ERK/p38 MAP-kinases and PI3K are involved in the differential regulation of B7-H1 expression in DC subsets

Authors


Abstract

Regulatory molecules of the B7-H-family expressed by DC are important for immune homeostasis, but their regulation is largely unknown. When investigating the pathways regulating B7-H1 expression in monocyte-derived DC (MoDC), freshly isolated myeloid DC (mDC) and plasmacytoid DC, respectively, we showed that in MoDC and mDC B7-H1 expression was upregulated by a standard cytokine cocktail, poly I:C or LPS. MoDC utilize ERK and PI3K pathways to upregulate B7-H1 in response to cytokines, whereas p38 kinase was predominantly utilized in response to poly I:C. In mDC, ERK and p38 pathways are involved in B7-H1 regulation, and similar to MoDC, mainly p38 signaling was required for poly I:C-induced expression of B7-H1. Plasmacytoid DC responded only to CpG with upregulation of B7-H1 and in addition to p38 also utilized the PI3K and ERK pathways to regulate B7-H1 expression. As a functional consequence of B7-H1 expression on DC, we demonstrate downmodulation of IFN-γ production in T cells. Thus, the differential regulation of B7-H1 on DC subsets may suppress immune responses variably, depending on the target DC population. Further analysis of the regulatory mechanisms may facilitate the development of new immunosuppressive therapies, utilizing the regulation of B7-H1 expression on DC.

Introduction

Mechanisms that limit undesirable T-cell activation are important for the maintenance of peripheral tolerance. In particular, the expression of positive or negative regulatory molecules on APC is crucial. In humans, the most potent type of APC are DC, which consist of myeloid and plasmacytoid subpopulations as well as Langerhans cells. Although originally defined by their potent T-cell activation capacity, DC are obviously also involved in the maintenance of peripheral tolerance and immune homeostasis, as several types of tolerogenic DC have been described 1, 2.

To provide signals for the induction of T-cell immunity, DC must be matured and activated with microbial agents or cytokines, among others, in order to become equipped with an appropriate set of costimulatory molecules 3. In contrast, immature DC lack cytokine production and surface expression of several T-cell costimulatory molecules and, hence, have been shown to tolerize T cells or to induce regulatory T cells 4, 5.

In addition, DC can also express T-cell regulatory molecules that limit immune responses and protect the body from overwhelming injury. One recently identified molecule, B7-H1 (CD274, PD-L1-programmed death receptor ligand 1), has been implicated in the downregulation of immune responses in different human diseases, such as hepatitis, chronic gastritis, renal inflammatory disease, autoimmunity and different types of cancer 6–12.

B7-H1 is a type 1 transmembrane glycoprotein and a member of the B7 superfamily and one of two ligands for the CD28 homologue programmed death-1 receptor (PD-1). At the mRNA level, it is detected on cells of lymphoid origin as well as on nonlymphoid tissue. However, protein expression is limited to macrophages, activated DC, endothelial and epithelial cells, activated T cells, B cells as well as to various tumor cells 13. Because of its frequent expression by tumors and its suppressive activity against tumor-specific T cells 14–17, B7-H1 was originally attributed to the T-cell inhibitory molecules. Nevertheless, B7-H1 is a regulatory molecule with dual functions on T-cell responses, as it, on one hand, can suppress T-cell activation and, on the other hand, also has T-cell activating properties 18.

Until now not much is known about the molecular mechanisms that regulate B7-H1 expression. Especially the regulation of B7-H1 in DC remains to be elucidated, as these cells govern the decision between tolerance and immunity. Our aim was, therefore, to investigate which signaling mechanisms are involved in the upregulation of B7-H1 in DC and if the application of specific signal transduction inhibitors is able to decrease the induced B7-H1 protein expression.

In this study, we show that B7-H1 expression always accompanied maturation not only of in vitro generated monocyte-derived DC (MoDC) but also of freshly isolated peripheral blood DC. However, further analysis of the subpopulations of blood DC revealed that it is predominantly upregulated on DC of myeloid origin. We also investigated signaling mechanisms leading to protein expression of B7-H1 and demonstrated that ERK/p38-MAPK and PI3K pathways are differentially involved in the regulation of B7-H1 expression in DC subtypes.

Results

Maturation of DC is accompanied by upregulation of B7-H1

We investigated the molecular mechanisms leading to the upregulation of B7-H1 molecules during activation and maturation of different DC subsets in vitro and ex vivo.

First, we compared the expression of B7-H1 molecules on different DC populations before and after stimulation with a standard cytokine cocktail (containing IL-6, TNF-α, PGE2 and IL1-β) 19. Figure 1A and B shows that both in vitro generated MoDC and freshly isolated peripheral blood DC strongly upregulate the surface expression of B7-H1 molecules during maturation with a standard cytokine cocktail (Fig. 1A and B). This upregulation of B7-H1 expression was also obvious during maturation of MoDC and total blood DC by means of the TLR ligands poly I:C and LPS (Fig. 1C and 2A).

