The transcription cofactor Hopx is required for regulatory T cell function in dendritic cell-mediated peripheral T cell unresponsiveness.

Induced regulatory T cells (iT(reg) cells) can be generated by peripheral dendritic cells (DCs) that mediate T cell unresponsiveness to rechallenge with antigen. The molecular factors required for the function of such iT(reg) cells remain unknown. We report a critical role for the transcription cofactor homeodomain-only protein (Hop; also known as Hopx) in iT(reg) cells to mediate T cell unresponsiveness in vivo. Hopx-sufficient iT(reg) cells downregulated expression of the transcription factor AP-1 complex and suppressed other T cells. In the absence of Hopx, iT(reg) cells had high expression of the AP-1 complex, proliferated and failed to mediate T cell unresponsiveness to rechallenge with antigen. Thus, Hopx is required for the function of T(reg) cells induced by DCs and the promotion of DC-mediated T cell unresponsiveness in vivo.

The monumental task of distinguishing between potentially harmful entities bearing countless antigenic determinants and multiple self-proteins is a complex process referred to as immune tolerance that begins in the thymus and continues in the peripheral lymphoid system 1–6. In the thymus some developing T cells are deleted or become natural regulatory cells (nTreg cells) that modulate immune responses in the periphery 1, 3–6. The active induction of tolerance can also continue extrathymically in the steady state when dendritic cells (DCs) constantly sample their cellular environment to process antigens for subsequent presentation to T cells and induction of immune tolerance 2. DC-induced tolerance to peripheral antigens is thought to be particularly important for limiting deleterious immune responses to antigens unavailable during T cell development in the thymus though it could also limit beneficial anti-cancer immunity 1–4.

Induction of antigen-specific regulatory T cells is an important mechanism of peripheral tolerance induced by DCs under tolerogenic steady state conditions 7. In the mouse, such induced regulatory T (iTreg) cells resemble thymic nTreg cells in many respects and share an expression of some of the Treg cell-specific genes including the transcription factor Foxp3 and the high affinity IL-2 receptor (CD25) 7. While the functions of thymus-derived nTreg cells have been the focus of intense research over the last years, the specific mechanisms of induction and suppression operating in extrathymically iTreg cells have only recently begun to be investigated 7–16. In vivo, the presentation of antigen by steady-state DCs leads to a burst of proliferation among the responding T cells but iTreg cells arise from the few cells that proliferate only minimally during this initial antigenic stimulation 7. Such a low level of the proliferation that leads to induction of iTreg cells by DCs is consistent with the classic anergic phenotype of Treg cells 17–20. Although Treg cells can proliferate and expand in vivo, the anergic phenotype of many Treg cells has been intimately linked with their regulatory function to suppress T cell responses particularly those induced under strongly immunogenic conditions in vivo 17, 21, 22.

Anergy of T cells has been linked to disruption of specific signal transduction and cell cycle-controlling pathways and one of the best-known mechanisms is negative regulation of the function of the AP-1 complex 23–25. The AP-1 complex comprised of Fos and Jun transcription factors regulates expression of multiple genes that control cell proliferation and growth 26. The inhibition of the Fos or Jun transcriptional activity in anergic T cells results in the loss of their proliferation response to antigenic stimulation. In thymic nTreg cells, inhibition of Fos or Jun transcription function by Foxp3 has been linked with their diminished proliferation upon T cell receptor (TCR) engagement but the specific mechanisms of anergy in iTreg cells remain unknown 20, 27–29. In addition to its well-established post-translational regulation, the AP-1 complex has been recently found to be subject to transcriptional control by Homeodomain Only Protein (Hop also known as Hopx) that inhibits AP-1 expression by interfering with initiation of Fos transcription by Serum Response Factor (SRF) 30–33. However, the role of Hopx and its control of AP-1 in immune cells has remained unknown. Here we show that Hopx is critically required for peripheral T cell tolerance mediated by DCs which regulate T cell unresponsiveness during an immunogenic challenge with antigen. In our work, we report that the maintenance of anergic and regulatory phenotypes by DC-induced Treg cell in vivo requires Hopx-dependent down-regulation of expression of Fos and Jun during a response to a secondary challenge with antigen. Thus, the transcription co-factor Hopx is required to mediate T cell unresponsiveness induced by steady-state DCs in vivo by playing a critical role for the function of iTreg cells.

