|Home | About | Journals | Submit | Contact Us | Français|
DR3 (TRAMP, LARD, WSL-1, TNFRSF25) is a death-domain containing tumor necrosis receptor primarily expressed in T cells. TL1A, the TNF-family ligand for DR3, can costimulate T cells, but the physiological function of TL1A-DR3 interactions in immune responses is not known. Using DR3-deficient mice, we identified DR3 as the receptor responsible for TL1A-induced T cell costimulation, and dendritic cells as the likely source for TL1A during T cell activation. Despite its role in costimulation, DR3 was not required for in vivo T cell priming, polarization into T helper 1 (Th1), Th2 or Th17 effector cell subtypes, or for effective control of infection with Toxoplasma gondii. Instead, DR3 was required on T cells for immunopathology, local T cell accumulation and cytokine production in Experimental Autoimmune Encephalomyelitis (EAE) and allergic lung inflammation, disease models that depend on distinct effector T cell subsets. DR3 may be an attractive therapeutic target for T cell mediated autoimmune and allergic diseases.
Interactions between TNF family ligands and receptors play an important role in shaping specific features of T cell responses. A subfamily of TNF receptors including CD30, TNFR2, OX40, CD27, GITR, HVEM, and 4-1BB is expressed on T cells. These receptors mediate distinct aspects of costimulation in specific T cell subsets (Croft, 2003; Watts, 2005). Death Receptor 3 (DR3), also known as TNFRSF25, TRAMP, LARD or WSL-1 is a death domain containing TNF-family receptor that, like its closest paralog TNFR1, binds the adaptor molecule TRADD through its cytoplasmic death domain. TRADD recruitment endows DR3 with dual signaling capability to activate NF–κB and MAP-Kinase signaling or alternatively trigger caspase activation and programmed cell death (Chinnaiyan et al., 1996; Screaton et al., 1997; Wen et al., 2003). However, unlike TNFR1, which is widely expressed, DR3 has been reported to be expressed primarily by T lymphocytes (Screaton et al., 1997; Su et al., 2004). The ligand for DR3 was identified in 2002 as the TNF-family member TL1A (Migone et al., 2002). When added to certain tumor cell lines, TL1A can induce apoptosis after addition of cycloheximide. However, in primary T cells TL1A has been reported to enhance proliferation and production of interleukin-2 (IL-2) and interferon-γ (IFN-γ) induced by TCR cross-linking (Migone et al., 2002; Papadakis et al., 2004). TL1A was originally reported to be expressed exclusively in endothelial cells (Migone et al., 2002), but more recently TL1A has been found to be highly expressed in dendritic cells (DCs) after activation in vitro and in Crohn's disease, rheumatoid arthritis, and mouse models of inflammatory bowel disease, (Bamias et al., 2003; Bamias et al., 2006; Cassatella et al., 2007). Genetic variants in TL1A and DR3 have also been associated with Crohn's disease and rheumatoid arthritis, respectively (Osawa et al., 2004; Yamazaki et al., 2005).
Although exogenous TL1A is a T cell costimulator, the role of DR3 in peripheral T cell responses is not known. DR3-deficient mice exhibit a mild defect in thymic negative selection, but peripheral T cell numbers and subsets are normal (Wang et al., 2001). Spleens from DR3-deficient mice also have normal numbers of myeloid DC (CD11b+CD11chiB220−PDCA−), plasmacytoid DC (CD11b−CD11cintB220+PDCA+), macrophages, monocytes and granulocytes (F.M. E.W., and R.M.S, unpublished observations). Understanding the physiological and pathophysiological roles of DR3 in immune responses is important in predicting the therapeutic and possibly deleterious consequences of blocking TL1A-DR3 interactions. Here we have analyzed peripheral T cell responses in mice rendered deficient in DR3 through gene targeting. Exogenous TL1A functioned as a costimulator of T cell proliferation and cytokine production in wild-type, but not DR3-deficient T cells, confirming the role of DR3 as an authentic and non-redundant costimulatory receptor for TL1A. However DR3-deficient T cells only displayed proliferative and cytokine production defects when activated in the presence of DCs, showing that DCs, rather than T cells are the physiologically relevant source of TL1A in T cell costimulation. Despite its role in costimulation, DR3-TL1A interactions were not required for polarization of naive CD4+ T cells into T helper 1 (Th1), Th2 or Th17 effector cell subtypes, and T cell priming and systemic production of effector cytokines were normal in response to a number of model antigens and pathogens. Strikingly however, immunopathology was dramatically reduced in the target organs of two different models of T cell mediated inflammatory disease. These studies reveal a specific, non-redundant function for DR3 in controlling the function of effector CD4+ T cells in the target organs of autoimmune and inflammatory diseases.
