SATB1 Plays a Critical Role in Establishment of Immune Tolerance

Identity of T cells is established by T cell–specific gene regulatory networks, which are composed of multiple transcription factors (1). T cell development starts in the thymus when multipotent progenitors or common lymphoid progenitors in the bone marrow (BM) transit to the thymus through blood flow and are recognized as CD4⁻CD8⁻ double-negative (DN) thymocytes (2–4). The most immature thymocytes, called early T cell progenitors (ETPs), in the DN thymocyte fraction have not only T cell potential but also B cell and myeloid potential (5, 6). In the thymic microenvironment, Notch signaling is activated in these ETPs, which initiates T cell differentiation and shuts off developmental potential to other lineages (7–9). T lineage commitment is established during the DN2 stage, more specifically at the transition from the DN2a to the DN2b stage (10, 11). Recombination of the TCRβ genes is completed in the DN3 stage. Only DN3 cells with productive TCRβ-chains can transit to the CD4⁺CD8⁺ double-positive (DP) stage after β-selection (12). DP cells rearrange the TCRα-chain gene, and their fate is determined by two subsequent selection events, positive and negative selection (2, 13). DP cells that successfully express TCRαβ-chains and that recognize MHC–peptide complexes with moderate affinity are positively selected and can advance to the next stage, either the CD4⁺ or CD8⁺ single-positive (SP) stage (14–16). If the TCR on the DP cells binds to MHC–peptide complexes with high affinity, these cells are autoreactive T cells and die by apoptosis (17–19). This elimination process is termed negative selection and is indispensable for establishment of central tolerance (13, 20, 21). Protection of DP cells from apoptosis after positive selection, which is supported at least in part by the IL-7/IL-7R system, is important in the transition from DP to SP cells (22). Because intrathymic T cell development is a multiple-step and complicated process, elucidation of gene expression patterns and their regulation by a network of multiple transcription factors is necessary to fully understand how T cell development and immune tolerance is regulated at the molecular level.

Transcription factors bind to cis-elements in the promoters of target genes and activate transcription of these genes (1). Because not only individual transcription factors but also a set of transcription factors that form transcriptional networks govern the identity of developing T cells at the various maturational stages in the thymus, it is important to understand how global epigenetic regulation of genes is ensured (1, 23). Several nuclear proteins are known to possess functions that regulate chromatin structure. One such protein family is the SWI/SNF–chromatin remodeling complex, which moves or removes nucleosomes and regulates nucleosome positioning and density at the proper genomic location (24). When the functions of SWI/SNF complexes are inhibited, the expression of CD4 and CD8 molecules in developing T cells in the thymus are misregulated (25), which demonstrates that the regulation of chromatin structures by SWI/SNF complexes plays a role in proper gene expression in T cells. However, the action of SWI/SNF complexes is necessary for a broad range of cell types including pluripotent stem cells (26). Therefore, the existence of another factor that regulates epigenetic gene expression in a T-lineage–specific manner has been postulated.

Special AT-rich binding protein 1 (SATB1) is an NF that binds AT-rich sequences (also known as base unpairing regions) and can function as a genome organizer (27, 28). SATB1 forms complexes with SWI/SNF factors and positively and negatively regulates the expression of a vast number of genes (28–30). SATB1 expression is observed in hematopoietic stem cells (HSCs), and its expression is abundant in thymocytes (31, 32). Accordingly, SATB1 and its target genes are involved in various cellular functions in different types of blood cells. For example, it has been demonstrated that SATB1 is necessary for the self-renewal of HSC (32). SATB1 also plays a role in lymphoid lineage specification and/or commitment (31). In SATB1 null mice, however, a dramatic change in cell phenotype is observed in the thymus, where T cell development is mostly stagnated at the DP stage (30). In addition, it has been reported that downregulation of SATB1 is necessary for exhibition of proper inhibitory actions by regulatory T (Treg) cells, which negatively regulate immune responses and are major players in establishment of peripheral tolerance (33). However, because SATB1 is expressed not only by hematopoietic cells but also by other types of cells such as neuronal cells, and because SATB1 null mice die by 3 wk after birth, it is difficult to investigate the precise role of SATB1 in T cell development and function (30).

