protect the fetus from maternal immune aggression 6–9. Because of their crucial role in vivo, CD4+CD25+ regulatory T lymphocytes are very good candidates as therapeutic agents for the regulation of transplantation tolerance and inhibition of autoimmunity. It has recently been shown that these cells can inhibit graft-vs.-host disease 10–14, rejection of transplanted tissue 2, 15, and autoimmune disease in experimental settings 16–18.
Probably the majority of (but not all) CD4+CD25+ regulatory T lymphocytes develop in the thymus 19–23. In this organ, CD25+ regulatory and CD25– effector T lymphocytes appear to have common CD4–CD8–24 and CD4+CD8+ (our unpublished data) precursors. Similar to effector T cells, regulatory T cells are positively selected via interaction with thymic cortical epithelial cells 25. Expression of high-affinity ligands by thymic epithelial cells has been reported to favor development of regulatory T cells 19, 26, 27. A recent report suggests that this may be due to deletion of CD4+CD25– but not CD4+CD25+ precursors upon recognition of their cognate ligand expressed by thymic epithelial cells 28. Interestingly, interaction with high-affinity/avidity ligands expressed by thymic antigen-presenting cells of bone marrow origin can lead to deletion of regulatory T cell precursors 29, 30. Consistent with these observations, we and others have previously shown that the peripheral repertoire of regulatory T lymphocytes is enriched in auto-specific cells 29, 31, 32.
Surprisingly, despite the generally appreciated crucial importance of dominant tolerance, little is known about genetic control of regulatory T cell development and function. Such potential genetic variations might modulate susceptibility to a large panel of pathologies. Moreover, they would help in providing information concerning fundamental issues such as lineage choice and selection of regulatory T cell precursors in the thymus and functioning of these cells in the periphery.
Only one very rare genetic polymorphism is known to modulate differentiation of regulatory T lymphocytes. The forkhead/winged-helix transcription factor FoxP3 is preferentially (but not exclusively) expressed by regulatory T lymphocytes 33–37. Transfection of effector T cells with constructs encoding this transcription factor causes these cells to exert potent suppressor-effector functions 33–36. Mice carrying a natural mutation in the gene encoding FoxP3 (“scurfy”) lack regulatory T lymphocytes and die after a few weeks of life 33, 34. In humans, a natural mutation in FoxP3 causes the rare lethal autoimmune disorder IPEX 38, 39. To our knowledge this is the only genetic polymorphism known to modulate regulatory T lymphocyte development.
We here present data indicating the existence of genetic polymorphisms causing quantitative differences in regulatory T lymphocyte development in common laboratory mouse strains. We show that genes outside the MHC and acting in a thymocyte-intrinsic manner modulate intrathymic differentiation of regulatory T lymphocytes. Ultimate identification of the responsible loci should prove important for the analysis of thymic regulatory T cell lineage choice and selection, may enable identification of still elusive regulatory T cell-specific markers, and may yield more insight in mechanisms modulating susceptibility to autoimmune disease.
Distinct proportions of CD25+ regulatory T cells in thymus of different inbred mouse strains
We analyzed the proportion of CD25+ regulatory cells among CD4+CD8– (CD4SP) TCRhigh thymocytes and peripheral blood lymphocytes in the inbred mouse strains C57BL/6 (B6), C57BL/10 (B10), BALB/c, DBA/2, DBA/1, and SJL (Fig. 1A). Since regulatory CD4SP T cells express high levels of CD25 while cells expressing intermediate levels of CD25 proliferate and produce IL-2 40, 41, we only considered thymocytes of CD25high phenotype.