Through searching the symatlas gene expression database, we confirmed that DR3 is primarily expressed on T cells in both mouse and man (Figure S1) Exogenous TL1A can costimulate human and mouse T cells, but whether DR3 is the sole costimulatory receptor for TL1A and what role endogenously produced TL1A plays in T cell activation is not known. To investigate this, we purified CD4+ T cells from spleens and lymph nodes of wild-type (WT) or age and sex-matched DR3-deficient (Tnfrsf25−/−) mice (Wang et al., 2001) on a C57BL/6 background and activated them through the TCR in presence or absence of recombinant murine TL1A. Costimulation by other TNF family members has been shown to be maximal when CD28-mediated costimulation is blocked (Croft, 2003). At concentrations of 10 ng/ml or above, TL1A increased T cell proliferation most dramatically in the absence of CD28-mediated costimulation (Figure 1A). When CD28-mediated costimulation was present, TL1A only costimulated proliferation at sub-optimal doses of anti-CD3 (Figure 1A). The increased thymidine incorporation was due to increased cell division and not enhanced survival, as we observed increased CFSE dilution and no significant changes in cellular viability induced by TL1A (data not shown). Importantly, DR3-deficient cells were unresponsive to TL1A, indicating that DR3 is the major receptor that mediates costimulation by TL1A (Figure 1A). However, stimulation of DR3 through endogenous T cell derived TL1A was apparently dispensable for T cell proliferation, because there were no deficits in proliferation in cultures of purified Tnfrsf25−/− T cells (Figure 1A). TL1A costimulation was largely dependent on increased IL-2 production, as TL1A-induced proliferation was greatly reduced in IL-2 deficient T cells or after the addition of an antagonistic IL-2Rα (αCD25) antibody (Figure 1B).
To investigate the spectrum of cytokines that can be costimulated by TL1A and the dependence of cytokine production on DR3, we measured IL-2, IFN-γ and IL-4 production in WT or Tnfrsf25−/− T cells activated in the presence or absence of recombinant TL1A. TL1A increased IL-2, IFN-γ and IL-4 production by WT but not by Tnfrsf25−/− T cells, with IL-4 being most prominently induced by TL1A in the presence of CD28 costimulation (Figure 1C). DR3-deficient T cells were unresponsive to TL1A, but had no defects in cytokine production compared to wild-type T cells. Thus, as with proliferative responses, DR3 is required for TL1A-induced costimulation, but endogenously produced TL1A is not necessary for cytokine production by activated T cells under these conditions.
TL1A has been reported to costimulate memory, but not naive T cells (Bamias et al., 2006). To address this issue, we purified CD62Lhi CD44lo naive CD4+ T cells from WT and DR3-deficient mice and activated them with or without exogenous TL1A. TL1A mildly enhanced proliferation with or without CD28 costimulation and strongly increased IL-2 (8-fold) and IFN-γ (10-fold) production in a DR3-dependent manner (Figure S2A, B), showing that DR3 can function in naive T cells at least by increasing cytokine production. Percentages of memory phenotype CD44hi CD4+T cells were also identical in age-matched Tnfrsf25−/− and control mice (Figure S2C), indicating that TL1A costimulation of total CD4+ T cells is unlikely to be due to differences in the percentages of memory and naive cells.
The lack of proliferative or cytokine production defects in purified DR3-deficient T cells suggested that other cell types may be the physiological source of TL1A. TL1A has been reported to be produced by human DCs and monocytes after a variety of stimuli, and DCs would be a source of TL1A produced at the appropriate time and place for T cell costimulation. To test this hypothesis, we measured upregulation of TL1A gene expression by Reverse Transcriptase Quantitative PCR (RT-qPCR) in purified splenic CD11c+ DCs and bone-marrow derived DCs stimulated with a variety of agents. LPS and Soluble Tachyzoite Antigen from Toxoplasma gondii (STAg) act through Toll-Like receptors (TLRs) that can induce expression of other TNF family members. LPS induced TL1A in bone-marrow derived DC and STAg-induced TL1A in splenic DC that express TLR11 required for STAg responsiveness (Yarovinsky et al., 2005). Expression peaked at up to 100-fold above baseline at 3 hours, and rapidly declining after that (Figure 2A). Interestingly, Schistosoma Egg Antigen (SEA) from Schistosoma mansoni, which triggers alternative activation of DCs to program T cells for Th2 differentiation, did not appreciably induce TL1A mRNA (Figure 2A, left panel). Stimulation of DCs deficient in TLR signaling components showed that LPS induction of TL1A was mediated by TLR4 in a manner dependent on MyD88 and TIRAP (Figure 2B). Immune complexes acting through low-affinity Fc receptors have recently been shown to be a potent stimulus for TL1A production (Cassatella et al., 2007; Prehn et al., 2007). Stimulation of murine DCs with plate-bound cross-linked mouse Ig (IC) also stimulated TL1A gene expression comparably to LPS (Figure 2C). Thus, like other TNF-family members, TL1A can be rapidly induced in DCs through TLR and immune complexes.
