Dendritic cell SIRT1–HIF1α axis programs the differentiation of CD4⁺ T cells through IL-12 and TGF-β1

Significance

Naive CD4⁺ T cells differentiate into diverse effector and regulatory subsets to orchestrate immunity and tolerance. Whereas the mechanism of T-cell intrinsic signals has been extensively studied, how T-cell lineage differentiation is controlled by innate immune signals remains unknown. Here we used loss-of-function mouse systems, combined with other complementary approaches and models, to define the role of dendritic cell (DC) sirtuin 1 (SIRT1) as a key regulator in orchestrating the orientation of T-cell differentiation via HIF1α signaling in a mammalian target of rapamycin–independent manner. DC-expressed SIRT1, a type III histone deacetylase, programmed reciprocal T helper 1 (T_H1) and regulatory T-cell (T_reg) differentiation by modulating IL-12–STAT4 and TGF-β1–SMAD3 axes and cytokine receptor expressions at the DC–T-cell interface.

Abstract

The differentiation of naive CD4⁺ T cells into distinct lineages plays critical roles in mediating adaptive immunity or maintaining immune tolerance. In addition to being a first line of defense, the innate immune system also actively instructs adaptive immunity through antigen presentation and immunoregulatory cytokine production. Here we found that sirtuin 1 (SIRT1), a type III histone deacetylase, plays an essential role in mediating proinflammatory signaling in dendritic cells (DCs), consequentially modulating the balance of proinflammatory T helper type 1 (T_H1) cells and antiinflammatory Foxp3⁺ regulatory T cells (T_reg cells). Genetic deletion of SIRT1 in DCs restrained the generation of T_reg cells while driving T_H1 development, resulting in an enhanced T-cell–mediated inflammation against microbial responses. Beyond this finding, SIRT1 signaled through a hypoxia-inducible factor-1 alpha (HIF1α)-dependent pathway, orchestrating the reciprocal T_H1 and T_reg lineage commitment through DC-derived IL-12 and TGF-β1. Our studies implicates a DC-based SIRT1–HIF1α metabolic checkpoint in controlling T-cell lineage specification.

CD4⁺ T cells are essential components of the adaptive immune system that regulate immune responses against foreign antigen. Upon antigen recognition, naive CD4⁺ T cells undergo activation and expansion, and, depending on inflammatory contexts and cytokine milieus, differentiate into functional and phenotypic T helper (T_H) subsets characterized by distinct cytokine production profile and function (1–3). T_H1 cells produce IFN-γ and elicit cellular immunity in responding to intracellular pathogens; T_H2 cells produce IL-4 and IL-5 and promote humoral immunity in responding to extracellular bacteria and helminthes; and T_H17 cells produce IL-17 and mediate antifungal defense and inflammation (4, 5). Additionally, regulatory T cells, often known as “induced regulatory T cells” (iT_reg cells), which act in synergy with naturally occurring T_reg cells (nT_reg cells), produce IL-10 and TGF-β1 and dampen immune responses elicited from T_H1, T_H2, and T_H17 (6–9).

Dendritic cells (DCs), an essential component in the innate immune system, play a critical role in initiating front-line primary immune responses and directing subsequent pathogen-specific adaptive immune responses (2). In addition to presenting antigens and modulating cell surface costimulatory molecules, DC-derived cytokines and chemokines can result in either a proinflammatory or antiinflammatory environment, engaging distinct T-cell differentiation programs on naive CD4⁺ T cells (1, 10–16). For example, DC-producing IL-12 can support T_H1 development, whereas DC-producing IL-10 or TGF-β1 can support T_reg development. Recent studies from us and others have shown that innate signaling in DCs mediated by G protein-coupled receptor S1P1, the mitogen-activated kinases (MAPKs), and Wnt–β-catenin plays important roles in stimulating adaptive immune responses through directing native CD4⁺ T-cell differentiation (17–20). However, other critical signaling components in DCs that may play a role in shaping T-cell lineage engagement remain to be identified.

SIRT1 is a mammalian homolog of the yeast NAD⁺-dependent protein deacetylase Sirt2, and plays a role in a variety of essential biological processes, including cell cycle progression, apoptosis, cell survival, gene silencing, heterochromatin formation, tumorigenesis, metabolism, and development (21, 22). SIRT1 has also been implicated in regulating immune responses. In T cells, SIRT1 is required to maintain T-cell tolerance (23, 24) and also play a role in inhibiting the function of T_reg cells in allograft survival (25). In myeloid cells, SIRT1 limits the inflammatory process by inhibiting the expression of proinflammatory cytokines (26, 27), while promoting DC maturation and T_H2 response in airway allergy (28). However, whether SIRT1 is involved in bridging the innate immune signal to adaptive immune response remains unexplored.

Here, we found that SIRT1 plays a critical role in determining the T-cell lineage fate by directing DC-derived cytokine production, which links innate and adaptive immune modulation. Largely through a HIF1α–dependent signaling pathway, SIRT1 is required for the reciprocal production of IL-12 and TGF-β1 production in DCs as well as the expression of IL-12Rβ2 and TGF-βR2 in responding T cells, resulting in a differential lineage engagement of T_H1 and iT_reg in the microbial-induced inflammation.