Results

Hopx-expressing Treg cells mediate T cell unresponsiveness

In the steady-state DCs targeted with antigens in vivo induce immune tolerance, characterized by unresponsiveness of T cells to secondary re-challenge with cognate antigen 34, 35. This was established using targeted delivery of antigens such as ovalbumin (OVA) to DCs in vivo using an engineered hybrid antibody to DC-specific CD205 (DEC205) defined here as αDEC-OVA. αDEC-OVA carries peptide antigens as a fusion protein linked to the C terminus of its heavy chains for their subsequent presentation by DCs to antigen-specific T cells under steady state conditions 34. To characterize the molecular mechanisms which promote DC-mediated T cell tolerance, we used the same strategy to deliver antigens to DCs for presentation to T cells under steady-state conditions. We performed gene-array experiments using DC-tolerized antigen-specific total CD4+ T cells in order to identify candidate genes responsible for DC-induced T cell unresponsiveness. Thus, we identified Homeodomain Only Protein (Hop or Hopx), a previously described co-factor involved in transcriptional regulation in non-lymphoid tissues 30, 31 to be induced in T cells tolerized by steady state DCs in vivo (data not shown). We verified the results of the gene-arrays by quantitative real-time PCR and found an approximately 5-fold increase of Hopx expression in OVA-specific CD4 T cells from OTII OVA-TCR transgenic (Tg) mice injected with αDEC-OVA (Fig. 1a). The targeted deletion of Hopx involving loss of all Hopx exons, results in a non-lethal defect in cardiac development but the immune system of the mutant mice develops without gross abnormalities 30, 31. We back-crossed Hopx -deficient (Hopx−/−) mice that were generated on the 129 × B6 mixed background 30 for 11 generations onto the C57BL/6 background and then crossed them with anti-OVA TCR Tg (OTII) animals 36. To examine if expression of Hopx was required for DC-mediated T cell unresponsiveness, we used Hopx -sufficient and Hopx -deficient anti-OVA T cells isolated from OVA TCR Tg Hopx+/− and Hopx−/− OTII littermate mice injected with (αDEC-OVA). We labeled such T cells with 5-(6)-carboxyfluorescein diacetate succinimidyl diester (CFSE) and measured their cell division in response to OVA in Complete Freund’s Adjuvant (CFA) after adoptive transfer. Consistent with previously published results 35, the Hopx-sufficient T cells were unresponsive to further challenge with antigen in vivo and did not proliferate when re-challenged with OVA in CFA (Fig. 1b). In contrast, the transferred Hopx -deficient total CD4+ T cells remained responsive to re-challenge with OVA in CFA (Fig. 1b). We conclude that Hopx is required for DC-mediated T cell unresponsiveness in vivo.

Figure 1

Hopx expressed in iTreg cells is required for T cell unresponsiveness

(a) Expression of Hopx is induced during T cell unresponsiveness.

Hopx transcripts were analyzed in CD4 cells from OTII mice pre-treated either with αDEC-OVA or saline control (PBS). Results in each group are measurements (N=3) from pools of 5 separate animals, normalized for expression of HPRT and standardized to Hopx expression by the dd CT method. Error bars represent SD. P<0.0008

(b) Hopx is required for DC-mediated T cell unresponsiveness.

Hopx−/− and Hopx+/− CD4 cells from αDEC-OVA-pre-treated OTII mice were adoptively transferred as indicated. Recipients were immunized with OVA peptide and T cells were analyzed by flow cytometry 3 days later. Histograms show CFSE intensity and percentages indicate cells that did not proliferate.

The results shown represent one of three similar experiments.

(c) Hopx is expressed in iTreg cells induced by DCs.

Hopx gene expression was measured in CD4 cells from αDEC-OVA pre–treated OTII mice as indicated. The results are measurements (N=3) from pools of 8 separate animals normalized for expression of Hprt and standardized to Hopx expression by the dd CT method. Error bars represent SD. P<0.0002

(d) Treg cells induced by DCs are required for T cell unresponsiveness.

CD4+CD25− T cells from αDEC-OVA-pre-treated OTII mice were transferred either in absence or presence of CD4+CD25+ iTreg cells as indicated. Recipients were immunized with OVA peptide and T cells were analyzed by flow cytometry 3 days later. Histograms show CFSE intensity and percentages indicate cells that did not proliferate.

The results shown represent one of three similar experiments.

Previous research from other investigators showed that targeting DCs with antigens such as anti-DEC-antigen hybrid antibodies leads to the induction of iT regs 7. To examine whether Hopx is expressed in iTreg cells induced by DCs in vivo, we isolated T cells from OVA TCR tg (OTII) mice treated with αDEC-OVA and determined Hopx expression by quantitative PCR in CD4+CD25+ iTreg cells and CD4+CD25− T cells of non-regulatory phenotype (Fig. 1c). Expression of Hopx in anti-OVA iTreg cells was about 20 times higher than in anti-OVA CD4+CD25− T cells isolated from the same αDEC-OVA-treated OTII mice (Fig. 1c). In contrast, the CD25+Foxp3+ nTreg cells that develop in the thymus express Hopx at levels about 3 times lower than in iTreg cells induced in vivo (Supplementary Fig. 1) We conclude that Hopx is preferentially expressed in iTreg cells during the induction of DC-mediated T cell unresponsiveness.