To test whether T cells could serve as an autocrine source of TL1A, purified T cells were stimulated through the T cell receptor (TCR) and TL1A mRNA expression was measured by RT-qPCR. TL1A mRNA was upregulated after TCR stimulation, but with delayed kinetics compared with DCs. Interestingly, TL1A upregulation was specifically dependent on DR3-expression, as DR3-deficient T cells showed dramatically reduced TL1A induction but normal upregulation of IL-2 mRNA after activation (Figure 2D). Taken together, these data show that T cells can produce TL1A that acts in an autocrine manner to sustain its own expression, but T cell derived TL1A may not be necessary for proliferation or cytokine production by isolated T cells.
To study the role of TL1A-DR3 interactions in a more physiological model of T cell activation, we backcrossed DR3-deficient mice to the Ovalbumin (Ova)-specific TCR transgenic line OT-II, and cultured naive T cells from Tnfrsf25−/− OT-II and OT-II control mice with Ova peptide and wild-type bone marrow derived DCs. Under these conditions, proliferation of Tnfrsf25−/− OT-II cells was diminished especially at low concentrations of Ova (Figure 3A) whether costimulation through B7 was blocked by adding CTLA4-Ig or not. The cytokine profile of T cells stimulated with Ova peptide and DCs is characteristically dependent on the dose of antigen, with higher doses favoring IFN-γ production and lower doses favoring IL-4 production (Tao et al., 1997). Tnfrsf25−/− OT-II cells produced less IL-2 at lower doses of Ova and lower amounts of IL-4 at all doses of Ova tested. By contrast, Tnfrsf25−/− OT-II cells produced marginally higher amounts of IFN-γ compared with controls at all doses tested (Figure 3B). These data suggest that during cognate interactions between antigen-specific T cells and DCs, TL1A-DR3 interactions function to costimulate T cell proliferation and production of IL-2, IL-4, but not IFN-γ.
These alterations in cytokine production and proliferation suggest that DR3 may influence T cell polarization. To test this, we activated naive CD4+ T cells from Tnfrsf25−/− or control mice in the presence of DCs under conditions optimal for differentiation of Th1, Th2 or Th17 effector T cells or under neutral conditions, and measured cytokine production after restimulation (Figure 4). In the absence of exogenous polarizing stimuli, both WT and Tnfrsf25−/− T cells exhibited mild skewing towards a Th1-IFN-γ secreting profile expected on the C57BL/6 background. In addition, Tnfrsf25−/− T cells could be polarized normally towards IL-4, IFN-γ or IL-17 producing cells (Figure 4A). We then set up cultures of Tnfrsf25−/− OT-II and control OT-II T cells stimulated with DCs and Ova and polarized them with cytokines or STAg, which induces TL1A production. These conditions also resulted in normal Th1 cell skewing by antigen specific Tnfrsf25−/− T cells (Figure 4B). Induction of the transcription factors T-bet, GATA-3, or RORγ by appropriate differentiation stimuli was also unaffected in Tnfrsf25−/− purified T cells (data not shown). Thus TL1A-DR3 interactions appear to be dispensable for the in vitro differentiation of naive T cells into Th1, Th2 or Th17 effector cell subtypes.
To determine the role of DR3 in T cell differentiation and effector function in the intact immune system, we studied disease models dependent on distinct T cell subsets in Tnfrsf25−/− mice. We first studied Experimental Autoimmune Encephalomyelitis (EAE), an autoimmune disease model dependent on Th17 and Th1 cell subsets. In four separate experiments, Tnfrsf25−/− mice exhibited dramatically reduced paralysis measured by clinical scores (Figure 5A). Despite profound resistance to EAE, T cells from draining lymph nodes of MOG-primed Tnfrsf25−/− mice proliferated normally in response to most does of MOG (Figure 5B). In accordance with the clinical scores, the total yield and percentage of CD4+ T cells was markedly reduced in spinal cord homogenates from pooled Tnfrsf25−/− mice compared with controls (Figure 5C). Within the T cell gate, the percentage of IFN-γ producing cells was also reduced by 50% in T cells from the spinal cords of Tnfrsf25−/− mice (Figure 5C). The percentage of IL-17 producing cells was normal in Tnfrsf25−/− mice within this gate, but overall was still reduced due to the decreased percentage of CD4+ T cells in the spinal cord. To examine the absolute amounts of these cytokines in the inflamed spinal cord, mRNA for IL-17 and IFN-γ was measured by RT-qPCR in spinal cord homogenates. Both cytokines were reduced in spinal cord preparations from MOG-primed Tnfrsf25−/− mice when normalized to the housekeeping gene β2-microglobulin, with IFN-γ being the most affected. However, when normalized to the expression of the T cell specific gene CD3δ, IL-17 and IFN-γ mRNA expression were not reduced in Tnfrsf25−/− spinal cord. (Figure 5D). Thus, DR3 is critical for immunopathology in EAE and appears to act principally by controlling the accumulation of T cells in the spinal cord.