Results

SIRT1 Is Highly Expressed in DCs and the Deficiency of SIRT1 in CD11c⁺ Cells Does Not Change DC Homeostasis.

To systematically profile SIRT1 expression in the immune system, we purified the mouse immune cells based on surrogate cell surface markers: macrophages (CD11b⁺F4/80⁺ cells), DCs (CD11c⁺MHCII⁺F4/80⁻Ly6G⁻NK1.1⁻CD19⁻TCR⁻ cells), neutrophils (CD11b⁺ Ly6G⁺ cells), natural killer (NK) cells (NK1.1⁺ CD49b⁺ cells), B cells (CD19⁺ cells), and T cells (TCR⁺ cells) (SI Appendix, Fig. S1A). Subsequent qPCR and immunoblot analysis showed that SIRT1 is highly expressed in DC cells (SI Appendix, Fig. S1 B–D). To gain insights into the role of SIRT1 in DCs, we generated a CD11c⁺ cell-specific SIRT1-deficient mouse model by crossing SIRT1^fl/fl with CD11c-Cre (referred to as SIRT1^CD11c−/− hereafter). In addition to DCs, the expression of CD11c in monocytes/macrophages, neutrophils, and some B cells has also been reported (29, 30). To examine the specificity and efficacy of SIRT1 deletion, we isolated various cell populations (indicated in SI Appendix, Fig. S1A) from wild-type (WT) and SIRT1^CD11c−/− mice and examined the expression of SIRT1. We revealed a reduction of SIRT1 protein in DCs but not in other immune cells (SI Appendix, Fig. S1C). Also, we observed a reduction of SIRT1 mRNA expression in CD11c⁺ DC cells but not in other immune cells isolated from SIRT1^CD11c−/− mice compared with their WT counterparts (SI Appendix, Fig. S1D). Notably, the deletion of SIRT1 in DCs is incomplete (about 5-fold reduction in protein level and about 2.5-fold reduction in mRNA level). We reason that both the potential limited efficacy of CD11c-Cre and impurity of DCs (95% in purity, SI Appendix, Fig. S1A) may contribute to the residue level of SIRT1 in our experiment. Nevertheless, these data suggested that the CD11c-Cre activity is largely restrained in DCs with a relatively limited efficacy.

Next, we examined the composition of DC subsets in vivo and revealed a comparable percentage and cell number of CD11c⁺ cells as well as CD8α⁺, CD11b⁺, and mPDCA1⁺ subsets in the spleen isolated from WT and SIRT1^CD11c−/− mice (SI Appendix, Fig. S2A). Also, the expression of costimulatory molecules: CD80, CD86, CD40, and CD54 in CD11c⁺ cells isolated from WT and Sirt1^CD11c−/−mice were comparable (SI Appendix, Fig. S2B). Moreover, acute deletion of SIRT1 in bone marrow cells (Sirt1^CreER model described in detail below) resulted in a similar number of bone marrow-derived DCs (BMDCs) in vitro (SI Appendix, Fig. S3). These data suggested that SIRT1 deficiency does not affect DC development and homeostasis.

SIRT1 Signaling in DCs Is Involved in Regulating T-Cell Differentiation.

As professional antigen-presenting cells (APCs), DCs present foreign antigens with major histocompatibility complexes (MHCs) to T cells and modulate T-cell immune function. We then examined whether SIRT1 regulates antigen presentation in DCs. For the analysis of antigen presentation in DC, we transferred naive CD4⁺ T cells isolated from OT-II TCR-transgenic donors into WT and Sirt1^CD11c−/− recipients, which were subsequently immunized with ovalbumin (OVA) and complete Freund’s adjuvant (CFA). At day 7 after immunization, draining lymph node (DLN) cells were isolated and stimulated with OVA ex vivo for 3 d. CD4⁺ T donor cells isolated from WT and Sirt1^CD11c−/− recipients display a comparable proliferation rate, indicating a similar capability of antigen presentation in both groups (Fig. 1A). However, our parallel assays on T-cell differentiation markers revealed increased IFN-γ⁺ but not IL-17A⁺ or IL-4⁺ donor CD4⁺ T cells isolated from Sirt1^CD11c−/− recipients (Fig. 1B). Also, the amount of secreted IFN-γ but not IL-17 was higher in donor T cells isolated from the immunized Sirt1^CD11c−/− recipients than from WT recipients, indicative of enhanced T_H1 differentiation (Fig. 1C). By contrast, we observed fewer Foxp3⁺ cells and lower level of Foxp3 mRNA expression in T cells from Sirt1^CD11c−/− recipients compared with those in WT recipients, indicative of decreased T_reg differentiation (Fig. 1D). Also, we observed similar alterations on T_reg differentiation in mesenteric lymph nodes (MLNs) from a model of oral antigen-induced iTreg generation, whereby Sirt1^CD11c−/− recipients were fed the antigen in the drinking water following adoptive transfer of OT-II T cells (Fig. 1E). Together, these results indicate that Sirt1 signaling in DCs may restrain inflammatory responses through modulating T-cell differentiation.

Fig. 1.