Research from other laboratories 7, 37 has established that CD4+CD25+ T cells induced by steady-state DCs have a regulatory phenotype characterized by expression of Foxp3, however the expression of Hopx by these cells remained unexplored. To analyze the function of such Hopx-expressing Treg cells induced by DCs, we depleted OTII CD4+ T cells isolated from αDEC-OVA-treated OTII mice of CD4+CD25+ double positive iTreg cells, and labeled the remaining (CD25−) cells with CFSE to track their cell division. CD25− T cells were then transferred into CD45.1 recipient mice either alone or after reconstitution with the originally depleted iTreg cells, to attain the original frequency (5%) of CD4+CD25+ T cells. The recipient mice were subsequently challenged with OVA in CFA. In the absence of iTreg cells, the transferred anti-OVA CD25− T cells proliferated to re-challenge with OVA in CFA, but the presence of iTreg cells restored T cell unresponsiveness (Fig. 1b and Fig. 1b). We conclude that iTreg cells are required for DC-induced T cell unresponsiveness in vivo.

Unaltered development of Hopx-deficient Treg cells

We found that both Hopx and the iTreg cells that express Hopx are required for DC-mediated T cell unresponsiveness in vivo (Fig. 1). We therefore, reasoned that Hopx might play a role in either the homeostasis or suppressor function of iTreg cells induced by DCs in vivo. To examine the role of Hopx in iTreg cells, we first compared the numbers of OVA-specific iTreg cells in the Hopx−/− and Hopx+/− mice treated with αDEC-OVA. We injected Hopx−/− and Hopx+/− OTII littermate mice with αDEC-OVA as in Fig. 1 and performed FACS analysis of isolated T cells. We found similar numbers (5.1% versus 5%) of OVA-specific CD4+CD25+ double-positive iTreg cells in αDEC-OVA-treated Hopx−/− and Hopx+/− littermates (Fig. 2a). Thus, Hopx does not determine the number of iTreg cells induced by DCs.

Figure 2

Hopx-deficient iTreg cells express Treg-specific genes but fail to mediate T cell unresponsiveness in vivo

(a) iTreg cells are induced by DC in the absence of Hopx.

Hopx−/− and Hopx+/− CD4 cells from αDEC-OVA-treated mice were analyzed by flow cytometry. The numbers indicate the percentage of double positive CD4+CD25+ cells.

(b) DC-induced Tregs maintain expression of Foxp3 in absence of Hopx.

Hopx−/− and Hopx+/− CD4+CD25+ cells from αDEC-OVA-pre-treated mice were analyzed by flow cytometry. The numbers indicate the percentage of Foxp3 positive cells. The results shown in a and b represent one of three similar experiments.

(c) Comparison of the Foxp3-dependent gene expression in Hopx−/− and Hopx+/− DC-induced Treg cells.

Gene expression was analyzed by quantitative real-time RT-PCR in Hopx−/− and Hopx+/− CD4+RFP+ cells from αDEC-OVA-pre-treated OTII- FOXP3-IRES-RFP mice. Results in each group (Hopx+/− and Hopx−/−) are measurements (N=3) from pools of 6 separate animals normalized for expression of Hprt, and standardized for each gene analyzed to its expression in Hopx+/− Treg cells by the dd CT method. Error bars represent SD. P>0.4.

(d and e) Hopx is required for suppressor function by DC-induced Tregs in vivo.

Hopx+/− (d) or Hopx−/− (e) CD4+CD25− T cells from αDEC-OVA-treated OTII mice were transferred either in absence or presence of Hopx−/− and Hopx+/− DC-induced Treg cells as indicated. Recipients were immunized with OVA peptide and T cells were analyzed by flow cytometry 3 days later. Histograms show CFSE intensity and percentages indicate cells that did not proliferate. The results represent one of three (d) and two (e) similar independent experiments

Foxp3 is a signature transcriptional regulator required for the function of Treg cells of thymic origin (i.e. nTreg cells) which is also expressed in iTreg cells 7, 38, 39. To determine if the expression of Foxp3 in Treg cells induced by DCs was affected by absence of Hopx, we performed Foxp3 intracellular staining in OVA-specific CD4+CD25+ double-positive iTreg cells from αDEC-OVA-treated Hopx−/− and Hopx+/− littermates. Both Hopx−/− and Hopx+/− iTreg cells expressed a similar level of Foxp3 and were present at a similar percentage (87% versus 88%) (Fig 2b).

In Treg cells of thymic origin, Foxp3 was proposed to amplify and maintain the genetic signature that is induced by various developmental and environmental cues and is required for the anergic and suppressive phenotype in Treg cells 20, 27, 28. To examine if Hopx affects the gene expression pattern that is characteristic for Foxp3 in Treg cells, we crossed Hopx−/− and Hopx+/− OTII mice with Foxp3-IRES-red fluorescent protein (RFP) reporter (FIR) mice 40 and determined the expression of several known genes specific for Foxp3+ Treg cells in Hopx−/− and Hopx+/− RFP+ (Foxp3+) iTreg cells from αDEC-OVA-treated OTII-FIR mice. We found similar levels of expression of the Foxp3+ Treg cell-associated molecules Nrp1, Icos, Ctla4 and Gitr in Hopx−/− and Hopx+/− iTreg cells (Fig. 2c). We conclude that absence of Hopx does not alter the numbers of iTreg cells induced by DCs, nor does it obviously disrupt the Foxp3+-specific gene expression pattern in these cells.

