A20 Restrains Thymic Regulatory T Cell Development

Animals

Mice with cell type–specific deficiency of A20 in T cells were generated by breeding A20^F/F mice (25) with CD4Cre transgenic mice (37) as previously described (38). C57BL/6 (H-2Kb, Thy-1.2) and BALB/c (H-2Kd, Thy-1.2) mice were purchased from Janvier Labs. Mice were between 6 and 12 wk of age at the onset of experiments if not indicated otherwise. Animal protocols were approved by the local regulatory agency (Regierung von Oberbayern, Munich, Germany).

Allogeneic hematopoietic stem cell transplantation and GVHD

Allogeneic bone marrow (BM) transplantations were performed as previously described (39). In brief, female, 6- to 8-wk-old BALB/c recipients were lethally irradiated with total body irradiation (TBI; 2 × 4.5 Gy). Directly after the second irradiation step mice received 5 × 10⁶ T cell– depleted C57BL/6 wild type (WT) BM cells for reconstitution. T cell depletion of freshly isolated BM cells was performed using CD90.2 microbeads (Miltenyi Biotec). In addition, some mice received 2 × 10⁵ CD8⁺ T cells and 2 × 10⁵ CD4⁺ CD25⁻ effector T cells with or without 6 × 10⁵ CD4⁺ CD25⁺ T_reg cells as indicated in the figure legends. All T cell subsets were isolated from complete splenocytes using the T_reg cell isolation kit (Miltenyi Biotec) and the CD8 T cell isolation kit (Miltenyi Biotec) according to the manufacturer's instructions.

Syngeneic hematopoietic stem cell transplantation and competitive T cell development

CD45.1⁺ CD45.2⁺ (double-positive) C57BL/6 WT mice were used for competitive T cell development experiments. Conditioning therapy of C57BL/6 recipient mice before the syngeneic BM transplantation was performed as previously described (40). After conditioning therapy (2 × 5.5 Gy), recipient mice received 2.5 × 10⁶ T cell–depleted C57BL/6 WT CD45.1⁺ CD45.2⁻ BM cells. In addition, recipient mice received either 2.5 × 10⁶ A20^F/F CD4^cre− or 2.5 × 10⁶ A20^F/F CD4^cre+ T cell–depleted C57BL/6 BM cells, expressing only CD45.2, but not CD45.1. Mice were analyzed 3 mo after transplantation. Exact ratios of transplanted and engrafted CD45.1⁺ and CD45.2⁺ BM cells were determined by analysis of CD45.1 and CD45.2 expression of CD4⁻ CD8⁻ cells. The engraftment ratio was calculated for all transplanted animals separately. Further analysis of cell subsets was normalized to this ratio.

Flow cytometry and Abs

All Abs were purchased from eBioscience, BioLegend, and Miltenyi Biotec. For cell viability analysis, a near-infrared Live/Dead reagent (Molecular Probes; Thermo Fisher Scientific) was used according to the manufacturer's instructions. The following Abs were used for cell surface or intracellular staining and flow cytometric analysis: CD4 (GK 1.5/RM4-5), CD8 (53-6.7), CD25 (PC61.5), CD44 (IM7), CD62L (MEL-14), c-Rel (REA397), Foxp3 (FJK16s), GITR (DTA-1), and KI-67 (16A8). Cells were stained in PBS supplemented with 2.5% BSA using fluorochrome-conjugated Abs against specific cell surface markers in different dilutions for 20 min at 4°C. For Foxp3 staining, the Foxp3/transcription factor staining buffer set (eBioscience) was used according to the manufacturer's instructions. Active caspase-3 stain was performed directly before further viability and cell surface staining using the CaspGLOW Fluorescein Active Caspase-3 Staining Kit (Thermo Fisher Scientific) according to the manufacturer's instructions. In vivo cell proliferation was analyzed using the Click-iT Plus EdU (5-ethynyl-2′-deoxyuridine) Alexa Fluor 647 Flow Cytometry Assay Kit (Thermo Fisher Scientific) according to the manufacturer's instructions. The assay was performed 24 h after i.p. injection of EdU. Treatment dose of EdU (50 mg/kg per mouse) was used as previously described (41). Analyses of cell populations were performed by flow cytometry with a FACSCanto II (BD). Data analysis was performed using FlowJo software (Tree Star). Flow cytometry sorting was performed with a FACSAria III (BD Biosciences). The anti-mouse IL-2 Ab (JES6-1A12; BioXCell) was used to neutralize IL-2 in vivo. In indicated experiments, 500 mg of anti–IL-2 was injected i.p. 6 and 3 d before analysis.