DC SIRT1 is required for regulating T-cell differentiation. (A–D) Naive OT-II (Thy1.1⁺) T cells were transferred into WT and SIRT1^CD11c−/− mice and immunized with OVA + CFA. DLN cells were analyzed at 7 d after immunization. (A) Proliferation of DLN cells after ex vivo stimulation with OVA for 72 h. (B) Expression of IFN-γ and IL-17A in donor-derived T cells after PMA + ionomycin stimulation. (Right) The proportion of IFN-γ⁺ population among donor cells. (C) Secreted IFN-γ and IL-17A after stimulation with 5 μg/mL OVA for 72 h. (D, Left) The intracellular expression of Foxp3. (Middle) The proportion of Foxp3⁺ population among donor T cells. (Right) The mRNA expression of Foxp3 in donor-derived CD4⁺T cells. (E) Naive OT-II (Thy1.1⁺) T cells were transferred into WT and SIRT1^CD11c−/− mice and then fed with OVA in the drinking water for 5 d. (Left) Foxp3⁺ population in donor-derived CD4⁺ T cells from MLN. (Middle) the proportion of Foxp3⁺ population among donor T cells. Right, the mRNA expression of Foxp3 in donor-derived CD4⁺ T cells. Data are shown as mean ± SD, n = 4, from one of three (A–D) or four (E) independent experiments. **P < 0.01 and ***P < 0.001 compared with the indicated groups.

DC SIRT1 Deletion Enhances Microbial Infection-Induced Inflammation.

Next, we examined the expression of SIRT1 in DCs in responding to various proinflammatory or antiinflammatory stimuli such as LPS, IFN-γ, TNFα, TGF-β1, and IL-10 (7). The proinflammatory and antiinflammatory stimuli readily suppress and promote SIRT1 expression, respectively (Fig. 2A and SI Appendix, Fig. S4).

Fig. 2.

Dendritic cell SIRT1 alters the differentiation of CD4⁺T-cell lineage against microbial-induced inflammation. (A) The protein level of SIRT1 in the DC cells following stimulation with LPS (10 ng/mL), TNFα (100 ng/mL), IL-12 (10 ng/mL), IFN-γ (50 ng/mL), IL-10 (20 ng/mL) and TGF-β1 (5 ng/mL). (B and C) SIRT1^CD11c−/− mice were infected with L. monocytogenes. The intracellular staining of IFN-γ, IL-17A and Foxp3 in the spleen was determined following LLO_189–201 stimulation at day 7. (D–F) T cells (CD4⁺TCR⁺CD45Rb^hiCD25⁻) were transferred into Rag1^−/− and SIRT1^CD11c−/−Rag1^−/− mice, and body weight was measured weekly (D) and representative colon histology (E). The intracellular staining of IFN-γ, IL-17A and Foxp3 in donor cells (CD45.1⁺ cells) isolated from the MLNs (F). Data (B–F) are shown as mean ± SD, from one of four (B–F) independent experiments (at least four mice per group). **P < 0.01 and ***P < 0.001 compared with the indicated groups.

To further explore the role of SIRT1 in relevant in vivo contexts, we examined the pathological progression and T-cell differentiation of WT and SIRT1^CD11c−/− mice in an inflammatory model. We challenged mice with Gram-positive bacteria Listeria monocytogenes, a prototypic model eliciting a strong T_H1 cell-polarized response (31). In this model, we found L. monocytogenes infection resulted in more T_H1 cells, comparable T_H17 cells, but fewer T_reg cells in splenic CD4⁺ T cells isolated from Sirt1^CD11c−/− mice compared with T cells isolated from WT mice, as demonstrated by FACS, qPCR, and ELISA analysis of IFN-γ, IL-17A, and Foxp3 (Fig. 2 B and C and SI Appendix, Fig. S5 A and B).

Furthermore, we tested the role of DC–SIRT1 in a colitis model, the pathogenic progression of which is largely determined by the balance of T_H1 and iT_reg in vivo (32–34). For this test, we transferred naive WT T cells (CD4⁺TCR⁺CD45Rb^hiCD25⁻) into Rag1^−/− or Sirt1^CD11c−/−Rag1^−/− mice. Sirt1^CD11c−/−Rag1^−/− recipients exhibited markedly accelerated weight loss (Fig. 2D) and pathological inflammatory injuries (Fig. 2E) compared with Rag1^−/− recipients. Moreover, we observed more IFN-γ⁺ cells, comparable IL-17⁺ cells, and fewer T_reg cells in Sirt1^CD11c−/−Rag1^−/− recipients than their WT counterparts (Fig. 2F). Altogether, our data from two different in vivo models indicate that the ablation of SIRT1 in DCs results in an altered balance of T_H1 and T_reg cells.

The Role of SIRT1 Signaling in DCs in Directing Antigen-Specific T_H1 and T_reg Differentiation.