In this study, we demonstrated that T cell development is severely impaired in the absence of SATB1 by using a mouse model in which the SATB1 gene was conditionally deleted in hematopoietic cells. These SATB1 conditional knockout (cKO) mice were created by crossing SATB1 floxed mice with Vav-Cre transgenic (TG) mice (34, 35). We showed that the T cell developmental defect in SATB1cKO mice is mainly due to dysfunctional positive selection, resulting in stagnation at the transition stage from the DP to the CD4⁺ or CD8⁺ SP stage. Notably, the mortality rate of SATB1cKO mice was higher than that of wild-type (WT) littermates. One cause of death of SATB1cKO mice may be autoimmune manifestation because the concentration of autoantibodies in the serum was increased in SATB1cKO mice after 16 wk of age. In addition, inflammatory lesions or lymphocyte infiltrates were observed in multiple organs in SATB1cKO mice. We also found that negative selection and a regulatory function of Treg cells were impaired in SATB1cKO mice, suggesting that SATB1 plays an important role in regulation of the gene expression that is critical for establishment of immune tolerance.

SATB1^fl/fl mice were generated as described (34). Vav-Cre mice (35), OT-I and OT-II TCR TG mice (36, 37), and RIPmOVA mice (38) were purchased from The Jackson Laboratory. Lck-Cre mice (39) were obtained from the Laboratory Animal Resource Bank at Nibiohn. HY-TCR TG mice were provided by W. Zhang at Duke University. C57BL/6 mice were purchased from CLEA Japan. RAG2^−/− and C57BL/6 (CD45.1) mice were maintained at the animal facility in Toho University. All mice were maintained on a C57BL/6 background and under specific pathogen-free conditions at the Toho University School of Medicine animal facility. All experiments using mice received approval from the Toho University Administrative Panel for Animal Care (15-55-262) and Recombinant DNA (15-52-260). Mice were used at 5–10 wk of age unless otherwise specified in the text.

Total RNA from various cell populations was extracted using the TRIzol reagent (Invitrogen). The High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems) was used for first-strand synthesis. Quantitative RT-PCR was performed according to the protocol of TaqMan gene expression assay kits (Applied Biosystems) with an ABI 7500 Fast system (Applied Biosystems) using the following primers: Hprt, Mm00446968_m1, and SATB1, Mmoo485920_m1. Results were normalized to the expression of Hprt mRNA.

Abs used for cell-surface and intracellular staining were as follows: anti–CD4-PE–Cy7 (GK1.5), anti-CD5–FITC (53-7.3), anti-CD8–Pacific Blue (53-6.7), anti-CD24–PE (M1/69), anti-CD25–PE (PC61), anti-CD44–PE (IM7), anti-CD45.2–allophycocyanin (104), anti-CD62L–FITC (MEL-14), anti-TCRβ–APC (H57-597), anti-Vα2–PE (B20.1), anti-Vβ5.1 5.2–FITC (PE-MR9-4), and anti-B220–FITC (RA3-6B2) were obtained from BioLegend. Anti-CD45.1–FITC (A20), anti-CD69–FITC (H1.2F3), anti-HY–APC (T3.70), anti–c-Kit–APC (2B8), anti-Thy1.2–PE (53-2.1), and anti-Foxp3–FITC (FJK-16a) were purchased from eBioscience. Anti-TCRβ–APC Abs (H57-597) were obtained from Tonbo Biosciences.

For staining, erythrocytes were depleted from a single-cell suspension derived from the various organs specified in the figure legends. After the cells were washed with FACS buffer (2% FBS in 1× PBS with 0.05% NaN₃), they were incubated with a mixture of fluorescence-conjugated Abs on ice for 20 min and then washed twice with FACS buffer. A Cytofix/Cytoperm kit (BD Biosciences) was used for intracellular staining by following the instructions of the kit. Flow cytometric analysis was performed using FACSCanto II (BD Biosciences) and FACSAria III (BD Biosciences). Cell sorting was done using FACSAria III (BD Biosciences). Dead cells were excluded as 7-aminoactinomycin D–positive (BD Biosciences) cells. Data were analyzed using FlowJo software (Tree Star).