Imaging flow cytometry

Thymocytes were subjected to MACS-based depletion of CD8⁺ T cells (CD8 microbeads; Miltenyi Biotec). Cells were then either left untreated or were stimulated with 100 nM PMA (Sigma) and 1 mM ionomycin (Calbiochem) for 30 min at 37°C, 5% CO₂. Extracellular and intracellular Foxp3 staining was performed as described earlier. Intracellular c-Rel staining was performed overnight. DAPI (Thermo Fisher Scientific) was applied as a nuclear stain subsequent to Ab stainings. Stained cell suspensions were acquired on an ImageStreamX Mark II (Amnis; Merck Millipore) imaging flow cytometer. Data analysis was performed using the wizard for nuclear translocation of the IDEAS software (Amnis), where determination of nuclear translocation is based on a similarity score that quantifies the correlation of nuclear stain and translocation probe intensities. High correlation and thus high similarity scores represent strong nuclear translocation, whereas low scores are indicative of cytoplasmic localization (42).

Phospho–NF-κB p65 (S536) phosphoflow

Cells were either left untreated or were stimulated with 100 nM PMA (Sigma-Aldrich) and 1 mM ionomycin (Calbiochem) for 30 min at 37°C, 5% CO₂. After washing the cells with PBS, live/dead cell staining was performed with a fixable viability dye (eBioscience) followed by surface staining. Cells were fixed with 2% PFA for 40 min, permeabilized by two washes with Perm/Wash buffer (eBioscience), and stained overnight with anti-Foxp3 and anti–phospho-NFKBp65 (Ser536) (93H1; Cell Signaling) Abs. After two washes with Perm/Wash buffer, cells were stained with an allophycocyanin-labeled anti-rabbit IgG Ab (A10931; Invitrogen).

Isolation of CD4⁺ CD25⁺ T_reg cells with magnetic cell sorting

CD4⁺ CD25⁺ T_reg cells were isolated from splenic single-cell suspensions using the MACS T_reg cell isolation kit (Miltenyi Biotec) according to the manufacturer's instructions.

T_reg cell suppression assay

MACS-sorted CD4⁺ CD25⁻ T effector cells were cultured in RPMI 1640 medium in a 96-well plate at a concentration of 1 × 10⁵ cells per well. Soluble anti-CD3 (1 μg/ml; clone 2C11; BD Pharmingen) and irradiated APCs were used in a 1:2 ratio. APCs as stimulator cells were prepared from syngeneic mice by depleting the CD4⁺ T cell fraction using anti-CD4 microbeads according to the manufacturer's instructions (Miltenyi Biotec). For suppression, CD4⁺ CD25⁺ MACS-sorted A20^F/F CD4^cre+ or A20^F/F CD4^cre− T_reg cells were added at the indicated ratio to A20^F/F CD4^cre− T effector cells. Proliferation of CD4⁺ CD25⁻ T effector cells was analyzed using the CellTrace Violet Cell Proliferation Kit (Invitrogen) according to the manufacturer's instructions. The percentage of proliferating cells was analyzed after 72 h.

T_reg cell expansion and differentiation of naive CD4⁺ T cells in vitro

Statistics

GraphPad Prism version 6 was used for statistical analysis. Survival was analyzed using the log-rank test. Differences between means of experimental groups were analyzed using the one- or two-tailed unpaired Student t test, corresponding to the distribution shape of our observations. We used ordinary one-way ANOVA for multiple comparisons and always performed Dunnett's test for multiple-test corrections. The applied statistical tests are indicated in the figure legends. For visual clarity, data are shown as mean ± SEM throughout.