To determine how DC–SIRT1 directs antigen-specific T-cell differentiation, we first primed DCs from WT and Sirt1^CD11c−/− with OVA and LPS, and subsequently injected these cells into two groups of recipient mice, which previously received naive OT-II T cells. Then, we analyzed the intracellular expression of T-cell differentiation markers in donor T cells 7 d after DC transfer. Evidently, Sirt1^CD11c−/− DC transfer induced more IFN-γ⁺, comparable IL-17A⁺, and fewer iT_reg donor T cells than those in the WT DC transferred group (Fig. 3A). Next, we extended experiments in the CD11c-Cre model to the acute deletion model. We generated Sirt1^fl/flRosa26-Cre-ER^T2 mice (referred to as Sirt1^CreER mice hereafter) by crossing Sirt1^fl/fl mice with Rosa26-Cre-ER^T2 mice (CreER^T2 fusion gene in the ubiquitously expressed Rosa26 locus). This model allows us to acutely delete targeted gene in a wide range of cell lineages by treating mice in vivo or treating isolated cells in vitro with tamoxifen. Also, the acute deletion of SIRT1 may avoid any potential adaptive and compensatory effects caused by CD11c-Cre–mediated deletion, a chronic process started at early development stage. In addition, the assessment of phenotypes following the deletion of the SIRT1 in CreER system will allow us to determine whether the phenotypes in CD11c-CRE are due to a CRE-specific effect. Both WT and Sirt1^CreER mice were pretreated with tamoxifen for 3 d, which resulted in an efficient deletion of SIRT1 in DCs (SI Appendix, Fig. S6). Then DCs from Sirt1^CreER and WT mice were isolated and treated, and these cells were injected into recipient mice as described above. Acute deletion of SIRT1 in DCs promotes T_H1 but suppresses T_reg differentiation as demonstrated by IFN-γ, IL-17A, and Foxp3 intracellular staining of donor T cells (Fig. 3B). Collectively, these results suggested that SIRT1 deficiency in DCs alters antigen-specific T_H1 and iT_reg cell differentiation in a reciprocal manner in vivo.

Fig. 3.

SIRT1 is required for adoptively transferred DCs to direct T_H1 and iT_reg immunity. Indicated splenic DCs were pulsed with LPS and OVA and injected s.c. into SIRT1^CD11c−/− (A) or Sirt1^CreER (B) recipient mice, which previously received naive OT-II (Thy1.1⁺) T cells 24 h before, as described in Materials and Methods. The intracellular staining of IFN-γ and IL-17A was determined in donor-derived CD4⁺T cells. (Right) The proportion of IFN-γ⁺, IL-17⁺, and Foxp3⁺ population among donor cells. Data are shown as mean ± SD, from one of three independent experiments (at least five mice per group). **P < 0.01 and ***P < 0.001 compared with the indicated groups.

SIRT1 Is Involved in a DC-Dependent Regulation of T_H1 and T_reg Cell Differentiation in Vitro.

Next, we applied an in vitro coculture system (composed of purified naive OT-II T cells, WT, or Sirt1^CD11c−/− splenic DCs and stimuli LPS and OVA) to determine the functional interaction between antigen-specific OTII T cells and DCs. We observed more IFN-γ⁺ T cells, higher expression of IFN-γ mRNA, fewer Foxp3⁺ T cells, and lower expression of Foxp3 mRNA in T cells cocultured with Sirt1^CD11c−/− DCs than in T cells cocultured with WT DCs (Fig. 4 A and B). Also, polyclonal T cells activated by α-CD3 and SIRT1-deficient DCs displayed similar alterations in T_H1 and iT_reg differentiation (Fig. 4 C and D). Notably, the expression of T_H2 and T_H17 cell differentiation markers are comparable in both groups (Fig. 4 B and D). Thus, SIRT1 signaling in DCs directly inhibits T_H1, whereas it instructs iT_reg cell differentiation in vitro.

Fig. 4.

DC–SIRT1 signaling directs T_H1 and T_reg differentiation in vitro. (A and B) Naive OT-II T cells were cocultured with WT or SIRT1^CD11c−/− DCs in the presence of antigen for 5 d. (A) Intracellular staining of IFN-γ and Foxp3 expression in cultured T cells. (Right) The proportions of IFN-γ⁺ and Foxp3⁺ proportion in CD4⁺ T cells. (B) The mRNA expression of indicated genes in T cells restimulated with α-CD3 for 5 h. (C and D) Naive T cells from C57BL/6 mice were cocultured with WT or SIRT1^CD11c−/− DCs in the presence of α-CD3 for 5 d. (C) The proportions of IFN-γ⁺ and Foxp3⁺ proportion in CD4⁺ T cells. (D) The mRNA expression of indicated genes in OT-II T cells. Data are representative of two (A and B) and three (C and D) independent experiments (three to five mice per group), **P < 0.01 and ***P < 0.001 compared with the indicated groups.

SIRT1 Modulates DC-Derived T-Cell Polarizing Cytokines.

We next sought to measure DC-derived cytokines that are known to regulate T_H1 and iT_reg cell differentiation, including IL-12 and TGF-β1. LPS stimulation of SIRT1^CD11c−/− DCs in vitro resulted in an increased intracellular Il-12p40 protein but decreased TGF-β1 protein compared with WT DCs (Fig. 5A). Also, a similar alteration of IL-12 (IL-12p70) and TGF-β1 in mRNA and protein level was observed in DCs after LPS stimulation in vitro (Fig. 5B) and in vivo (Fig. 5C and SI Appendix, Fig. S7). On the other hand, the expression of IL-12p35 and IL-23p19, two known partners of IL-12p40, was not significantly altered in SIRT1^CD11c−/− DCs (SI Appendix, Fig. S8A). Also, the expression of TGF-β2 and TGF-β3 is comparable in SIRT1^CD11c−/− DCs and WT DCs after LPS stimulation in vitro (SI Appendix, Fig. S8B). Moreover, we were unable to find any significant changes of IL-12 and TGF-β1 in other immune cells isolated from Sirt1^CD11c−/− or WT mice following in vitro LPS stimulation (SI Appendix, Fig. S9).