BM cells isolated from SATB1cKO (CD45.2) mice, their WT littermates (CD45.2), and congenic C57BL/6 (CD45.1) mice were stained with anti-B220–FITC, anti-Thy1.2–PE, and anti–c-Kit–APC Abs. Subsequently, c-Kit⁺Thy1.2^−/loB220⁻ hematopoietic progenitor cells (HPCs) with HSCs were purified using cell sorting on an FACSAria III (BD Biosciences). HPCs (1 × 10⁶ cells) from SATB1cKO mice or a mixture of HPCs from SATB1cKO mice and congenic C57BL/6 (CD45.1) mice at a ratio of 1:1 were i.v. injected into lethal dose–irradiated (900 rad) C57BL/6 (CD45.1) mice. Splenocytes from these mice were analyzed on FACS machines at 8 wk after injection of the HPCs.

Mouse tissues were fixed in 10% formalin solution (Wako) and embedded in paraffin. Sections (6 μm) were stained with H&E and observed using a BX61 microscope (Olympus).

Anti-dsDNA Abs in the serum were measured using a Mouse Anti-dsDNA ELISA kit (Shibayagi).

CD25⁻CD4⁺ conventional T (Tconv) cells and CD25⁺CD4⁺ Treg cells from spleen and peripheral lymph nodes were purified by cell sorting. Responder Tconv cells (CD45.1, 2 × 10⁵ cells) were labeled with CFSE according to the manufacturer’s instructions (Invitrogen) and cultured for 4 d with irradiated splenocytes (1 × 10⁵ cells; CD45.2) and anti-CD3ε Ab (5 μg/ml; 145-2C11; eBioscience) in the presence or absence of CD25⁺CD4⁺ Treg cells (CD45.2) either from SATB1cKO mice or their WT littermates. CFSE levels on Tconv cells were then measured by gating on CD45.1⁺ cells in FACS analysis.

Rag2^−/− mice were injected i.v. with CD25⁻CD4⁺CD45RB^hi T cells (2 × 10⁵) in the presence or absence of CD25⁺CD4⁺ Treg cells (2 × 10⁵). The mice were weighed weekly, assessed for clinical signs of colitis for 5 wk, and then killed to determine the presence of colitis in colon sections. Colons were fixed in 10% formalin solution and embedded in paraffin. Sections (6-μm thick) were stained with H&E.

WT and SATB1 cKO mice were i.p. injected with 25 μg anti-CD3ε Abs in 100 μl PBS. After 48 h, thymocytes were isolated, and the expression of CD4 and CD8 was analyzed using FACS.

Statistical analysis was performed with the mean difference hypothesis of Student t test or the Mann–Whitney U test with assumption of different variances and a confidence level of 95%.

SATB1 expression is observed in hematopoietic cells as early as the HSC stage (32). In this study, real-time quantitative RT-PCR analysis of thymus tissue indicated that, although all thymocytes of WT mice expressed SATB1 mRNA, the expression level was the highest at the DP stage and decreased thereafter along with maturation (Fig. 1A). This result implies that SATB1 expression is important at the DP stage during intrathymic T cell development.

To investigate the effect of SATB1 deficiency on T cell development, we generated SATB1cKO mice by crossing mice with an SATB1 floxed allele (SATB1^fl/fl mice) with Vav-Cre TG mice (34, 35, 40). As in the case of SATB1 null mice (30), T cell development was severely impaired in the SATB1cKO mice (Fig. 1B, 1C). Thymocyte numbers in SATB1cKO mice (0.55 ± 0.18 × 10⁸; p < 0.05) were ∼40% of those in WT littermates (1.3 ± 0.14 × 10⁸). Reduction in the number of thymocytes in SATB1cKO mice was significant in the SP population (Fig. 1B, 1C), suggesting that transition from the DP to the SP stage, in which positive selection is required, is impaired (13).

The decrease in the number of thymocytes in SATB1cKO mice was observed even for the most immature population of thymocytes, the ETPs (Fig. 1D). A significant reduction was observed in the number of common lymphoid progenitor cells but not in HSCs in BM of SATB1cKO mice compared with WT mice (Supplemental Fig. 1A). These findings are consistent with our previous report that SATB1 deficiency might affect specification and/or commitment of the lymphoid lineage (32). Indeed, no significant difference was observed in the number of ETPs between WT mice and SATB1 floxed mice with Cre transgenes under the control of the Lck proximal promoter (Lck-Cre), which are active in thymocytes after the DN2 stage (41), although the number of ETPs in Lck-Cre-SATB1^fl/fl mice was lower than that in WT mice (Supplemental Fig. 2A).