Increased numbers of T_reg cells in mice with T cell–specific A20 deletion

To investigate the role of A20 in T cells, we analyzed A20^F/F CD4^Cre+ mice, which specifically lack A20 in T cells. The population of Foxp3⁺ T_reg cells was quantitatively enlarged in thymus, spleen, and inguinal lymph nodes of naive A20^F/F CD4^Cre+ mice (CD4 A20^−/−) as compared with A20^F/F CD4^Cre− mice (CD4 A20^+/+) (Fig. 1A–E, Supplemental Fig. 1A). At the same time, the total thymic CD4⁺ and CD8⁺ T cell population was not significantly different between CD4 A20^−/− mice and controls. In the spleen, both total CD4⁺ and CD8⁺ T cell populations were reduced in CD4 A20^−/− mice (Fig. 1B, 1D, Supplemental Fig. 1B, 1C), which is in line with previous observations (38). To exclude a bias of age on the development of T_reg cells (43), we analyzed splenocytes of young (12-d-old) and adult (50-d-old) CD4 A20^+/+ and CD4 A20^−/− mice. We found frequencies and absolute numbers of T_reg cells to be significantly increased in both young and adult mice (Fig. 1F).

A20-deficient T cells show reduced T_reg cell expansion and differentiation in vitro

To investigate whether this absolute increase of peripheral T_reg cell numbers in CD4 A20^−/− mice was due to enhanced sensitivity of T_reg cells to external proliferation signals, we stimulated FACS-purified CD4⁺ CD25⁺ T_reg cells from CD4 A20^+/+ and CD4 A20^−/− mice in vitro with anti-CD3, anti-CD28, and IL-2 (44). Although IL-2 induced a significant increase in the absolute numbers of both control and A20-deficient T_reg cells after 4 or 8 d of culture, the in vitro expansion of A20-deficient T_reg cells was significantly less pronounced (Fig. 2A, Supplemental Fig. 2A). Next, we wanted to know whether enhanced sensitivity to peripheral differentiation signals of T_reg cells was responsible for the enlarged T_reg cell population in CD4 A20^−/− mice. However, stimulation of FACS-purified CD62L^high CD44^low naive T cells with anti-CD3, anti-CD28, and either TGF-β alone or TGF-β and IL-2 resulted in decreased proportions and absolute numbers of A20-deficient compared with control T_reg cells (Fig. 2B, Supplemental Fig. 2B). In summary, the sensitivity to expansion or differentiation signals of A20-deficient T_reg cells or CD4⁺ T cells in vitro was reduced compared with controls, and thus could not explain the increase of total T_reg cell numbers in CD4 A20^−/− mice.

A20 deficiency in T cells drives thymic T_reg cell development cell-intrinsically

We assumed that an intrinsic effect could account for increased T_reg cell development and enhanced T_reg cell numbers in CD4 A20^−/− animals. To test this hypothesis, we established a model of competitive BM engraftment and T cell development in which CD45.1⁺ CD45.2⁺ double-positive C57BL/6 recipient mice were lethally irradiated (TBI) to eradicate the recipient hematopoietic system. Then T cell–depleted BM cells mixed in a 1:1 ratio from both CD45.1 (WT) and CD45.2 (either CD4 A20^+/+ or CD4 A20^−/−) donors (Fig. 3A) were introduced for hematopoietic reconstitution.