Hopx-deficient Treg cells lose suppressor function

To investigate directly the role of Hopx in the suppressor function of iTreg cells induced by DCs in vivo, we compared the suppressor function of Hopx−/− and Hopx+/− iTreg cells from αDEC-OVA-treated Hopx−/− and Hopx+/− OTII littermate mice. We first CFSE-labeled Hopx-sufficient OVA-specific CD25− T cells and reconstituted them with Hopx-sufficient or Hopx-deficient OVA-specific CD25+ iTreg cells to obtain the original frequency (5%) of CD4+CD25+ T cells which is shown in Fig. 2b or with saline control (PBS) only. We adoptively transferred these different combinations of cells into CD45.1 recipient mice that we then immunized with OVA in CFA. To measure the response of CD25− T cells to re-challenge with OVA, we followed their proliferation in vivo by tracking CFSE dilution as in Fig. 1 (Fig. 2d). Similar to the results from Fig. 1, we found that in the presence of Hopx-sufficient iTreg cells, CD25− T cells from αDEC-OVA-treated mice were unresponsive to re-challenge with OVA in CFA in vivo but proliferated in the absence of the iTreg cells (PBS only) (Fig. 2d and compare Fig. 1c). In contrast, CD25− T cells responded to re-challenge with antigen when reconstituted with Hopx-deficient iTreg cells (Fig. 2d). In contrast to CD25− T cells, the vast majority of CD25+ iTreg cells expressed Foxp3 (Fig 2b and Supplementary Fig. 2). In a separate experiment (Supplementary Fig. 3) we used pure Foxp3+ iTreg cells from αDEC-OVA treated Hopx−/− and Hopx+/− OTII-FIR littermate mice and determined that only Hopx- sufficient but not Hopx-deficient Foxp3+ iTreg cells could suppress proliferation to re-challenge in vivo (Supplementary Fig. 3). Our results are also consistent with the lack of DC-induced T cell unresponsiveness observed in transferred total Hopx-deficient CD4+ T cells after re-challenge with OVA in CFA (Fig. 1b). We conclude that expression of Hopx is required for the suppressor function of Treg cells induced by DCs. In contrast, Hopx is not required for the function of nTreg cells because Hopx-deficient CD25+Foxp3+ nTreg cells remain functional (Supplementary Fig. 4).

To exclude a direct function of Hopx in non-regulatory (CD4+CD25−) T cells during DC-mediated T cell unresponsiveness, we examined whether Hopx-deficient CD25− T cells were susceptible to suppression by Hopx-sufficient iTreg cells. We obtained iTreg cells and CD25− T cells as in Fig. 2d and transferred the CFSE-labeled Hopx-deficient CD25− T cells that were reconstituted with either Hopx-sufficient or Hopx-deficient CD25+ iTreg cells or with PBS only. We then measured cell division of Hopx-deficient CD25− T cells in response to immunization with OVA in CFA. Hopx-deficient OVA-specific CD25− T cells in the presence of Hopx-sufficient iTreg cells became unresponsive to further challenge with antigen in vivo and did not proliferate when re-challenged with OVA in CFA (Fig. 2e). By contrast, in the absence of iTreg cells or in the presence of Hopx-deficient iTreg cells, the Hopx-deficient OVA-specific CD25− T cells remained responsive and proliferated to re-challenge with antigen (Fig. 2e). We conclude that Hopx is not required for CD25− non-regulatory T cells to be susceptible to suppression by Treg cells induced by DCs. We further conclude that a Hopx-dependent suppressor function of iTreg cells is both necessary and sufficient for DC-mediated T cell unresponsiveness to antigenic re-challenge in vivo.

Hopx-dependent gene expression in iTreg cells

Hopx is a recently identified, small homeodomain protein, evolutionarily highly conserved between rodents and humans, that has an indirect transcription-regulatory activity 30, 31. Hopx interferes with transcription of some immediate early genes such as Fos and it can influence expression of other diverse genes in various non-lymphoid tissues and organs including heart, skeletal muscles, lungs and nervous system 30–33. We found that the absence of Hopx in the steady-state does not affect the generation of Treg cells by DCs or their expression of Foxp3+-specific genes (Figs. 2a–c). However, during re-challenge with antigen in vivo the suppressor function of such Treg cells is compromised in the absence of Hopx (Fig. 2d and e). To define the role of Hopx in regulating gene expression in iTreg cells during the antigenic re-challenge, we compared total gene expression in the presence or absence of Hopx in iTreg cells in response to immunization. We immunized with OVA in CFA the Hopx−/− and Hopx+/− OTII-FIR littermate mice that we first pre-treated with αDEC-OVA, and then compared global gene expression in isolated Hopx−/− and Hopx+/− anti-OVA CD25+RFP+ (Foxp3+) double-positive iTreg cells. We identified several proliferation-related genes that were altered in the absence of Hopx (Fig. 3). This result is in agreement with the established role of Hopx as a negative transcriptional regulator of the expression of the AP-1 complex and/or Fos 30–33. We conclude that Hopx alters expression of proliferation-promoting genes in iTreg cells induced by DCs upon re-challenge with antigen in vivo.