We next analyzed peripheral T cells in SATB1cKO mice and their WT littermates. Reflecting the observed reduction in the number of SP thymocytes, both CD4⁺ and CD8⁺ T cells were significantly decreased in the spleen of SATB1cKO mice (Fig. 2A). As in the case of SATB1 null mice (30), CD4⁺CD8⁺ DP cells were observed in the spleen of SATB1cKO mice (Fig. 2A), suggesting that SATB1 is involved in the downregulation of CD4 and/or CD8 molecules in peripheral T cells. In contrast to the T cell defect in SATB1cKO mice, the B cell number was not significantly different between SATB1cKO and WT littermates (Fig. 2B).

T lymphopenia may lead to activation of T cells as the result of cell expansion by homeostatic proliferation (42). This also appears to be the case for SATB1cKO mice because the number of CD62L^highCD44^−/lo naive T cells was drastically reduced and that of CD62L⁻CD44^high effector/memory T cells was significantly increased in both CD4⁺ and CD8⁺ populations in the spleen of SATB1cKO mice (Fig. 2C, 2D). The mechanism that underlies this phenomenon is unclear; however, whereas naive T cells in WT mice showed from low to negative expression of CD44, naive T cells in the SATB1cKO mice were mostly negative for CD44 (Fig. 2C). Activated T cells in SATB1cKO mice seem to enhance humoral immunity because germinal center B cells, which are positive for GL7, were detected in the spleen and lymph nodes of SATB1cKO mice (data not shown).

To address whether T cell activation in SATB1cKO mice was truly due to homeostatic proliferation under a lymphopenic condition, we generated chimeric mice by injecting BM cells from SATB1cKO (CD45.2) and WT C57BL/6 (CD45.1) mice into lethally irradiated WT mice (CD45.1). At 6 wk after BM reconstitution in the mice injected with SATB1cKO BM alone, ∼10% of the CD4⁺ cells derived from SATB1cKO BM were naive cells (Fig. 3A, left panel, Fig. 3B). When SATB1cKO BM cells were injected with WT BM cells, the percentage of naive CD4⁺ T cells was increased to ∼20% (Fig. 3A, right panel), which is comparable to the number of CD4⁺ naive T cells from WT BM (Fig. 3B). This result demonstrates that the activated phenotype of T cells in SATB1cKO mice is a result of homeostatic proliferation in a T lymphopenic condition.

We observed a marked decrease in the number of CD4⁺ and CD8⁺ SP cells in the thymus of SATB1cKO mice (Fig. 1B). To determine if this decrease in the number of SP thymocytes was a result of a defect in positive selection, we examined the developmental fate of thymocytes expressing fixed TCRs. We crossed SATB1cKO mice with OT-I (36) and OT-II (37) TCR TG mice. These TCRs recognize OVA peptides presented by MHC class I and II, respectively, and therefore, OT-I⁺ and OT-II⁺ DP cells give rise to CD8⁺ and CD4⁺ T cells, respectively, upon positive selection in the TCR TG mice (Fig. 4A, 4B, left panels). However, OT-I⁺ DP thymocytes in a SATB1cKO background did not differentiate into CD8⁺ SP cells (Fig. 4A, middle and right panels). Similarly, fewer OT-II⁺CD4⁺ SP cells were observed in the SATB1cKO background than in the SATB1 sufficient OT-II⁺ mice (Fig. 4B), although the developmental block from the DP to the SP stage was less severe in OT-II⁺ SATB1cKO mice than in OT-I⁺ SATB1cKO mice. These results clearly demonstrate that positive selection is severely impaired in the absence of SATB1.

In accordance with this defect in positive selection, upregulation of CD69 and TCRβ, which normally occurs after successful positive selection at the DP stage, was impaired in SATB1cKO mice (Fig. 4C). Finally, we analyzed the CD5 expression level on thymocytes from SATB1cKO mice and their WT littermates. The expression of CD5 was obviously impaired in SATB1cKO mice, especially in DP thymocytes (Fig. 4D). This decrease in CD5 expression persisted at the CD4⁺CD8^int stage in SATB1cKO mice (Fig. 4D). The CD5 expression level on thymocytes, especially on CD4⁺CD8^int cells, is known to reflect the avidity of the interaction of the TCR with self-peptide/MHC complexes (43). Therefore, impairment of positive selection might be due, at least in part, to insufficient TCR engagement of DP thymocytes in SATB1cKO mice.