In the thymus, statistically significant different percentages of CD25high cells were observed between B6 and B10 mice on one hand, and DBA/2, BALB/c, DBA/1, and SJL strains on the other (Fig. 1B). These differences reached, in the strains analyzed, up to 1.7-fold (DBA/1 vs. B6). The quantitative differences might be caused by distinct CD25– effector (rather than CD25+ regulatory) T cell percentages. To evaluate this possibility, we analyzed the ratio of mature CD4SP TCRhighCD25high regulatory T cells to their CD4+CD8+ precursors in B6, DBA/2, and SJL mice. This ratio was significantly higher in DBA/2 and SJL mice than in B6 animals (Fig. 1C, top). On the other hand, the ratio of CD4SP CD25– to CD4+CD8+ thymocytes was similar in all three mouse-strains (Fig. 1C, bottom). These data indicate that the increased proportions of CD25+ regulatory T cells among CD4SP TCRhigh cells correspond to increased production from immature precursors.
To evaluate whether the CD4SP CD25high thymocytes found in the different mouse strains belong to the same regulatory T cell population, we assessed their surface phenotype (Fig. 1D). All CD4SP CD25high thymocytes were TCRhigh in all mouse strains studied. Interestingly, CD25high cells expressed relatively low heat-stable antigen (HSA, CD24) and CD69 levels, clearly distinguishing them from their CD25– and CD25intermediate counterparts. Moreover, in all mouse strains all CD25high cells expressed very high levels of the glucocortocoid-induced TNF-like receptor (GITR), characteristic for regulatory T cells. Most importantly, all CD4SP CD25high thymocytes expressed FoxP3. These data indicate that the CD25high cells found in the different mouse strains all belong to the same regulatory T lymphocyte lineage.
It has previously been shown that regulatory CD4+CD25+ T lymphocytes can develop in the periphery from CD25– precursors. Recent data suggest that CD25– precursors for CD25+ regulatory cells express FoxP3 42. Moreover, CD4+CD25–FoxP3+ T cells inhibit T cell activation in vitro41. We therefore analyzed the percentage of FoxP3+ cells among CD4SP thymocytes by flow cytometry. As shown in Fig. 1E, substantially higher percentages of FoxP3-expressing cells were observed in DBA/2 and SJL mice than in B6 animals, confirming and extending our data on CD25high thymocytes.
We also analyzed levels of regulatory T cells in the periphery. As shown in Fig. 1F, we failed to observe a direct correlation of CD4+CD25+ percentages in thymus vs. PBMC. Similar data were obtained for secondary lymphoid organs (not shown).
Distinct proportions of thymic regulatory T cells are caused by differences in their differentiation
Distinct proportions of regulatory CD25+ cells among CD4SP thymocytes may be due to differences in their development or in thymic retention of mature thymocytes. To study the former possibility, we analyzed the kinetics of regulatory T cell development by measuring the appearance of bromodeoxyuridine (BrdU)+ cells in mice continuously fed with this nucleotide analog in their drinking water. As shown in Fig. 2, more CD4+CD25+ regulatory T lymphocytes differentiated from their dividing precursors in SJL mice than in B6 thymi. This result establishes that significant differences in thymic differentiation of these cells exist between these two mouse strains. However, it formally does not exclude the possibility that differences in thymic retention of regulatory T cells may also exist.
Distinct regulatory T cell proportions are not caused by differences in MHC haplotype
We next studied whether the MHC haplotypes of the distinct mouse strains analyzed were responsible for the quantitative differences in thymic regulatory T cell generation. SJL mice have significantly higher percentages of CD25+ CD4SP thymocytes than B10 mice (Fig. 3A). Congenic B10.S mice (which carry the H-2s locus from SJL mice on a B10 genetic background) have a similar proportion of thymic regulatory T cells as B10 mice. Similarly, B10.D2 mice (carrying the DBA/2-derived H-2d locus) have CD4+CD25+ percentages similar to those in B10 mice (Fig. 3B). Therefore, the distinct MHC class I and II haplotypes of these mouse strains are not responsible for the different proportions of thymic regulatory T cells, and the genetic loci involved are not linked to the MHC.