Figure 1.

B7-H1 is upregulated during maturation of MoDC and blood DC. (A, B) B7-H1 and CD86 expression on MoDC (A) and total blood DC (B) from healthy donors was measured by flow cytometry before (DC) and after maturation with TNF-α, IL-6, IL-1β and PGE2 containing cytokine cocktail (DC cyt). In vitro-generated MoDC or freshly isolated total blood DC were matured with cytokine cocktail (cyt) and stained with PE-labeled α-B7-H1 and α-CD86 mAB. MLR: Allogeneic CD4+ T-cells were isolated and cocultured with MoDC (A) or total blood DC (B) before (DC) and after their maturation with cytokine cocktail (DC cyt). After 72 h, the proliferative response was assessed by means of 3H-Thymidin incorporation. Data are representative of three experiments and expressed as mean±SD. (C) B7-H1 expression on MoDC measured by flow cytometry after stimulation with cytokine cocktail (cyt), poly I:C, LPS or medium only (control). FACS histogram data represent one typical result out of three independent experiments. (D) The effect of single cytokines on the induction of B7-H1 expression on MoDC. B7-H1 expression was measured by flow cytometry after stimulation with single cytokines (TNF-α, IL-6, IL-1β and PGE2), complete cytokine cocktail (cyt) or medium only (con). Presented in MFI. Data are representative of three experiments and expressed as mean±SD.*p<0.05, **p<0.01.

Figure 2.

Upregulation of B7-H1 expression on total blood DC subpopulations by means of cytokines and TRL ligands. (A) Flow cytometry analysis of B7-H1 expression on freshly isolated total blood DC after incubation for 24 h with cytokine cocktail (cyt), LPS, poly I:C and medium only (control). The different blood DC subpopulations (mDC and pDC) were identified by FITC-labeled CD123 and APC-labeled BDCA-1 mAB. FACS histogram data represent one typical result out of three independent experiments. (B) Enrichment of total blood DC, mDC and pDC from peripheral blood. Respective populations were isolated from the peripheral blood of healthy donors, using MACS beads and stained with APC-BDCA1, APC-BDCA2 and FITC-CD123 mAb. (C) B7-H1 expression on freshly isolated mDC and pDC treated for 24 h with cytokine cocktail (cyt), poly I:C, LPS, CpG and medium only (control) was analyzed by flow cytometry. FACS histogram data represent one typical result out of three independent experiments. (D) B7-H1 expression on freshly isolated mDC and pDC treated with cytokine cocktail (mDC cyt) or CpG (pDC CpG), respectively, and on control cells (mDC, pDC) was analyzed by flow cytometry. Data are presented separately for eight different donors.

To identify how single cytokines, within the standard maturation cocktail, stimulate the induction of B7-H1 expression, we activated MoDC separately with IL-6, TNF-α, PGE2 or IL-1β and assessed B7-H1 expression by FACS analysis. We found that TNF-α and IL-6 were generally the most potent inducers of B7-H1 upregulation (Fig. 1D). Nevertheless, the highest level of B7-H1 expression could only be achieved with the complete four cytokine-containing maturation cocktail, which we therefore used for all further experiments.

In line with previous results, mature DC stimulated allogeneic CD4+ T cells better than their immature counterparts [19], although increased amounts of the “inhibitory” molecule B7-H1 were detected on their surface. These data indicate that B7-H1 might play only a minor role as a T-cell suppressive molecule on activated DC.

B7-H1 expression is differentially induced in DC subpopulations by cytokines and TLR ligands

Although MoDC represent a homogeneous population of myeloid origin, total blood DC comprise a mixed (myeloid DC, mDC and plasmacytoid DC, pDC) population with different repertoires of TLR receptors and cell surface markers 20, 21. Therefore, we investigated whether the B7-H1 expression induced by maturation of blood DC equally affects mDC and pDC, respectively. For this purpose, total blood DC were isolated by paramagnetic beads, stimulated with various TLR ligands, and B7-H1 expression of the individual subpopulations was assessed by FACS. To distinguish the mDC and pDC subpopulations within these cultures, we used the marker BDCA1 and CD123, respectively. We observed that LPS, poly I:C as well as cytokines enhanced expression of B7-H1 in the BDCA1+ mDC population vigorously (Fig. 2A), whereas in CD123+ pDC none of the stimuli was able to induce such a substantial surface expression. To investigate whether LPS, cytokines and poly I:C directly induce B7-H1 expression on mDC or whether a mixture of both DC types is required for effective B7-H1 upregulation, we isolated respective subpopulations using paramagnetic beads (Fig. 2B) and stimulated the subpopulations separately with TLR ligands or the cytokine cocktail. As depicted in Fig. 2B and C, mDC represented approximately 70% of the total blood DC population and could be directly stimulated to express B7-H1 by incubation with the cytokine cocktail, poly I:C and LPS. In contrast, pDC, comprising 25–30% of blood DC, responded only to CpG and to a limited extent to cytokine stimulation with upregulation of B7-H1 (Fig. 2C). We also observed that, despite some individual differences among blood donors, freshly isolated DC demonstrated minimal B7-H1 expression, which was upregulated upon stimulation (Fig. 2D). These results show that the induction of B7-H1 expression during maturation of total blood DC is more pronounced in mDC, whereas pDC show only a moderate, CpG-dependent upregulation of this molecule.