In this model, tracing the genetic markers by FACS analysis allows to quantify whether one of the BM compartments transferred to the host at day 0 (WT plus CD4 A20^+/+ or WT plus CD4 A20^−/−) would outperform another or would favor the development of certain T cell or other immune cell subsets. Ninety days after transplantation, recipients that had received CD45.1 WT + CD45.2 CD4 A20^+/+ BM showed CD45.2 and CD45.1 T_reg cell fractions with a balanced ratio. In contrast, recipients that had received CD45.1 WT + CD45.2 CD4 A20^−/− BM showed an ∼2.6-fold (spleen) and 3.8-fold (thymus) increased ratio between CD45.2⁺ A20^−/− and CD45.1 WT T_reg cell fractions (Fig. 3B, 3C). Ratios of CD45.2/CD45.1 populations of splenic dendritic cells, B cells, and total CD4⁺ T cells did not differ significantly between CD4 A20^+/+ BM or CD4 A20^−/− BM (Fig. 3B). Accordingly, we found increased fractions of Foxp3⁺ T_reg cells within the CD4⁺ T cell compartment in spleen and thymus of mice that had received CD45.1 WT plus CD4 A20^−/− BM compared with mice that had received CD45.1 WT plus CD4 A20^+/+ BM (Fig. 3D), whereas populations of CD11c⁺, B220⁺, and CD4⁺ T cells were similar in both cohorts (Supplemental Fig. 2C). Together with our findings that in vitro T_reg cell differentiation and expansion through common external stimuli were not enhanced in CD4 A20^−/− cells, these data suggest that early events in the BM or thymus favor T_reg cell development in CD4 A20^−/− mice via cell-intrinsic mechanisms.

A20 deficiency reduces the dependence of thymic T_reg cells on IL-2

IL-2 was shown to play a crucial role in thymic T_reg cell development by stabilizing the Foxp3⁺ phenotype and counterbalancing Foxp3-induced apoptosis (45). To investigate effects of IL-2 signaling on A20^−/− T_reg cells, we used in vivo application of an anti–IL-2 Ab that antagonizes effects of IL-2 (46). We injected anti–IL-2 into naive CD4 A20^+/+ and CD4 A20^−/− mice, and analyzed T cell frequencies and numbers 6 d later. The number of total thymocytes and the frequency of CD4 single-positive (CD4SP) T cells were unchanged after 6 d in all groups (Fig. 4A, Supplemental Fig. 2D). However, frequencies and numbers of T_reg cells in the thymus of CD4 A20^+/+ mice were significantly reduced 6 d after injection of anti–IL-2 (Fig. 4). In contrast, CD4 A20^−/− mice showed unchanged T_reg frequencies after neutralization of IL-2 (Fig. 4). We examined the expression of the integral IL-2R component IL-2RA (CD25) in the population of Foxp3⁺ CD4⁺ T_reg cells to test whether differences in IL-2R expression, which could possibly lead to altered sensitivity for IL-2, contributed to this phenotype. We found similar proportions of CD25⁺ and CD25⁻ Foxp3⁺ T_reg cells in CD4 A20^+/+ or CD4 A20^−/− mice and marginally decreased CD25 expression levels of CD25⁺ Foxp3⁺ T_reg cells in CD4 A20^−/− mice (Supplemental Fig. 2E). Overall, these data indicate that the dependence of thymic T_reg cell development on IL-2 is reduced in the absence of A20.

Thymic A20-deficient T_reg cells show unchanged proliferation and apoptosis

We next wanted to analyze the role of A20 for T_reg proliferation and apoptosis in vivo. First, we treated CD4 A20^−/− and CD4 A20^+/+ mice with EdU, a nucleoside analog that is incorporated into DNA during active DNA synthesis and can be analyzed by flow cytometry. We found similar frequencies of EdU⁺ thymic T_reg cells in CD4 A20^−/− and CD4 A20^+/+ mice 24 h after EdU injection (Fig. 5A). We then tested the expression rate of the proliferation marker KI-67 (47) in thymic T_reg cells and again found no significant difference between CD4 A20^−/− and CD4 A20^+/+ mice (Fig. 5B). Next, we determined the frequency of thymic T_reg cells that stain positive for activated caspase-3 to evaluate the role of apoptosis in T_reg cells but found no significant difference between CD4 A20^−/− and CD4 A20^+/+ mice (Fig. 5C). In contrast, when analyzing Foxp3⁻, CD4SP thymic cells, we observed reduced frequencies of EdU⁺, similar frequencies of KI-67⁺, and increased frequencies of activated caspase-3⁺ cells in CD4 A20^−/− versus A20^+/+ mice (Supplemental Fig. 3). Our data indicate that A20 does not regulate T_reg cell proliferation or apoptosis in vivo under the conditions tested in our study and suggest that these mechanisms are not responsible for the enlarged T_reg cell population in CD4 A20^−/− mice.