Fig. 5.

SIRT1-dependent IL-12 and TGF-β1 production in DCs regulates T_H1 and T_reg cell differentiation. (A) The intracellular staining of indicated genes in WT and SIRT1^CD11c−/− splenic DCs 5 h following LPS stimulation in vitro. (Right) The proportion of IL-12p40⁺ and intracellular TGF-β1 expression (mean fluorescence intensity, MFI) among CD11c⁺ cells. (B and C) The mRNA expression of IL-12p70 and TGF-β1 in WT and SIRT1^CD11c−/− splenic DCs following LPS stimulation in vitro (B) and in vivo (C) were detected. (D) The intracellular staining of IFN-γ and Foxp3 in T cells cocultured with WT or SIRT1^CD11c−/− splenic DCs with indicated treatment. (Right) The proportion of IFN-γ⁺ and Foxp3⁺ proportion in CD4⁺ T cells. Data are representative of four independent experiments (three to five mice per group). *P < 0.05, **P < 0.01, and ***P < 0.001 compared with the indicated groups.

Next, we applied a DC–T-cell coculture system (as described above) to determine whether SIRT1 signaling in DCs modulates T-cell differentiation through intercellular cytokine signaling. In T cells cocultured with SIRT1^CD11c−/− DCs, we observed an increased STAT4 phosphorylation and decreased SMAD3 phosphorylation compared with T cells cocultured with WT DCs (SI Appendix, Fig. S10). The alteration of these phosphorylation events is consistent with the fact that STAT4 and SMAD3 are downstream signaling targets of IL-12 and TGF-β1, respectively. Moreover, we were able to reverse the impact of SIRT1^CD11c−/− on T_H1 cell differentiation and T_reg differentiation in our coculture system by adding anti–IL-12 neutralizing antibodies and adding TGF-β1, respectively (Fig. 5D). Together, our data suggested that enhanced IL-12 and diminished TGF-β1 production in SIRT1^CD11c−/− DCs contributes to the alteration of T_H1 and T_reg differentiation.

HIF1α Mediates SIRT1-Dependent Regulation of IL-12 and TGF-β1 in DCs.

To gain more mechanistic insights on SIRT1-dependent regulation of DC-derived cytokines, we assessed the predominant signaling molecules in DCs that are involved in LPS-mediated proinflammatory signaling pathways. DCs isolated from SIRT1^CD11c−/− and WT display a comparable level of phosphorylation on ErK, JNK, p38, and AKT, indicative of unaltered signaling. By contrast, SIRT1 deficiency in DCs results in an enhanced S6 phosphorylation (pS6), indicative of the mammalian target of rapamycin (mTOR) signaling, and an accumulated transcriptional factor, HIF1α (Fig. 6A). Conversely, pharmacological stimulation of SIRT1 activity by SRT1720 (an activator of SIRT1) significantly decreases HIF1α expression (Fig. 6A).

Fig. 6.

SIRT1 regulates IL-12 and TGF-β1 production through HIF1α but independent of mTOR. (A) Intracellular staining and immunoblot of indicated proteins in DCs following LPS stimulation in vitro. (B) Intracellular staining of IL-12p40 and TGF-β1 in LPS-stimulated WT and SIRT1^CD11c−/− splenic DCs pretreated with rapamycin (100 μM) for 1 h. (Left) The proportions of IL-12p40⁺ and intracellular TGF-β1 expression (MFI) in DCs. (Right) Intracellular staining of phosphorylation of S6 and HIF1α. (C–E) The intracellular staining of IL-12p40 and TGF-β1 in DCs isolated from indicated mice 4 h following i.p. injection of 10 mg/kg LPS (C). The proportion of IL-12p40 and MFI of TGF-β1 in DCs (D). The serum IL-12 and TGF-β1 levels indicated in mice (E). Data are representative of four (A and E) and two (B–D) independent experiments (three to five mice per group). *P < 0.05, **P < 0.01, and ***P < 0.001 compared with the indicated groups.

To determine whether mTOR signaling is involved in SIRT1-dependent regulation on DC-derived cytokines, we applied a pharmacological approach (rapamycin) to block mTOR activity in DCs. Whereas rapamycin treatment is sufficient to reduce the level of pS6 in SIRT1^CD11c−/− DCs to a comparable level as in WT DCs following LPS stimulation, it failed to reverse the cytokine production and HIF1α level (Fig. 6B). Thus, our results suggested that mTOR is unlikely involved in mediating SIRT1-dependent regulation on cytokines and HIF1α in DCs.