To investigate whether DR3 on T cells is responsible for mediating these effects, we performed transfer experiments in which Tnfrsf25−/− or WT CD4+ T cells were transferred into TCRα-deficient mice (Figure 5E). The mice were then immunized with MOG. Mice receiving DR3-deficient T cells exhibited dramatically reduced paralysis compared to recipients WT T cells. As with EAE in Tnfrsf25−/− mice, there were fewer CD4+ T cells in the spinal cord homogenates of mice receiving DR3-deficient T cells induced to develop EAE, with the percentages of CD4+ T cells correlating closely with disease score in each mouse (data not shown). Thus, DR3 is required on T cells for immunopathology in this mouse model of autoimmune demyelinating disease.
We next investigated a Th2 cell dependent model of lung inflammation in which mice are primed systemically with Ova and Alum and then locally challenged with Ova (Gavett et al., 1994). In three independent experiments, histological analysis showed that the airways in Tnfrsf25−/− mice lung had less inflammation, including mucin production and peribronchial inflammation (Figure 6A). Standardized histopathology scores and cell counts in BAL were reduced in OVA-sensitized and challenged Tnfrsf25−/− mice compared with DR3 WT mice sensitized and challenged in parallel with OVA (Figure 6B). The percentage of CD3+ and CD4+ T cells, invariant Vα14+ NK T cells, and eosinophils were all significantly reduced in lung cell preparations from Ova-sensitized and challenged Tnfrsf25−/− mice compared with controls (Figure 6C). Localization of CD3+ cells in lung tissue from Ova-sensitized DR3-deficient mice by immunohistochemistry revealed fewer interstitial and peribronchial T cells compared with controls, and increased perivascular localization. Similar increases in perivascular infiltrates were observed for macrophages (Figure S3). Expression of mRNA for IL-5 and IL-13, which are critical for Th2-mediated lung pathology were markedly reduced in Tnfrsf25−/− Ova-sensitized lungs, whereas IL-10 and IFN-γ were produced equally (Figure 6D). By contrast, when Tnfrsf25−/−spleen cells from these mice were restimulated with Ova there was normal production of IL-5 and IL-13, indicating that systemic priming of OVA-specific Th2 T cells was independent of DR3 (Figure 6E). In addition, Tnfrsf25−/− splenocytes proliferated normally in response to OVA (data not shown). Systemic Th2 cell function as assessed by the production of Ova-specific IgG1 and Ova-specific IgE after Ova priming was also normal in Tnfrsf25−/− mice (Figure 6F). Thus, in this model of Th2-mediated lung inflammation, DR3 is required for Th2 effector cells to accumulate at the site of inflammation but not for systemic differentiation of Th2 T cells. Decreased T cells in the lung may result in defective recruitment of eosinophils and iNKT cells to the site of inflammation as was observed in the Tnfrsf25−/− lung.
To investigate the role DR3 in the trafficking and local proliferation of antigen specific T cells in more detail, and to determine whether the requirement for DR3 is T cell intrinsic in this disease model, we performed transfer experiments in which Tnfrsf25−/− OT-II Ova specific T cells and OT-II controls were activated in vitro under Th2 cell differentiation conditions and then transferred into naive congenic host mice. Recipients were then given a respiratory challenge with Ova and responses analyzed after one or two days. In this system, when antigen was only delivered locally to the lung, we saw a significant defect in accumulation and proliferation of primed Ova-specific T cells in the lung, with an even more profound defect in the recovery and proliferation of Tnfrsf25−/− OT-II T cells in the draining mediastinal lymph nodes (Figure 7A, B). Inflammatory cells in the BAL were correspondingly reduced in recipients of Tnfrsf25−/− OT-II T cells (Figure 7C). These experiments clearly show that DR3 expression on T cells is necessary for proper local responses of primed T cells to cognate antigen. It is possible that this phenotype could stem from a generalized requirement for DR3 for T cells to acquire the ability to migration to extralymphatic tissues. However, when naive DR3-deficient T cells were transferred into congenic hosts and Ovalbumin was delivered systemically with LPS used as an adjuvant, T cells migrated normally to lung, liver and intestinal sites as efficiently as wild-type OT-II T cells (Figure S4). Thus, DR3 on T cells appears to be critical for tissue homing and proliferation only for locally restimulated T cells.