Figure 3

Hopx-dependent expression of proliferation-related genes in DC-induced Treg cells

Hopx−/− and Hopx+/− OTII FOXP3-IRES-RFP littermate mice were treated with αDEC-OVA and after 7 days re-challenged by immunization with OVA peptide.

TCRα2+CD4+CD25+RFP+ (Foxp3+) DC-induced Treg cells were isolated 24h after immunization with OVA in CFA by cell sorting and total gene expression was measured using Affymetrix Mouse Genome 430 2.0 Arrays. Results show change of a particular gene expression in Hopx-deficient DC-induced Treg cells compared to its expression level in Hopx-sufficient DC-induced Treg cells. 1=expression level in Hopx sufficient. Shown are proliferation-related genes (expression altered in absence of Hopx) and selected Foxp3-dependent and T cell lineages-specific genes. The results shown are from one of two experiments.

In contrast, we found multiple genes that are known targets of other transcription factors such as Foxp3 or which are associated with separate T cell lineages including nTreg, TR1, TH1, TH2 and TH17 cells, to be expressed at similar levels in the presence and absence of Hopx in iTreg cells after the re-challenge with antigen (Fig. 3, Table 1 and Supplementary Figs. 5–7).

Table 1

Hopx-dependent expression of T cell lineage-specific genes in DC-induced Treg cells

Hopx−/− and Hopx+/+ OTII FOXP3-IRES-RFP littermate mice were treated with 50 ng of αDEC-OVA and after 7 days re-challenged by immunization with 100 µg OVA peptide in CFA. TCRα2+CD4+CD25+RFP+ (Foxp3+) DC-induced Treg cells were isolated 24h after immunization with OVA in CFA by cell sorting and total gene expression was measured using Affymetrix Mouse Genome 430 2.0 Arrays as in Fig. 3. Results show fold-change of a particular gene expression in Hopx-deficient DC-induced Treg cells compared to its expression level in Hopx-sufficient DC-induced Treg cells of multiple major Foxp3-dependent and T cell lineages-specific genes. 1=expression level in Hopx sufficient cells.

Gene	FoldChange	Gene	FoldChange	Gene	FoldChange	Gene	FoldChange
Abcb1a	1.15	Ifng	0.81	Il21	1.17	Pdcd1	1.17
Ahr	0.95	Ifngr1	1.22	Il21r	1.03	Pdcd2	0.9
Arl6	0.78	Ifngr2	1.13	Il22ra2	0.85	Phxr4	1.05
Atf1	1.01	Ikzf2	1.29	Il23a	1.08	Prdm1	1.13
Batf	1.03	Il10	1	Il24	1.01	Prkch	1.21
Batf3	0.98	Il10ra	1.12	Il2ra	0.88	Rhoh	1.25
Ccl4	1.14	Il10rb	1.35	Il4	1.04	Rorc	0.77
Ccl5	1.26	Il12rb2	0.9	Il4ra	1.14	Runx1	0.8
Ccr6	1.1	Il13	1.07	Il5	1.02	Runx3	1.01
Cd2	1.15	Il13ra1	1.06	Il5ra	1	S100a10	1.21
Cd200	1	Il17a	1.02	Il6ra	1.2	Snx9	1.04
Cd44	1.14	Il17b	1.16	Il9	1	Socs2	0.9
CD83	1.18	Il17d	0.77	Il9r	0.76	Stat3	1.12
Crem	1.25	Il17ra	1.15	Klrg1	1.17	Stat4	1.22
Ctla4	1.26	Il17rb	1.28	Lag3	0.86	Stat5a	0.9
Cxcr4	0.86	Il17rc	1.2	Maf	0.86	Stat5b	1.14
Ebi3	1.06	Il17rd	1.01	Map3k8	1.05	Stat6	1.01
Foxp3	1.21	Il17re	1.02	Notch1	1	Tbc1d4	1.29
Gata3	1.11	Il1r1	0.81	Notch2	1	Tbx21	0.7
Gitr	0.97	Il1r1	0.81	P4ha1	1.28	Tgfbr1	1.19
Icos	1.26	Il1r2	0.92	P4ha1	1.28	Tgfbr2	1.25