Next, we sought to assess the role of HIF1α in DCs. The treatment of LPS-stimulated DCs with CoCl2, which resulted in an accumulation of HIF1α, readily increases IL-12 and decreases TGF-β1 (SI Appendix, Fig. S11A). Conversely, 2-ME2, an inhibitor of HIF1α, significantly reversed the up-regulated IL-12 expression and down-regulated TGF-β1 expression, which resulted from SIRT1 deficiency in DCs (SI Appendix, Fig. S11B). Furthermore, the genetic deletion of HIF1α in SIRT1^CD11c−/− DCs (DKO) reversed the alterations on IL-12 and TGF-β1 caused by SIRT1 deficiency following LPS stimulation (Fig. 6 C–E). Together, our results suggested that HIF1α signaling mediates SIRT1 effects on IL-12 and TGF-β1 in DCs.

DC SIRT1 Signaling Modulates IL-12Rβ2 and TGF-βR2 Expression in T Cells.

T-cell polarizing cytokines often induce the expression of their corresponding cytokine receptors on T cells, resulting in a robust programming of cell fate determination (1, 16, 20, 35). This led us to investigate the expression of cytokine receptors of IL-12 and TGF-β in T cells activated by WT and SIRT1^CD11c−/− DCs. Expression of IL-12Rβ1, TGF-βR1, and TGF-βR3 protein and mRNA is comparable in T cells induced by WT or SIRT1^CD11c−/− DCs (SI Appendix, Fig. S12). In contrast, SIRT1^CD11c−/− DCs induced more IL-12Rβ2 but less TGF-βR2 in T cells compared with WT DCs. Also, either the genetic deletion of HIF1α (SI Appendix, Fig. S13) and the IL-12 neutralizing antibodies or TGF-β1 treatment could reverse the changes of IL-12Rβ2 or TGF-βR2 in T cells cocultured with SIRT1^CD11c−/− DCs, respectively (SI Appendix, Fig. S14).

Next, we sought to determine whether the alteration of cytokine receptors mediates the change of T_H1 and T_reg differentiation. For this inquiry, we applied shRNA approaches to knockdown IL-12Rβ2 and TGF-βR2 expression in OTII T cells (SI Appendix, Fig. S15), which were subsequently stimulated by either WT or SIRT1^CD11c−/− DCs. Then, we determined mRNA expression of IFN-γ and Foxp3, indicative of T_H1 and iT_reg differentiation. Knockdown of IL-12Rβ2 or TGF-βR2 in OT II T cells selectively reversed the alteration of T_H1 or iT_reg differentiation caused by SIRT1 deficiency in DCs, respectively (SI Appendix, Fig. S16). Collectively, IL-12Rβ2 and TGF-βR2 expression in T cells is required for the alteration of T_H1 and iT_reg cell differentiation induced by SIRT1^CD11c−/− DCs.

Pharmacologically Targeting SIRT1 in Mouse and Human DCs Modulates T-Cell Differentiation.

Next, we sought to apply a pharmacological approach to target SIRT1 in both mouse and human DCs and determine whether we can recapitulate our finding in genetic targeting SIRT1. We first applied Ex527, a selective inhibitor for SIRT1 but not for other histone deacetylases (36, 37) to the mouse DC–T-cell coculture system. In consistent with the results from genetic targeting SIRT1, we observed that SIRT1 inhibitor significantly enhanced T_H1 but suppressed iT_reg differentiation (Fig. 7 A–C). Then, we extended our inhibitor experiment to the human DC–T-cell coculture system, whereby DCs were derived from human peripheral blood monocytes and T cells were isolated from human cord blood. The pharmacological inhibition of SIRT1 in human DCs largely recapitulated what we observed in genetic and pharmacological targeting of mouse DCs in terms of the alterations of human T-cell IFN-γ, Foxp3, IL-12, TGF-β1, IL-12Rβ2, and TGF-βR2 expression (Fig. 7 D–F). Thus, our data indicated that SIRT1 mediated an evolutionarily conserved signaling pathway in both mouse and human DCs.

Fig. 7.

Pharmacologically targeting SIRT1 in mouse and human DCs. As described in Materials and Methods, mouse DCs (A–C) or human DCs (D–F) pulsed with LPS (100 ng/mL) were cocultured with mouse or human T cells, respectively, in the absence or presence of EX-527 (10 μM). The intracellular staining of IFN-γ and Foxp3 in T cells (A or D); the mRNA expression of indicated genes in T cells (B or E); the level of IL-12 and TGF-β1 in indicated supernatant (C or F) were detected. Data are shown as mean ± SD, n = 3–5, from one of two independent experiments. ***P < 0.001 compared with the indicated groups.

Discussion

DCs play a central role in initiating front-line innate immunity and inducing subsequent adaptive immunity in the process of host defense against infection (38, 39). Particularly, DCs shape antigen-specific adaptive immune response through presenting antigens, modulating cell surface costimulatory molecules, and producing cytokines and chemokines (40, 41). Fine tuning a myriad of DC intrinsic signaling pathways is required for eliciting an effective adaptive immune response without triggering inflammation-induced host damage (41, 42). Our current study revealed that an integrated SIRT1–HIF1α signaling axis in DCs directs the generation of two particular subsets of T cells, T_H1 and iT_reg cells, under infectious inflammation. Whereas SIRT1 is not involved in regulating antigen presentation in DCs, SIRT1–HIF1α axis in DCs instructs T_H1 and iT_reg differentiation through modulating the production of DC-derived T-cell polarizing cytokines, including IL-12 and TGF-β1. The altered IL-12Rβ2/TGF-βR2 expression and downstream STAT4/SMAD3 signaling in responding T cells further confer a robust DC–T-cell cross-talk, dictating the programming of T_H1 and iT_reg differentiation (SI Appendix, Fig. S17).