During the breeding and maintenance of SATB1cKO mice and their littermates, we noticed that the SATB1cKO mice died earlier than their WT littermates (Fig. 5A). We found that levels of autoantibodies, such as anti-dsDNA Abs, in the serum of SATB1cKO mice were higher than those of WT mice (Fig. 5B), suggesting that SATB1cKO mice are prone to autoimmunity. In addition to the presence of autoantibodies, apparent inflammatory cell infiltration was observed in various organs of the SATB1cKO mice (Fig. 5C). It seems that this autoimmune disorder is caused by T cells in the absence of SATB1 because autoimmune phenotypes were observed in Lck-Cre-SATB1^fl/fl mice, in which SATB1 is deleted only in T cells (Supplemental Fig. 2B, 2C).

Immune tolerance is the key for blocking autoimmune manifestation in healthy conditions. Once immune tolerance is disturbed, the bodies’ immune system may attack itself and various organs may get damaged, resulting in autoimmune diseases (44). Treg cells, which are CD4⁺ T cells that express the transcription factor Foxp3, play a central role in establishment of peripheral tolerance and immune homeostasis (45). Treg cell deficiency and dysfunction may cause a T cell–mediated autoimmune disorder (45). Therefore, to determine the involvement of Tregs in the autoimmunity of the SATB1cKO mice, we first compared the frequency and the cell number of splenic Treg cells in SATB1cKO mice with those of WT littermates. In SATB1cKO mice at 5–10 wk of age, the frequency of Treg cells was higher than that in WT mice (Fig. 6A); however, the absolute number of splenic Treg cells in SATB1cKO mice was less than that of WT mice (Fig. 6B). A reduction in the number of Treg cells was also observed in inguinal and mesenteric lymph nodes of SATB1cKO mice compared with WT mice (Fig. 6B). Similarly, slightly fewer Treg cells were present in older (24-wk-old) SATB1cKO mice than in WT mice (Supplemental Fig. 1B). Also, no significant difference in the number of Treg cells was observed between healthy and sick SATB1cKO mice (data not shown). However, Treg cell numbers were significantly different between SATB1cKO and WT mice at 2 wk after birth (Supplemental Fig. 1B), suggesting that the loss of SATB1 may affect Treg cell development.

We next analyzed the suppressive ability of SATB1-deficient Treg cells in vitro. To determine whether SATB1-deficient Treg cells can suppress the proliferation of CD4⁺CD25⁻ Tconv cells, we cultured activated Tconv cells from WT congenic mice (CD45.1) with or without CD4⁺CD25⁺ Treg cells, all of which expressed Foxp3, from either WT or SATB1cKO mice (CD45.2) for 4 d. After culture, we analyzed the proliferation of the Tconv cells by CFSE dilution on a FACS machine and found that the suppressive ability of SATB1-deficient Treg cells was slightly reduced compared with that of WT Treg cells (Fig. 6C, 6D). We also examined suppressive abilities of Treg cells derived from SATB1cKO mice in in vivo settings by using a T cell–induced colitis model. Upon transfer of naive CD4⁺CD45RB^high cells into RAG2^−/− mice, the mice became wasted and developed colitis, whereas cotransfer of either WT or SATB1-deficient CD4⁺CD25⁺ Treg cells prevented development of T cell–induced colitis (Fig. 6E, 6F). These results suggest that although the suppressive ability of SATB1-deficient Treg cells was not clearly observed in in vivo assays, Treg function is impaired in SATB1cKO mice in terms of reduction in the number and impairment of in vitro suppressive ability of Treg cells in SATB1cKO mice.