The defective autoimmune and inflammatory responses in Tnfrsf25−/− mice suggest that responses to infectious organisms might also be dependent on DR3. We have investigated this by studying the role of DR3 in immune responses to Toxoplasma gondii, a parasite for which IFN-γ secreting Th1 cells are critical for host defense and survival of infected mice. After infection with T. gondii, Tnfrsf25 −/− mice all survived acute infection. Spleen cells isolated from Tnfrsf25−/− infected mice at seven weeks after infection produced comparable amounts of TNF-α, IFN-γ and, IL-10 in response to STAg compared with controls (Figure S5A). Immune responses to toxoplasma infection were also intact, as indicated by similar counts of toxoplasma cysts in the brains of wild-type and Tnfrsf25−/− animals (Figure S5B). Infiltrating CD3+ T cells were present in similar amounts in histological sections of WT and Tnfrsf25−/− brain (data not shown). In addition Tnfrsf25−/− Toxoplasma infected mice survived similarly to WT infected mice if not better at later stage of infection (Figure S5C). These data indicate that at least in the case of T. gondii, priming and maintenance of effector T cells responsible for controlling acute and chronic infections are not dependent on DR3.
TNF receptors expressed on T cells have been found to play distinct roles in costimulation, post-activation survival, and formation of T cell memory (Croft, 2003; Watts, 2005). Here we have found that TL1A-DR3 interactions can enhance T cell proliferation and cytokine production, but are largely dispensable for differentiation into diverse effector T cell subsets in vitro and in vivo. Our findings provide strong evidence that DR3 is required for effective T cell immune responses in the target organs of T cell mediated autoimmune and inflammatory diseases, whereas DR3 is not required for an effective response to the parasite T. gondii.
Our results suggest that the cellular source of TL1A influences the biological outcome of DR3 engagement on T cells. DCs are the most likely source of TL1A for T cells at the time of antigen presentation, and DR3 is required for costimulation by DC-derived TL1A. TL1A is rapidly upregulated in DCs through TLR or FcR, and then rapidly returns to baseline expression. T cells are also able to upregulate TL1A after stimulation through the TCR, but mRNA upregulation is slower and sustained for at least 48 hours after activation. TL1A produced by T cells can bind to DR3 and enhance TL1A production through an autocrine feedback loop. This contrasts with OX40L and FasL, which are hyper-expressed by T cells in the absence of their cognate receptors (Soroosh et al., 2006) (Hao et al., 2004). TL1A can be cleaved off the plasma membrane by metalloproteases in endothelial and dendritic cells, but not in T cells ((Kim and Zhang, 2005) and F.M., unpublished observations). TL1A produced by T cells in an autocrine manner may sustain costimulation, although we did not find any functional consequences of DR3 deficiency in isolated T cells in terms of proliferation, cytokine production or differentiation. Even in the presence of DCs and TL1A-inducing stimuli, DR3 was not required for T cell polarization of Th1, Th2 or Th17 T cell subsets in vitro or in vivo. However, these conclusions must be tempered by the fact that we have only examined TL1A expression in mouse T cells at the RNA level, and there may be very little TL1A produced under these conditions, and even less by Tnfrsf25−/− T cells.
Despite the relatively modest role of DR3 in T cell activation in vitro, our data with T cell dependent animal disease models establish for the first time an essential role of DR3 in T cells within target tissues during autoimmune and allergic inflammation. Immunopathology and clinical disease were decreased both in a Th2-driven model of Ova-induced lung inflammation and in EAE, which depends on Th1 and Th17 T cells. In both models, the accumulation of T cells in the target tissue and the production of effector cytokines specific for that model were decreased. After submission of this manuscript, two other independent studies using TL1A deficient mice and DR3 dominant-negative mice have implicated TL1A-DR3 interactions in Th17 T cell responses (Pappu et al., 2008) and in allergic lung inflammation (Fang et al., 2008) respectively. Our results show that both disease models require DR3 irrespective of the specific Th subtype involved. Remarkably, unlike local T cell responses, we found that induction of systemic immune responses to Ova and MOG were independent of DR3 as measured by peripheral T cell proliferation, cytokine production or T cell help for antibody responses. In both lung hypersensitivity and EAE, the requirement for DR3 was T cell intrinsic, as indicated by the transfer experiments in which DR3 was deficient only on T cells.
The requirement for DR3 in T cell mediated inflammatory disease may be due to the induction of TL1A in the target organ or draining lymph node during secondary T cell responses in these animal models. TL1A produced in the area of inflammation may enhance survival, migration or proliferation of T cells in target tissues through DR3 signaling. Enhanced survival due to TL1A seems unlikely as we have not observed an increase in apoptotic cells in preparations of lung or spinal cord homogenate from Tnfrsf25−/− mice during lung hypersensitivity or EAE (F.M. and R.M.S., unpublished observations). DR3 may also function to enhance effector cytokine production by T cells at the site of inflammation, although defects in production of effector cytokines on a per-cell basis do not appear to explain the dramatic effects of DR3 deficiency on the severity of EAE. Although we detected fewer cells in the lung and spinal cord of mice with inflammation in these organs, this does not stem from a requirement for DR3 in extralymphatic trafficking of activated T cells, given the normal ability of T cells to home to the lung, liver and gut after systemic priming. Instead, the dramatic reduction in proliferation and expansion of DR3-deficient Ova-specific T cells in the mediastinal lymph nodes and lungs of mice after pulmonary challenge with Ova suggests that a critical function of stimulation of DR3 is to co-stimulate antigen-induced expansion of primed T cells in the target organ of T cell mediated autoimmune and inflammatory diseases. By contrast, DR3 may not be required for the acute Th1 response to T. gondii because this model does not require secondary re-expansion of T cells. DR3 may not be required for control of brain cysts in the chronic phase of T. gondii infection since control of brain cysts has been shown to be more dependent on CD8+ than CD4+ T cells (Parker et al., 1991; Scorza et al., 2003). Thus far, we have not found any defects in primary or secondary CD8+ T cell responses in Tnfrsf25−/− mice. (F.M. and R.M.S., unpublished observations). The role of DR3 in T cell immunity to other infectious agents will need to be tested further to know how applicable our findings are to other infections.