Hopx controls anergy of DC-induced Treg cells

Unresponsiveness to antigen (anergy) is an integral part of the function of some Treg cells but the molecular mechanisms that induce and maintain anergy in such cells remain unclear 17–20, 41. Since our results suggested a role for Hopx in Treg cells induced by DCs in controlling expression of Fos, Jun and cyclin genes all of which are known to be involved in mediating T cell proliferation and anergy 23–25, we examined expression of these genes in such iTreg cells both in the steady-state and after re-challenge with antigen. Expression of Fos and Jun was similar in Hopx-sufficient and Hopx-deficient anti-OVA CD25+RFP+ (Foxp3+) double-positive iTreg cells that were isolated from Hopx−/− and Hopx+/− OTII-FIR littermate mice treated with αDEC-OVA. We then compared gene expression of Fos and Jun in Hopx−/− and Hopx+/− anti-OVA CD25+RFP+ (Foxp3+) double-positive iTreg cells that were isolated from Hopx−/− and Hopx+/− OTII-FIR littermate mice first pre-treated with αDEC-OVA and then immunized with OVA in CFA 12 or 24 hours prior to isolation of the cells (Fig. 4a and b). Consistent with the known down-regulation of AP-1 (Fos and Jun) complexes that occurs during the induction of anergy 23–25, we observed a 3–4 fold decrease in expression of Fos and Jun in Hopx-sufficient iTreg cells 24h after re-challenge with OVA in CFA as compared to the level of expression in the steady state (Fig. 4a and b). In contrast, the expression levels of Fos and Jun in Hopx-deficient iTreg cells remained at high levels after the re-challenge with antigen (Fig. 4a and b). We conclude that Hopx is required to down-regulate expression of proliferation-promoting Fos and Jun in iTreg cells in vivo upon antigenic re-challenge and this result suggests a role for Hopx in regulating anergy of such cells.

Figure 4

Hopx maintains the state of anergy in DC-induced Treg cells

(a–c) Hopx-dependent down-regulation of proliferation-promoting genes in DC-induced Treg cells upon re-challenge with antigen

Hopx−/− and Hopx+/− OTII FOXP3-IRES-RFP littermate mice were treated with αDEC-OVA and 7 days later re-challenged by immunization with OVA peptide as in Fig. 3. DC-induced Treg cells were isolated by cell sorting either from αDEC-OVA-treated but un-immunized mice (NA) or 12 or 24h after immunization with OVA as indicated. Transcripts of Fos (a), Jun (b) and Cyclin D1 (c) were analyzed by real-time RT-PCR. The results are multiple measurements (N=3) from pools of 6–8 separate animals in each group and are normalized for expression of Hprt and standardized by the dd CT method for each gene analyzed to its expression in Hopx-sufficient DC-induced Treg cells from mice unimmunized with OVA in CFA. (**P ≤ 0.01 and ***P ≤ 0.001). Error bars represent SD.

(d) Hopx−/− DC-induced Treg cells proliferate in vivo upon re-challenge with antigen Hopx−/− and Hopx+/− OTII FOXP3-IRES-RFP littermate mice were treated with αDEC-OVA and after 7 days DC-induced Treg cells were isolated by cell sorting, labeled with CFSE and transferred into CD45.1+ C57BL/6 mice. Recipients were immunized with OVA peptide or were un-immunized (NA) and T cells were analyzed by flow cytometry 2 or 3 days later. Histograms show CFSE dilution and percentages indicate cells that did not proliferate. The results shown represent one of two similar experiments.

Cyclin D1 is one of the major Fos targets during initiation of cell proliferation 26. In the steady-state, expression of Cyclin D1 was about 35 % lower in Hopx-deficient iTreg cells induced by DCs (Fig. 4c). We compared cyclin D1 gene expression in Hopx-sufficient and Hopx-deficient iTreg cells in response to re-challenge with OVA in CFA (Fig. 4c). At 12 h following re-challenge, Cyclin D1 expression was decreased by ~80% in Hopx-sufficient iTreg cells but its expression remained at high levels in the absence of Hopx. We conclude that down-regulation of cyclin D1 gene expression occurs concomitant with decreased expression of Fos and Jun in Hopx-sufficient iTreg cells but expression of cyclin D1 remains high in the absence of Hopx.

The Hopx-dependent down-regulation of proliferation-promoting genes after re-challenge with antigen in Treg cells induced by DCs suggests that Hopx is required to maintain anergy of these cells during the acute phase of an antigenic re-challenge response. To examine directly whether Hopx controls proliferation of iTreg cells in response to re-challenge with OVA in CFA, we adoptively transferred CFSE-labeled Hopx-sufficient and Hopx-deficient anti-OVA iTreg cells from αDEC-OVA-treated OTII mice into CD45.1 recipient mice that we then immunized with OVA in CFA; we followed proliferation of these transferred iTreg cells in vivo by tracking dilution of CFSE (Fig. 4d). Consistent with the anergic phenotype of Treg cells, we found that the majority of transferred Hopx-sufficient iTreg cells induced by DCs did not proliferate (63% of CFSE-high [undivided] cells remaining 3d after the re-challenge with antigen) (Fig.4 d). In contrast, Hopx-deficient iTreg cells induced by DCs responded by proliferating to re-challenge with OVA in CFA (39% of CFSE-high [undivided] cells remaining 3d after the re-challenge with antigen) (Fig. 4d). Thus, consistent with its role in down-regulating proliferation-promoting genes (Fig. 4a–c), Hopx is required for the maintenance of the continuous state of anergy in iTreg cells cells upon re-challenge with antigen.