Thymic A20-deficient T_reg cells show enhanced RelA phosphorylation and reduced nuclear translocation of c-Rel

Because A20 was shown to be a central negative regulator of NF-κB activity in several cell types, we then asked whether deleting A20 in CD4⁺ T cells led to increased NF-κB activity in those cells. In fact, both Foxp3⁺ CD4SP thymic T_reg cells and Foxp3⁻ CD4SP thymic T cells displayed increased NF-κB activity as measured by p65 phosphorylation intensity after stimulation. In Foxp3⁺ CD4SP thymic T_reg cells, p65 phosphorylation intensity was already increased at baseline without stimulation, which was also the case for Foxp3⁻ CD4SP thymic T cells (Fig. 6A, Supplemental Fig. 4A). Because the NF-κB transcription factor c-Rel was shown to play a pronounced role in thymic T_reg cell development (17), we assessed c-Rel translocation to the nucleus of thymic T_reg cells on single-cell levels. Although c-Rel translocation was somewhat decreased in thymic Foxp3⁺ CD4SP A20^−/− T_reg cells compared with A20^+/+ T_reg cells, thymic T_reg cells were characterized by a high degree of nuclear c-Rel translocation in both genotypes even in the absence of stimulation (Fig. 6B). Foxp3⁻ CD4SP thymic A20^−/− T cells also showed a trend for reduced nuclear c-Rel translocation (Supplemental Fig. 4B).

A20 limits development of thymic T_reg cell progenitors

Considering the observations that NF-κB transcription factors and especially c-Rel modulates various steps of T_reg cell development including the generation of thymic T_reg cell progenitors before Foxp3 expression (48), we examined thymic T_reg precursors in the absence of A20. Mice with A20^−/− T cells exhibited increased frequencies of CD25⁺ Foxp3⁻ CD4SP thymocytes, revealing an enhanced T_reg cell precursor compartment (Fig. 7A, 7B). Furthermore, we observed an increased fraction of GITR⁺ cells within this population, resulting in a ∼2-fold increased T_reg cell progenitor (CD4SP CD25⁺ GITR⁺ Foxp3⁻) compartment (Fig. 7C). In addition, GITR expression of the CD25⁺ Foxp3⁻ CD4SP A20^−/− thymic T_reg precursors was enhanced compared with Foxp3-expressing T_reg cells (Fig. 7C, Supplemental Fig. 4C). This population increase at the T_reg precursor differentiation stage in CD4 A20^−/− mice suggests that early intrinsic mechanisms preceding Foxp3 expression are responsible for the developmental advantage of the CD4 A20^−/− versus CD4 A20^+/+ T_reg cell population.

A20-deficient T_reg cells are functional

Having demonstrated that deleting A20 in CD4 T cells induces a specific T_reg cell population increase, we next analyzed whether A20-deficient T_reg cells are functional. In vitro, we found that CD4⁺ CD25⁺ T_reg cells purified from CD4 A20^+/+ or CD4 A20^−/− mice both efficiently suppressed proliferation of A20^+/+ T effector cells to a similar extent (Fig. 8A, 8B). To analyze T_reg cell function in vivo, we used a model of major mismatch allogeneic hematopoietic stem cell transplantation (allo-HSCT), in which the hematopoietic system of recipient mice is eradicated by TBI and then reconstituted with BM and T cells from allogeneic donors. Lethal effects of acute GVHD, mediated by allogeneic effector T cells, can be significantly reduced in this model by cotransplantation of T_reg cells (49). Survival of mice that received either A20^+/+ or A20^−/− T_reg cells together with WT donor BM and T cells was similarly increased in both groups compared with controls that did not receive additional T_reg cells (Fig. 8C). This was attributable to the function of individual T_reg cells and not due to numerical changes throughout the course of the experiment, because the absolute number of T_reg cells in animals cotransplanted with effector and T_reg cells was not significantly different when CD4 A20^+/+ or CD4 A20^−/− T_reg cells were used (Fig. 8D). Together, our data show that loss of A20 does not impair the suppressive functions of T_reg cells.