Other TNF family members that costimulate T cells have been reported to contribute to EAE and allergic inflammation, but blockade of most other TNFR family members generally affects systemic immune responses as well as target-organ specific pathology. For example, OX40 has been shown to be important in the generation of a systemic as well as a local Th2 T cell response in asthma (Arestides et al., 2002; Salek-Ardakani et al., 2003). A blocking antibody against CD70 (ligand for the CD27 costimulatory receptor on T cells) ameliorated EAE but also depressed systemic immune responses (Nakajima et al., 2000). Antibody blockade of OX40L was reported to decrease influx of T cells into the brain in EAE (Nohara et al., 2001). Interestingly, like TL1A, OX40L expression is induced in endothelial cells by pro-inflammatory stimuli.
These results make DR3 a relatively unique TNF family receptor in terms of its role in promoting T cell accumulation and pathology in autoimmune and inflammatory disease models depending on a diverse set of cytokines and T helper cell subsets. The role of DR3 in promoting T cell mediated immunopathology strikingly parallels the role of its paralog TNFR1 as an amplifier of innate immune responses. The signaling pathways by which DR3 accomplishes this as well as the molecular mechanisms underlying this specific function of DR3 will be an interesting topic for further study. Because primary T cell responses and host defense are relatively preserved in Tnfrsf25−/− mice, blocking the action of TL1A may be less generally immunosuppressive than blocking other TNF or non-TNF cytokines. This makes blocking TL1A-DR3 interactions a promising strategy for therapy in a wide variety of T cell dependent autoimmune diseases.
LPS from E. Coli was obtained from Sigma. Soluble Tachyzoite Ag (STAg) was prepared from sonicated Toxoplasma gondii tachyzoites, and SEA was prepared from Schistosoma mansoni eggs as previously described (Grunvald et al., 1996). C57BL/6 mice were obtained from Jackson Laboratories. DR3-deficient mice, generated as previously described (Wang et al., 2001), were back-crossed to the C57BL/6 background for at least eight generations. DR3-deficient OT-II mice were generated by crossing DR3-deficient mice to OT-II TCR transgenic mice (Taconic). Il-2−/− mice were a generous gift from Pushpa Pandiyan, NIAID. All antibodies were purchased from BD pharmingen unless indicated otherwise. CD1d/PBS57 tetramers that recognize Vα14 iNKT T cells were prepared by the NIH tetramer core facility. B6.CD45.1 congenic mice were obtained from Jackson Labs or NCI/Frederick. TCRα-deficient mice were obtained from Taconic. Mice were maintained in the NIAMS and NIAID animal facilities under animal study protocols approved by the appropriate NIH institute animal care and use committee.
Splenic dendritic cells were sorted for high expression of CD11c+ on a MoFlo FACS sorter (Dako Carpenteria, CA) from liberase-digested spleens. The purity of CD11c+ DC cells was at least 97%. T cells were purified from spleen and lymph node cell suspensions by magnetic depletion of CD11b, PanNK, B220, NK1.1, CD24, CD16/32, GR-1, I-Ab using FITC conjugated mAb to these antigens (BD pharmingen), and anti-FITC microbeads (Miltenyi). To purify CD4+ T cells, anti-CD8-FITC was added to the above antibodies. For naive T cells, the CD62L+ CD44− population of CD4+ purified cells was sorted after staining with PE-Cy5 anti-CD44 and PE anti-CD62L. Bone marrow dendritic cells were generated by culture with RPMI/10% FCS supplemented with 10 ng/ml of murine GM-CSF (PeproTech, Rocky Hill, NJ). T-depleted APC were obtained by incubating spleen cell suspensions with anti-Thy1.1 for 10 min on ice followed by incubation with low-tox-M rabbit complement (Cedarlane laboratories) for 30 min at 37°C. Cells were washed and incubated with 25 μg/ml of mitomycin C (Sigma) for 30 min at 37°C.