Discussion

DCs were initially discovered to induce immune responses under inflammatory conditions and they have been fittingly called “nature’s adjuvants” 42–44. In the steady-state however, DCs are powerful antigen presenting cells that are constantly sampling their cellular environment to process antigens for subsequent presentation to T cells and induction of immune tolerance 2. Induction of antigen-specific iTreg cells is an important mechanism of peripheral tolerance induced by DCs under tolerogenic steady-state conditions but the specific molecular factors required for the function of DC-induced Treg cells have remained unknown 7–13. We have identified Hopx 30, 31 expression by iTreg cells induced by DCs in the periphery to be indispensable for these lymphocytes to mediate DC-dependent T cell unresponsiveness. Hopx-deficient iTreg cells fail to remain anergic and suppress T cell responses during immunogenic challenge with antigen.

In contrast, the CD25+Foxp3+ nTreg cells that develop in the thymus express Hopx at levels about 3 times lower than in iTreg cells induced in vivo. Moreover, Hopx is not required for the function of nTreg cells because Hopx-deficient CD25+Foxp3+ nTreg cells remain functional Also, Hopx mutant mice appear to have normal tolerance to systemic antigens presented in the thymus 30, 31.

The seminal discovery of Foxp3 as a transcription factor specific for nTreg cells has triggered great interest in these lymphocytes although the precise molecular mechanisms responsible for their anergic-suppressive phenotype still remains incompletely understood 20, 27, 28, 38, 39. The classic anergic phenotype is consistent with the minimal proliferation required for the initial generation of iTreg cells by DCs in vivo 7. Although Treg cells can proliferate and expand in vivo particularly under homeostatic conditions, anergy has been intimately linked with the regulatory function of many Treg cells to suppress acute T cell responses induced under strongly immunogenic conditions 17, 21, 22. The precise molecular link between anergy in nTreg cells and their various proposed mechanisms of direct suppression remains to be fully understood, but it is well established that Foxp3 is required for anergy 20, 27–29, 45. Absence of Foxp3 results in increased proliferation of nTreg cells upon TCR engagement and Foxp3 has also been shown to inhibit proliferation of nTreg cells by binding to Jun and diminishing its transcription function 20, 29.

We now find that DC-induced Treg cells express Hopx which then maintains a state of non-proliferation in those cells upon re-challenge with antigen under immunogenic conditions. Hopx acts in the presence of Foxp3 but the specific function of Hopx appears to be distinct from the transcriptional role of Foxp3 because both the generation of Foxp3+ T cells and expression of Foxp3-dependent genes in DC-induced Treg cells are unaltered in the absence of Hopx. Foxp3 was proposed to amplify and maintain the genetic signature of nTreg cells induced by various developmental and environmental cues, rather than act as a “master switch” that turns on an anergic-suppressor phenotype 20, 27, 28 and it is unknown at present if Hopx’s function could be affected by products of the Foxp3 transcription program.

We also examined the expression of several surface antigens and immunomodulatory cytokines that are known to participate in T cell regulation and suppression 13, 46, 47 and found no difference in their expression in absence of Hopx. In contrast, the previously unappreciated molecular function of Hopx in DC-induced Treg cells is consistent with its established direct role as a negative transcriptional regulator of proliferation-promoting genes such as Fos whereby it prevents ternary complex factors such as SRF 48 from transcribing Fos and other immediate-early genes upon cellular activation 30, 31. This function of Hopx in iTreg cells further provides a molecular explanation for the established mechanism of T cell anergy which is a down-regulation of the function of the AP-1 complex even though the role of Hopx in iTreg cells generated by other approaches in the absence of DC remains unknown at this point 23–25. While the complete explanation of a deficient suppression by non-anergic Treg cells will require considerable further study it is well established that Foxp3 regulates both anergy and suppression in nTregs cells 17, 20, 29. Analogous to this established function of Foxp3 to mediate both anergy and suppression, Hopx is required to maintain suppressive function by anergic iTreg cells during acute immunogenic responses. Following antigenic re-challenge, Hopx-deficient iTreg cells cannot down-regulate expression of Fos, Jun and other proliferation-promoting genes and they both lose their anergic phenotype and fail to inhibit proliferation of antigen-specific T cells. Thus, Hopx maintains the state of T cell unresponsiveness induced by steady-state DCs in vivo.

Methods

Gene arrays

Total gene expression in freshly isolated DC-tolerized and un-activated T cells was measured using Affymetrix Mouse Genome 430 2.0 Arrays. The final processing of RNA, hybridizations and primary data collection were performed at the Keck Affymetrix Facility of Yale University. The data obtained from all experimental groups were analyzed using the generic GCOS Affymetrix software, ArrayAssist software from Stratagene and GeneSpring. For the analysis we used gene expression profiles of the un-activated T cells from PBS injected mice as the baseline and in our initial analysis we focused primarily on genes involved in transcriptional regulation. To identify Hopx-dependent genes, we compared total gene expression in Hopx-deficient and Hopx-sufficient TCRα2+CD4+CD25+RFP+ (Foxp3+) DC-induced Treg cells using the same methodology as above.