Because SATB1cKO mice displayed a modest defect in a function of Treg cells, we next examined whether negative selection, which is required for establishment of central tolerance, is defective in these mice. For this purpose, we crossed SATB1cKO mice with HY-TCR TG mice (14). Because female mice do not have HY Ags, HY-TCR⁺ DP cells on the WT background were positively selected and gave rise to CD8⁺ T cells (Supplemental Fig. 3, left panel). Because positive selection is perturbed in the absence of SATB1, fewer CD8⁺ SP cells were observed in the HY-TCR⁺ SATB1cKO female mice versus the HY-TCR⁺ WT mice (Supplemental Fig. 3, right panel). In the male mice, most of the HY-TCR⁺ thymocytes were deleted because of the presence of HY Ags, resulting in very few CD8⁺ SP cells in the thymus (Fig. 7A, left panel). On a SATB1cKO background, however, more HY-TCR⁺ CD8⁺ SP cells were observed in the thymus than on the WT background (Fig. 7A, 7B). These results suggested that negative selection was impaired in SATB1cKO mice. It should be noted that there were two different CD4/CD8 staining patterns in HY-TCR⁺ SATB1cKO mice; one pattern was similar to that of the HY-TCR⁺ WT mice (Fig. 7A, pattern 1), and in the other pattern, the CD8⁺ SP population was more prominent (pattern 2). The cause of these different thymocyte staining patterns in HY-TCR⁺ SATB1cKO mice is unclear; we did not observe a significant difference in the peripheral phenotype of these mice compared with WT mice (data not shown). Whatever the cause of these different staining patterns, the results shown in Fig. 7A and 7B suggest that negative selection is impaired in the absence of SATB1 and that more HY-TCR⁺ thymocytes have escaped from clonal deletion in male mice.

We also examined whether negative selection of CD4⁺ T cells was impaired in SATB1cKO mice. For this purpose, we crossed SATB1cKO mice to OT-II × RIP-mOVA double TG mice according to the method used by Anderson et al. (46). RIP-mOVA TG mice express a membrane-bound form of OVA under the control of the insulin promoter (38). In addition to expression in the pancreatic islet, this transgene is also expressed by medullary thymic epithelial cells in the thymus under the control of a nuclear protein, AIRE. Large-scale deletion of OT-II TCR⁺ thymocytes is seen when both the OT-II TCR transgene and RIP-mOVA transgene are present (46). Indeed, we observed significant but partial deletion of CD4⁺ SP cells in the OT-II TG mice in the presence of RIP-mOVA (from 87.6 to 54%; Fig. 7C). As shown above, fewer CD4⁺ SP cells were observed in OT-II TG-SATB1cKO mice than in OT-II TG mice on the WT background due to a defect in positive selection (Fig. 7C). In the presence of RIP-mOVA, however, the number of CD4⁺ SP cells was increased in the OT-II SATB1cKO mice (Fig. 7D). These results demonstrate that SATB1 is necessary for establishment of central tolerance. Although self-reactive T cells should be in the periphery of SATB1cKO mice, we could not reproduce autoimmune phenotype in the naive mice by transferring SATB1-deficient T cells, maybe because of low activation potential of T cells in SATB1cKO mice due to weak TCR signal strength as shown in Supplemental Fig. 4. Because T cells in patients with systemic lupus erythematosus have some abnormalities, such as T cells’ reduced ability to produce cytokines (47), T cells in SATB1cKO mice might share some characteristics with T cells in human autoimmune patients.

Finally, we examined the sensitivity of the DP thymocytes in SATB1cKO mice to clonal deletion induced by strong TCR signals. We i.p. injected anti-CD3 Abs into SATB1cKO and WT littermates and analyzed thymocytes from each mouse 4 d later. Although almost all DP thymocytes were depleted in WT mice by this treatment, DP cells in SATB1cKO mice were resistant to apoptosis induced by anti-CD3 Ab treatment (Fig. 8). Therefore, dysregulation of negative selection as well as reduction in the suppressive function of Treg cells in the absence of SATB1 might be causes of autoimmune disorder in the SATB1cKO mice. Overall, the results in this study demonstrate that SATB1 plays a role in the establishment of immune tolerance in T cells.

In this study, we demonstrated that SATB1 is required for both positive and negative selection events during intrathymic T cell development. As previously reported by Alvarez et al. (30) for SATB1 null mice, T cell development was severely impaired at the DP stage in the absence of SATB1. Although no obvious T cell developmental defect at the DN stage is seen in SATB1 null mice, we showed that, in SATB1cKO mice, T cell development was obstructed even at the stage of the earliest thymocyte population, the ETP population (Fig. 1D). SATB1 plays a role in the maintenance of self-renewal potential in HSC (32) and in specification/commitment of HSC/multipotent progenitors to the lymphoid lineage as we previously reported (31), T cell development might be affected at a time point as early as the HSC stage in SATB1cKO mice. However, we did not observe a significant difference in the number of HSCs in the BM between SATB1cKO mice and their WT littermates, which may not be consistent with a previous report by Will et al. (32). Because SATB1 null mice die by 3 wk after birth, HPCs from fetal liver or BM in infants were used in the experiments in the previous report, whereas we analyzed adult mice in this study. Therefore, the difference between the previous studies and our present report might stem from the ages of the mice used in the investigations. More detailed analysis is necessary for clarification of the precise requirement of SATB1 at the various maturational stages of hematopoietic cells.