Recent studies suggested that SIRT1 is involved in regulating immune response in various inflammatory models (23–28, 43, 44). Mechanistically, transcriptional factor has been implicated as a critical proinflammatory-signaling module in myeloid leukocytes (45–50). In consistent with recent findings that SIRT1 is responsible for the deacetylation and destabilization of HIF1α (51, 52), we found that HIF1α is not only regulated by SIRT1 but also is involved in mediating the effects of SIRT1 deficiency on DCs.

Metabolic regulation and cell signaling are tightly and ubiquitously linked with immune responses (53–55). The effective immune response requires DCs to function in various conditions, including the alteration of extracellular or intracellular metabolic states due to the migration to a nutrient and/or oxygen-deficient environment (tumor microenvironment and inflammatory sites) or an ongoing metabolic reprogramming (resulted from inflammatory stimulation), respectively. The adaptation of DCs to changing metabolic states resulted from a mechanism of “metabolic checkpoint,” an active signaling process involved in sensing metabolic alteration and subsequently signaling transaction and execution (56). Recent studies indicated that a Toll-like receptor signaling-mediated metabolic reprogramming is required for DC maturation and antigen presentation (57–59). Moreover, the dysregulated mTORC1 or the adenosine monophosphate (AMP)-activated protein kinase (AMPK) activity results in an impaired DC development and maturation, indicating a metabolic checkpoint sensing amino acids and intracellular ATP during DC development and maturation (58, 60). Our current study further implicated a metabolic checkpoint in DCs requiring the interplay of SIRT1 and HIF1α, two metabolic sensors of redox and oxygen states, respectively (61–63). Finally, the above scenarios indicate that the DC-directed adaptive immunity requires the coupling of two evolutionarily conserved stress responses, inflammatory response, and metabolic stress response.

Materials and Methods

Mice and Bacterial Infection Model.

All animal experiments were performed with the approval of the Animal Ethics Committee of Fudan University. C57BL/6 Sirt1^fl/fl, CD11c-Cre, Rosa26-Cre-ER^T2, and HIF1α^fl/fl mice were obtained from The Jackson Laboratory. Thy1.1 and OT-II TCR-transgenic mice were obtained from the Center of Model Animal Research at Nanjing University. Rag1^−/− and CD45.1 mice were obtained from the Beijing University Experimental Animal Center. C57BL/6 mice were obtained from the Fudan University Experimental Animal Center. All mice had been backcrossed to the C57BL/6 background for at least eight generations and were used at an age of 6–12 wk. WT control mice were of the same genetic background and, where relevant, included Cre⁺ mice to account for the effects of Cre (no adverse effects due to Cre expression itself were observed in vitro or in vivo). For bacterial infection, mice were injected i.v. with 4 × 10⁴ L. monocytogenes-expressing OVA. After 7 d, splenocytes were harvested and stimulated with listeriolysin O_189–120 (LLO_189–120) for 5 h for intracelluar staining.

Cell Adoptive Transfer.

Naive T cells 2 × 10⁶ (CD4⁺CD62L^hiCD44^loCD25⁻) from OT-II TCR-transgenic mice were sorted and transferred into Sirt1^fl/fl-CD11c-Cre⁻ (WT) and Sirt1^fl/fl-CD11c-Cre⁺ (Srit1^CD11c−/−) mice. After 24 h, they were injected s.c. with OVA_323–339 in the presence of CFA (Difco) and LPS (Sigma). At day 7–8 after immunization, DLN cells were harvested and stimulated with the cognate peptide for 2–3 d for cytokine mRNA and secretion analyses, or pulsed with phorbol 12-myristate 13-acetate (PMA) and ionomycin for 5 h for intracellular staining from donor-derived T cells. For oral antigen stimulation, after adoptive transfer of 2 × 10⁶ naive OT-II T cells, mice were fed with water supplemented with 20 mg/mL OVA protein (grade VI; Sigma-Aldrich) for 5 d. For T-cell transfer-induced colitis, 3 × 10⁵ T cells (CD4⁺TCR⁺CD45Rb^hiCD25⁻) from WT mice were transferred into Rag1^−/− or SIRT1^CD11c−/−Rag1^−/− mice and body weight was measured weekly. For DC transfer, splenic DCs (2 × 10⁵) were pulsed with 50 μg/mL OVA and 500 ng/mL LPS for 8 h, washed, and injected s.c. into C57BL/6 mice that had received 2 × 10⁶ naive OT-II T cells 24 h before. Mice were killed 7–8 d later for analysis.

Cell Cultures and Flow Cytometry.