Mice

Hopx−/− mice 30 on a mixed 129/B6 background were backcrossed 11 generations with C57BL/6 mice before crossing them with OTII TCR Tg 36 and Foxp3-IRES-RFP (FIR) 40 animals. Experimental animals were produced by breeding the Hopx+/− with Hopx+/− or Hopx−/− mice. Age (7–10 weeks old) and sex-matched Hopx-deficient and sufficient littermates were used for the experiments. All mice were maintained in our facility under specific pathogen-free conditions and used in accordance with committee-approved protocols of the Yale University School of Medicine animal facility

Production of hybrid antibodies

DNA coding for the part of the OVA protein (320–347) containing the OTII peptide (323–339) with spacing residues on both sides was cloned into the C terminus of the heavy chains of a hybrid anti-DEC-205 34 using synthetic oligonucleotides:

Forward:

5’ CTAGCGACATGGCCAAGAAGGAGACAGTCTGGAGGCTCGAGGAGTTCGGT

AGGTTCGAGAGCCTGAAGATATCTCAAGCT

GTCCATG

5’CAGCACATGCAGAAATCAATGAAGCAGGCAGAGAGGTGGTAGGGTCAGCA

GAGGTCATGGAGGAGAGGTAATAGGC

Reverse:

5’GGCCGCCTATTACCTCTCCTCCATGACCTCTGCTGACCCTACCACCTCTCTG

CCTGCTTCATTGATTTCTGCATGTGCTG

CATGGACA

5’GCTTGAGATATCTTCAGGCTCTCGAACCTACCGAACTCCTCGAGCCTCCAG

ACTGTCTCCTTCTTGGCCATGTCG

Hybrid antibodies were expressed in 293T cells by transient transfection using a calcium phosphate method. Cells were grown in serum free DMEM supplemented with Nutridoma SP (Roche). Antibodies were purified on Protein G columns (Roche). The concentrations of antibodies were determined by ELISA using goat anti-mouse IgG1 and mouse IgG1 antibodies as a standard (Jackson Immunotech). Hybrid antibodies were injected in PBS intraperitoneally.

Flow cytometry and antibodies used for staining

Anti-CD4- (L3T4), anti-CD25 (PC61 and 7D4), anti-Valpha2 (B20.1), anti-Vbeta5.1, 5.2 (MR9-4), anti-CD45.2 (104), antu-Il-10 (JES5-16E3) were from BDPharmingen. Anti-Foxp3 (FJK16s), anti-CD73 (TY/11.8), anti-CD39 (24DMS1) were from eBioscience. Flow cytometry and cell sorting (including separation based on RFP expression) was performed on FACSCALIBUR and FACS Vantage (BD). For intracellular staining cells were fixed and permeabilized using Fixation-Permeabilization buffers from eBioscience and BD according to manufacturers’ manual. For cytokine stainings cells were additionally re-stimulated with PMA and Ionomycin.

Adoptive transfer

CD4, CD4+CD25− and CD4+CD25+ cells from OTII transgenic mice were isolated by depletion of CD8+, B200+, CD11c+, CD11b+ and NK1.1+ cells with magnetic microbeads (Miltenyi) and subsequent automated cell sorting FACS Vantage (BD). CD4+ and CD4+ CD25- or CD4+ RFP- cells were labeled with 3 µM CFSE (Molecular Probes) in 5% FCS RPMI at 37°C for 25 min and washed 3× with PBS and 5×106 cells injected intravenously per mouse. In experiments shown in Figs. 1D, 2D and 2E, 5 × 106 CD4+CD25− or CD4+ RFP- cells were reconstituted with 0.3 × 106 CD4+CD25+ or CD4+RFP+ cells before they were injected intravenously per mouse. In experiments shown in Fig. 4D, 0.3 × 106 CFSE-labeled CD4+RFP+ cells were transferred.

Immunizations

Synthetic OVA peptide (323–339) (Yale Keck Protein Synthesis Facility) was injected in Complete Freund’s Adjuvant (Difco) subcutaneously.

Real-time RT- PCR analysis

RNA was isolated from the sorted peripheral CD4 T cells using TRIZOL Reagent (Invitrogen) and Qiagen mRNAEasy kit (Qiagen). Total RNA was reverse transcribed and the cDNA was subsequently used for real-time PCR on an ABI Prism instrument using commercial primer-probe sets (Applied Biosystems): HOPX- Mm01217753_m1, ICOS- Mm00497600_m1, Nrp1- Mm00435379_m1, CTLA-4- Mm00486849_m1, GITR- Mm00437136_m1, Helios-Mm00496108_m1, Foxp3- Mm00475165_m1, Fos- Mm00487425_m1, Jun- Mm00495062_s1, Ccnd1- Mm00432359_m1, IL12a(p35)- Mm00434165_m1, Il-10-Mm00439616_m1, Tgfb1- Mm00441724_m1, Hypoxanthineguanine phosphoribosyltransferase (HPRT) was included as an internal control. The results of Q-PCR were standardized to the HPRT expression levels and analyzed by the dd CT method.

Statistical analysis

P values were calculated using Student’s t-test, error bars in graphs represent SD