Because T cell development of the SATB1 null mice is severely blocked at the DP stage in the thymus, it was hypothesized that SATB1 deficiency might lead to impairment of positive selection (48). In this study, we demonstrate that positive selection is indeed dysfunctional in the absence of SATB1 (Fig. 4). It should be noted that this developmental arrest is more severe for CD8⁺ T cells than for CD4⁺ T cells. Signal cascades via the TCR are triggered by a cytoplasmic tyrosine kinase, Lck, which associates more efficiently with CD4 than with CD8 (49). Therefore, different usage of Lck between MHC class I– and class II–restricted positive selection events might affect the cell fate of DP thymocytes in the absence of SATB1. Expression patterns of various cytokine receptor subunits in thymocytes are changed in SATB1 null mice compared with WT mice (30). Because intrathymic cytokines, such as IL-7, affect the specification of CD4 and CD8 lineage choice at the DP stage (50), it is also possible that this dysregulation of cytokine receptor expression might result in different severity in the phenotype of CD4⁺ T and CD8⁺ T cells in SATB1cKO mice. Further investigation is necessary for identification of the reason why SATB1 is required for the process of proper positive selection events.

A significant finding of this study was that SATB1 plays a role in establishment of immune tolerance. SATB1cKO mice died earlier than WT mice (Fig. 5A) and displayed multiple characteristics of autoimmune prone mice, such as increased levels of autoantibodies and infiltration of immune cells into various organs (51, 52). Thus, impairment of negative selection might be one of the causative reasons of development of autoimmune disorders in the absence of SATB1. Hwang et al. (53) have demonstrated that a weak TCR signal in mice with a mutated ITAM in the TCRζ chains leads to impairment of both positive and negative selection. Therefore, it is possible that dysregulation of both positive and negative selection in SATB1cKO mice is also caused by weak TCR signal strength at the DP stage. However, it has been reported that symptoms of autoimmune diseases are not observed in TCRζ mutant mice because of increased numbers of Foxp3⁺ Treg cells (53). SATB1cKO mice have impairment of Treg functions compared with WT mice at adulthood (Fig. 6B). Therefore, Treg cell numbers and functions might be critical for prevention of autoimmune diseases, when autoreactive T cells are increased in the periphery.

Reduction in the number of Treg cells was significantly different from WT in SATB1cKO mice at 2 wk after birth (Supplemental Fig. 1B). Because it has been recently demonstrated that there are distinct Treg populations that are generated in an age-dependent manner and have different TCR repertoires (54), it might be possible that SATB1-dependent and SATB1-independent Treg cells are present. In addition to the decreased number, the suppressive functions of Treg cells were slightly impaired in the absence of SATB1 (Fig. 6C, 6D). However, our preliminary results suggest that Treg cells in SATB1cKO mice express higher levels of GITR than the Treg cells of WT mice. GITR is observed in preactivated Treg cells that have higher suppressive functions than resting Treg cells (55–57). It is therefore not clear why the suppressive function of SATB1-deficient Treg cells is impaired. Further investigation is necessary for clarification of the roles of SATB1 in development and in the functions of Treg cells.

We demonstrated in this study that SATB1 deficiency in hematopoietic cells results not only in T cell deficiency but also in autoimmune manifestations. It has been shown that homeostatic proliferation of T cells is an important event for development of autoimmune diseases due to an increase in the number of autoreactive T cells and the activation of such T cells (58, 59). Therefore, the T lymphopenia that was observed in the SATB1cKO mice might accelerate the onset of autoimmune disorder even as early as 16 wk of age in these mice under the condition in which Treg cell development is also affected in the absence of SATB1. Because SATB1 globally regulates the expression of numerous genes by modification of chromosomal structure (28, 29), clarification of the gene regulatory network modulated by SATB1 in developing and mature T cells should provide new insights into understanding the establishment of immune tolerance, which may lead to the development of a new treatment of autoimmune diseases.

We thank Dr. Weiguo Zhang (Duke University Medical Center) for HY-TCR TG mice.

The authors have no financial conflicts of interest.