For costimulation studies, CD4+ or naive CD4+ cells were stimulated with plate-bound anti-CD3 mAb (5 μg/ml or at the indicated concentration, 145-2C11; BD Pharmingen) in the presence or absence of plate-bound anti-CD28 mAb (5 μg/ml) (37.51; BD Pharmingen). Recombinant mouse TL1A (R&D systems) was added at 10 ng/ml. For studies with Il2−/− mice, purified T cells were cultured as above, but in the absence or presence of 10 U/ml of IL-2. For DC-T cell co-culture studies, 104 bone-marrow derived DCs were cultured with 105 OT-II or DR3-deficient OT-II naive CD4+T cells per well and the indicated concentration of OVA323–339 peptide, with or without 10 μg/ml of mouse CTLA4/Fc (Chimerigen). On day 3, culture supernatants were collected for cytokine measurement and cells were pulsed with 1μCi of 3H-thymidine. After an additional 16–20 hours, 3H-thymidine incorporation was measured with a scintillation counter. For polarization studies, 8×105 T-depleted APCs were cultured with 2×105 naive CD4+ T cells from C57BL/6 or DR3-deficient mice. Th1 polarization was driven with rIL-12 (20 ng/ml) (PeproTech, Rocky Hill, NJ) and anti-IL-4 (10μg/ml), Th2 polarization with rIL-4 (20 ng/ml) (PeproTech, Rocky Hill, NJ), anti-IL-12 (10μg/ml) and anti IFN-γ (10μg/ml), Th17 polarization with rhTGFβ (5ng/ml) (eBioscience), IL-6 (20ng/ml) (eBioscience), anti-IL-12 (10μg/ml), anti IFN-γ (10μg/ml) and anti-IL-4 (10μg/ml), Th0 with anti-IL-12 (10μg/ml), anti IFN-γ (10μg/ml) and anti-IL-4 (10μg/ml). After 4 days of culture, intracellular cytokine staining was performed as described below. For polarization studies with STAg, 5 × 104 splenic DCs were cultured with 105 OT-II or DR3-deficient OT-II naive CD4+T cells per well with 1μM of OVA323–339 peptide. Th1 polarization was driven with rIL-12 (10 ng/ml), Th2 polarization with rIL-4 (10 ng/ml), and STAg polarization with 5μg/ml of STAg. After 72-h culture, supernatants were replaced with fresh medium containing 10 U/ml of rIL-2, and after an additional 2–3 days, intracellular cytokine staining was performed as described below.
Mice were immunized subcutaneously with myelin oligodendrocyte glycoprotein (MOG) 35–55 peptide in CFA with pertussis toxin administrated IP on days 0 and 2 to induce EAE. Five to eight mice were included per group and were scored. For transfer experiments in EAE, 2×107 CD4+ T cells from C57BL/6 or DR3-deficient mice were transferred into TCRα-deficient mice. Mice were immunized with MOG peptide 24h after the transfer and mice were scored by clinical assessment. Clinical assessment of EAE was performed daily according to the following criteria: (0), no disease; (1), tail paralysis; (2), hind leg weakness; (3), full hind leg paralysis; (4), complete hind limb paralysis plus front limb paraparesis; (5), death. Cells from the spinal cord were isolated using the Neural Tissue Dissociation Kit from Miltenyi Biotec according to the manufacturer's recommended protocol. Lymph node and spleen cells from MOG sensitized animals were isolated using CD4 beads. The cells were restimulated in the presence of irradiated T-depleted splenocytes as APCs and the indicated concentrations of MOG peptide in 96 well plates. On day 3 the cells were pulsed with 3H-thymidine for 6h and then harvested and counted on a scintillation counter.
On days 0 and 7, mice were sensitized systemically via a 200-μl intraperitoneal (i.p.) injection containing either 100 μg Chicken Ova (Sigma) or PBS emulsified in an equal volume mixture with alum (Pierce Laboratories, Rockford, IL). For assessment of pulmonary inflammation, mice were challenged with 100 μg Ova or PBS/30 μl inoculum intratracheally (i.t.) on day 14 and intranasally (i.n.) on day 15. Mice were euthanized 48-72h after the final challenge to evaluate cell infiltration, cellular inflammation in the lung, and cytokine levels in the sera and bronchoalveolar lavage fluid (BALF). For transfer experiments, OT-II or DR3-deficient OT-II CD4 T cells were polarized under Th2 condition for 5 days. The cells were CFSE-labeled and 2×106 cells were transferred into CD45.1 mice. Mice were challenged with 100 μg Ova or PBS/30 μl inoculum i.t. after 24h and i.n. after 48h, or i.t. after 48h. Mice were euthanized 24h after the final challenge to evaluate cell infiltration, cellular inflammation in the lung, and cytokine levels in the sera and bronchoalveolar lavage fluid (BALF). BAL fluid was obtained by direct cannulation of the lungs with a 20-gauge intravenous catheter and lavage with 500 μl 1% fetal bovine serum (FBS) in PBS (for cytokine analysis) and with 750 μl 1% FBS in PBS (for analysis of cellular infiltration). Samples for cytokine analysis were stored at −80°C. Samples for cellular analysis were prepared as a cytospin (Thermo-Shandon, Pittsburgh, PA) for differential cellular analysis after staining with Kwik-diff (Thermo-Shandon), and a portion was used to determine total cell counts. Lung histology was scored by a reader with experimental conditions masked as described previously (McConchie et al., 2006). Inflammatory lung score:(0) normal lung, minor inflammation, no goblet cells; (1) minor perivascular inflammation (PVI)/cuffing (PVC); (2) moderate PVI and peribronchial inflammation/cuffing (PBC); (3) increased PVI and PBI, increased goblet cells smaller airways; (4) severe PVI and PBI, goblet cell infiltration of both small and large airways (practically solid lungs).