Thymocyte-intrinsic factors determine the distinct proportions of regulatory T cells
Differences in the development of regulatory T cells may be due to thymocyte-intrinsic factors or to variations in the thymic microenvironment. To distinguish between these two possibilities we generated mixed bone marrow chimeras in which thymocytes derived from the different donor mouse strains differentiate simultaneously in the same thymic microenvironment. (B6 × DBA/2)F1 (B6D2F1) hosts were lethally irradiated and reconstituted with a 1:1 mixture of B6 and DBA/2 bone marrow cells (B6 + DBA/2 <$>\rightarrow<$> B6D2F1 chimeras). Six weeks later the thymi of these chimeras were analyzed by flow cytometry.
As shown in Fig. 4A, C, among B6-derived cells from these mixed chimeras the same (lower) proportion of thymic regulatory T cells was found as in the parent strain. Among DBA/2-derived thymocytes the (higher) percentage of CD25+ cells in the CD4SP population was similar to that found in the DBA/2 parent strain. In B6 + SJL <$>\rightarrow<$> B6SJLF1 mixed bone marrow chimeras we observed a proportion of B6-derived regulatory T cells similar to that observed in the parent strain (Fig. 4B, C). Interestingly, among SJL-derived thymocytes significantly more regulatory T cells were observed than among B6-derived cells but also than in the SJL parent strain (compare Fig. 4C and 1B). The exceptionally high percentage of SJL-derived regulatory thymocytes was also observed in SJL <$>\rightarrow<$> F1 chimeras (not shown). While we currently do not have a satisfactory explanation for the high levels of SJL regulatory T cells in bone marrow chimeras, this result suggests that thymocyte-extrinsic (i.e. environmental) factors can also modulate CD4+CD25+ regulatory T cell development. Whatever the precise explanation is, these results indicate that the different levels of regulatory T cells in the distinct mouse strains studied are caused by thymocyte-intrinsic factors.
The data presented in this paper demonstrate that polymorphic genetic factors quantitatively control intrathymic generation of CD4+CD25+FoxP3+ regulatory T lymphocytes. In the mouse strains studied the distinct regulatory T cell levels are caused by differences in their thymic differentiation from immature precursors. Moreover, we show that thymocyte-intrinsic factors modulate regulatory T cell development. Finally, we report that the genes responsible for modulation of regulatory T cell development in the mouse strains studied are located outside the MHC locus.
Several hypotheses may explain our observation that thymocyte-instrinsic genetic factors cause quantitative differences in regulatory T cell differentiation. They may be caused by quantitative differences in commitment to the regulatory T cell lineage. The gene encoding FoxP3, located on the X chromosome (http://www.ensembl.org/Mus_musculus/geneview?gene=ENSMUSG00000039521), would therefore be among the candidate genes (see Introduction). Another candidate gene would be Notch3. Transgenic expression of a constitutively active form of Notch3 also leads to strongly increased thymic generation of regulatory T cells 43. However, Notch3 is closely linked to the MHC locus (http://www.ensembl.org/Mus_musculus/geneview?gene=ENSMUSG00000038146) and is therefore unlikely to be involved in the differences in regulatory T cell development in the inbred mouse strains reported here.
Alternatively, differences in regulatory T cell positive and/or negative selection may be responsible. While initial reports suggested that thymic CD4 vs. CD8 lineage commitment is independent of TCR specificity, more recently it has become clear that selection mechanisms are responsible. The processes of lineage commitment and selection therefore actually seem to be very closely linked 44. In a still unresolved manner, TCR-mediated signals appear to control expression of Th-POK, a zinc finger transcription factor, as well as of the chromatin remodeling protein Runx, recently identified as binary switches regulating CD4 vs. CD8 lineage commitment, respectively 45–47.
Also the distinct proportions of regulatory cells among mature CD4+ thymocytes may be a consequence of differences in thymic selection and/or lineage commitment. Since thymocyte-intrinsic factors determine quantitative variations in regulatory T cell development, adhesion or signaling molecules may be involved. These molecules would probably also play important roles in function of peripheral regulatory T cells. Thus, one of the many candidate regions is the diabetes susceptibility locus Idd5. Within this locus three genes are located that encode proteins expressed by regulatory T cells: CD28, CTL-associated antigen (CTLA)-4, and inducible costimulator protein (ICOS). CD28 is known to play a crucial role in regulatory T cell development and homeostasis 48–50. Very closely linked is the gene encoding CTLA-4, which is critically involved in regulatory T cell function 51. Moreover, Ctla-4 is a diabetes susceptibility gene in humans 52, and has been reported to modulate thymic negative selection of effector T cells in mice 53. A third gene within the Idd5 locus, ICOS, is also known to play an important role in regulatory T cell function 18, 54. However, none of these genes has thus far been shown to modulate regulatory T cell development.