Spleens were digested with collagenase D, and DCs (CD11c⁺TCR⁻CD19⁻NK1.1⁻F4/80⁻Ly6G⁻) were sorted on a FACSAria II (Becton Dickinson). Lymphocytes were sorted to enrich for naive T cells. For DC–T-cell cocultures, DCs and T cells (1:10) were mixed in the presence of 1 μg/mL OVA_323–339 peptide and 100 ng/mL LPS. After 5 d of culture, live T cells were stimulated with PMA and ionomycin for intracellular cytokine staining or with plate-bound α-CD3 to measure cytokine secretion and mRNA expression. T-cell proliferation was determined by pulsing of cells with ³H-thymidine for the final 12–16 h of culture. For drug treatments, cells were incubated with vehicle, CoCl2 (200 μM; Calbiochem), 2-ME2 (2 μM; Calbiochem), rapamycin (100 μM; Calbiochem), or EX-527 (10 μM; Sigma) for 0.5–1 h before stimulation. For antibody or cytokine treatment, cultures were supplemented with 5 ng/mL TGF-β1 (R&D Systems), anti–IL-12 mAb (R&D Systems). Flow cytometry was performed with antibodies from eBioscience or BD Biosciences. Anti-mouse TGF-βR3 (ab78421) is from Abcam. Anti-mouse HIF1α mAb (241812), anti-mouse TGF-β1 mAb (9016), anti-mouse TGF-βR1 mAb (141231), anti-mouse TGF-βR2 (21813), anti-mouse IL-12Rβ1(16161), and anti-mouse IL-12Rβ2 (16162) are from R&D Systems. Flow cytometry data were acquired on a FACSCalibur (Becton Dickinson) or an Epics XL bench-top flow cytometer (Beckman Coulter) and data were analyzed with FlowJo (Tree Star). For inducible deletion of loxP-flanked alleles in vivo, mice were treated with 2 mg tamoxifen (Sigma-Aldrich) daily for 3 d.

RNA and Protein Analysis.

RNA was extracted with an RNeasy kit (Qiagen), and cDNA was synthesized using SuperScrip III reverse transcriptase (Invitrogen). An ABI 7900 real-time PCR system was used for quantitative PCR, with primer and probe sets obtained from Applied Biosystems. Results were analyzed using SDS 2.1 software (Applied Biosystems). The cycling threshold value of the endogenous control gene (Hprt1, encoding hypoxanthine guanine phosphoribosyl transferase) was subtracted from the cycling threshold (ΔC_T). The expression of each target gene is presented as the fold change relative to that of control samples (2-^ΔΔCT), as described previously (64). For detection of phosphorylated signaling proteins, purified cells were activated with LPS (Sigma), immediately fixed with Phosflow Perm buffer (BD Biosciences), and stained with phycoerythrin or allophycocyanin directly conjugated to antibody to Erk phosphorylated at Thr202 and Tyr204 (20A; BD Biosciences), p38MAPK phosphorylated at Thr180 and Thr182 (D3F9; Cell Signaling Technology), JNK phosphorylated at Thr183 and Tyr185 (G9; Cell Signaling Technology), STAT4 phosphorylated at Tyr701 and Ser727 (58D6; Cell Signaling Technology), SMAD3 phosphorylated at Tyr705 and Ser727 (D3A7; Cell Signaling Technology), AKT phosphorylated at Ser473 and Thr308 (587F11; Cell Signaling Technology), or S6 phosphorylated at Ser235 and Ser236 (D57.2.2E; Cell Signaling Technology), as described previously (9). Immunoblot analysis was performed as described (64) using anti-HIF1α (H1α67; Sigma-Aldrich), anti-Sirt1 (D60e1; Cell Signaling Technology), and anti–β-actin (AC-15; Sigma-Aldrich) antibodies.

IL-12Rβ2 and TGF-βR2 Knockdown with RNA Interference.

A gene-knockdown lentiviral construct was generated by subcloning gene-specific short hairpin RNA (shRNA) sequences into lentiviral shRNA expression plasmids (pMagic4.1) as described before (65). Lentiviruses were harvested from culture supernatant of 293T cells transfected with shRNA vector. Sorted OT-II CD4⁺ T cells were infected with recombinant lentivirus, and green fluorescent protein expressing cells were isolated using fluorescence sorting 48 h later. The IL-12Rβ2 and TGF-βR2 expression was confirmed using real-time PCR. The sorted T cells with either control or shRNA vectors were used for functional assay.

Human DC and T-Cell Cultures.

For assays of human DC-mediated T-cell activation and differentiation, normal human DCs (CC-2701; Lonza) were cultured and their populations were expanded for 5 d with human granulocyte-macrophage colony-stimulating factor and IL-4 (R&D), followed by treatment with EX-527 and stimulation for 24 h with LPS. DCs were washed extensively and cultured with human cord blood CD4⁺ T cells (2C-200; Lonza) at a ratio of 1:10. After 7 d of culture, live T cells were purified and then stimulated either with PMA and ionomycin for intracellular cytokine staining for 5 h or with plate-bound anti-CD3 for analysis of mRNA expression.

Statistical Analysis.

All data are presented as the mean ± SD. Student’s unpaired t test was applied for comparison of means and to compare differences between groups. Comparison of the survival curves was performed using the log-rank (Mantel–Cox) test. A P value (alpha-value) of less than 0.05 was considered to be statistically significant.