T. gondii cysts from the ME-49 strain were prepared from the brains of infected C57BL/6 mice. For experimental infections, mice were inoculated i.p. with an average of 20 cysts/animal. At 7 weeks post-infection the number of cysts in the brain of individual infected animals was determined. Spleen cells were harvested, cultured and stimulated with either anti-CD3 and anti-CD28 or 5μg/ml of STAg. Supernatants were harvested after 72h and analyzed for cytokine production.
5×106 of CD4+ T cells from DR3-deficient OT-II or OT-II control mice were transferred into CD45.1 congenic recipients, and the next day mice were challenged with 100μg OVA protein and 5μg LPS injected i.p. Three days later, lymphocytes were prepared from the indicated organs and analyzed by flow cytometry.
Detection of IFN-γ-, IL-4-, and IL-17- producing cells was determined by intracellular cytokine staining using anti IFN-γ-APC, anti IL-4-PE, anti-IL-17-PE (BD Biosciences). Briefly, cells were stimulated for five hours with anti-CD3 and anti-CD28 or phorbol myristate acetate and ionomycin, with monensin added after two hours. Cells were fixed in 3% paraformaldehyde, permeabilized in 0.1% saponin and analyzed on a FACS Calibur flow cytometer (Becton Dickinson). Cytokine production in cell culture supernatants was analyzed by Cytometric Bead Array (BD Biosciences). Serum immunoglobulins were measured by ELISA following the manufacturer's instructions (Bethyl Labs) and OVA-specific IgG1 and IgE were measured by IgG1 or IgE-specific ELISA using plates coated with 50 μl of OVA (100 μg/ml).
Bone-marrow derived DCs, or splenic CD11c+ DCs from C57BL/6 mice and the indicated knock-out mice were cultured and stimulated for the indicated time with or without 100 ng/ml of LPS, 20 μg/ml of SEA or 10 μg/ml of STAg. Stimulation with Ig cross-linking was performed by coating plates with 0.5mg/ml of mouse IgG (Jackson Immunoresearch) for 1h at 37°C, followed by 50μg/ml of sheep anti-mouse IgG (Jackson Immunoresearch) for 1h at 37°C. Purified T cells were stimulated with 5μg/ml of anti-CD3 and anti-CD28 for the indicated time.
Total RNA was isolated from cells using TriZOL and the pure link™ Micro-to midi kit (Invitrogen). Quantitative RT-PCR was performed using an ABI PRISM 7700 sequence detection system using SuperScript One-Step RT-PCR System (Invitrogen). Pre-designed Primer/probe sets were from Applied Biosystems with the exception of TL1A, which was detected with primers designed to recognize full-length TL1A (forward: CCCCGGAAAAGACTGTATGC; reverse: GGTGAGTA-AACTTGCTGTGGTGAA; probe: TCGGGCCATAACAGAAGAGAGATCTGAGC). Probes specific for β2-microglobulin or CD3δ were used as internal controls. Each measurement of TL1A is normalized to expression of β2-microglobulin or CD3δ (delta Ct) and then measurements of stimulated and unstimulated are compared (delta delta Ct). The inverse log of the delta-delta Ct is then calculated to give the fold change. The level of TL1A in each sample was normalized to the TL1A level in the same cells without treatment.
This research was supported in part by the Intramural Research Programs of the NIAMS an NIAID, National Institutes of Health. Françoise Meylan was supported in part by the Swiss National Science Fundation FNRS Project no: PBGEA—104622. Michelle Kinder was supported by the Molecular and Cell Biology Training Grant at the University of Pennsylvania. We would like to thank John O'Shea for critical reading of the manuscript, Bill Paul, Pam Schwartzberg, and Terri Laufer for helpful advice, Hillary Norris, Brittany Wetzel, Satish Madala, Marcus Hodges, Shweta Trivedi, and Agnieszka Boesen for technical assistance, Patricia Zerfas for assistance with immunohistochemistry, Dr. Allen Cheever for analysis of brain sections, and Joseph Woo and the NIAMS animal facility for assistance with animal care. There are no conflicts of interest.