Identification of the responsible gene(s) may also reveal entirely novel factors critically involved in regulatory T cell development and function, and thus enable better understanding of these processes. Such factors may also constitute unique markers for regulatory T cells, which have thus far proven elusive, and become clinical targets. Whatever the precise explanation for genetic modulation of regulatory T cell development is, it may have important consequences for regulatory T cell repertoire and/or function and thus modulate susceptibility, e.g. to autoimmune diseases. It would therefore be important to assess regulatory T cell differentiation in autoimmune-prone animals, and to identify responsible genetic loci.
Materials and methods
B6, SJL, DBA/2, DBA/1, BALB/c, (B6 × DBA/2)F1 (B6D2F1), and (B6 × SJL)F1 (B6SJLF1) females of 5–7 wk of age were purchased from Janvier (Le Genest St Isle, France). B10 mice were purchased from Charles River (Les Oncins, France). B10.D2, B10.S, and B10.D1 (B10.Q) mice were bred in our facilities. All experiments involving animals were performed in compliance with the relevant laws and institutional guidelines (INSERM; approval No. 31-13, ethical review No. MP/02/32/10/03).
The following Ab and secondary reagents were used for phenotypic analysis: PE-Cy7- or allophycocyanin-labeled anti-CD4 (GK1.5), FITC- or allophycocyanin-labeled anti-CD8 (53.6.7), allophycocyanin- or PE-labeled anti-CD25 (PC61), FITC-labeled anti-HSA (M1/69), FITC-labeled anti-CD69 (H1.2F3), FITC-labeled anti-TCRβ, FITC-labeled anti-CD45.1, PE-labeled anti-FoxP3, FITC-labeled anti-CD5.1, PE-Cy5.5-labeled streptavidine (eBioscience, San Diego, CA). Biotin-labeled anti-GITR was purchased from R&D (Lille, France).
Bone marrow chimeras
Bone marrow from femurs and tibias was collected in Dulbecco's modified Eagle's medium supplemented with 10% FCS. Thy1+ cells were eliminated using AT83 hybridoma supernatant and rabbit complement (Saxon Europe, Suffolk, UK). Cells from each donor were injected intravenously into lethally γ-irradiated hosts (8.5 Gy; 137Cs source, 6.3 Gy/min) that were kept on antibiotic-containing water (0.2% of Bactrim; Roche) for the complete duration of the experiment (6 wk).
Thymocytes or PBMC were incubated 30 min on ice in 2.4G2 (anti-FcγR mAb) hybridoma supernatant. Cells were then incubated 20 min with saturating concentrations of Ab. Intracellular FoxP3 staining on CD8-depleted thymocytes (using anti-CD8 mAb 31 M and complement) was performed according to the instructions of the manufacturer. Labeled cells were analyzed using a FACSCalibur cytometer and CellQuest software (BD Biosciences, San Diego, CA).
BrdU incorporation studies
Mice were continuously exposed to the thymidine analogue BrdU (0.8 mg/mL) in their drinking water. Extracellular staining of thymocytes with mAb against CD4, CD8, and CD25 was performed as described above. Cells were fixed, permeabilized, and stained with FITC-labeled anti-BrdU using the BrdU Flow Kit (BD Pharmingen, Heidelberg, Germany).
Statistical significance of the data was analyzed using Student's t-test.
We thank the staff of the IFR30 animal facility, and in particular Maryline Calise, for expert animal husbandry, and Drs. Gilbert Fournié and Jean-Charles Guéry for critical reading of the manuscript.