B7-H1 expression in DC is MAPK dependent

To determine the signaling pathways involved in the induction of B7-H1, we applied different signal transduction inhibitors to the DC cultures. First, we set out to analyze the cytokine- and poly I:C-induced B7-H1 expression, as these signals mimic the DC activation and maturation during viral infection. MoDC were preincubated with different signal transduction inhibitors before maturation stimuli were added to the cultures. Pretreatment and coincubation of MoDC with the MEK1/2-specific inhibitor UO126 strongly reduced the cytokine- and poly I:C-induced B7-H1 expression (Fig. 3A), indicating that the MEK/ERK pathway is crucial for B7-H1 surface expression in MoDC. In support of this observation, we further demonstrated that poly I:C as well as the cytokine cocktail induced phosphorylation of ERK42/44 in MoDC (Fig. 3B).

Figure 3.

MEK-ERK pathway is important for B7-H1 upregulation in DC. (A) FACS analysis of B7-H1 expression on MoDC after 24 h of incubation with cytokine cocktail (cyt) or poly I:C with or without 1-hour pretreatment with signal transduction inhibitors, i.e. 20 μM UO126 (MEK1/2), 10 μM SB203580 (p38MAPK) or 25 μM LY294002 (PI3K). FACS histogram data represent one typical result out of three independent experiments. Data in graphs show the mean of the MFI±SD of three experiments. (B) Western blot analysis of ERK42/44 and p38 phosphorylation (pERK42/44 and pp38 for phosphorylated form) in MoDC before and after activation with poly I:C and cytokine cocktail (cyt) for indicated periods of time. The numbers represent the average of gray level intensity recalculated by MetaVue software for p38 and ERK44 bands.

In parallel to the MEK/ERK pathway, cytokine receptors have been shown to signal also through p38 or through PI3 kinase (PI3K), both of which might act independently from the MEK/ERK pathway. Therefore, we also tested whether PI3K and/or p38, respectively, are involved in controlling the expression of B7-H1. Here, we show that the incubation of the MoDC with the PI3K inhibitor LY294002 during maturation with the cytokine cocktail significantly reduced the B7-H1 expression (Fig. 3A). In contrast, the p38-specific inhibitor SB203580 did not affect the cytokine-induced B7-H1 upregulation to such an extent. Both poly I:C and the cytokine cocktail induced phosphorylation of p38 MAPK in MoDC (Fig. 3B). However, as opposed to effects exerted by cytokines, B7-H1 upregulation was stronger abrogated when p38 pathway was blocked during poly I:C treatment of the DC. As demonstrated in Supporting Information Fig. 1, the initial expression of B7-H1 was not significantly downregulated by MEK/ERK, p38 or PI3K inhibitors. In aggregate, our data show that cytokine-induced expression of B7-H1 in MoDC is strongly modulated via the MEK/ERK pathway as well as by the PI3K. B7-H1 upregulation triggered by poly I:C, however, is more crucially dependent on p38.

Activation of ERK and p38 MAPK pathways in mDC is required for upregulation of B7-H1

Considering the fact that in vitro generated MoDC do not always reflect the phenotype and function of their in vivo counterparts and that total blood DC are heterogeneous and consist of at least two main cell subpopulations, i.e. pDC and mDC, we next tested the importance of both ERK and p38 MAPK pathways for B7-H1 upregulation on isolated mDC subpopulations.

We show that the incubation of isolated mDC with the MEK-specific inhibitor UO126 led to the abrogation of cytokine-induced B7-H1 expression (Fig. 4A). The same abrogating effect was observed by coincubating the mDC with the p38-specific inhibitor SB203580. To confirm that ERK and p38 play a decisive role in the regulation of B7-H1 expression in mDC during their maturation with cytokines, we stimulated freshly isolated mDC with the standard cytokine cocktail and assessed the phosphorylation of ERK and p38 by Western blot analysis. As shown in Fig. 4B, phosphorylation of ERK42/44 and p38 was induced upon activation with the cytokine cocktail. In contrast to in vitro generated MoDC, pretreatment of the isolated mDC with the PI3K inhibitor LY294002 had only a minor effect on cytokine-induced B7-H1 expression (Fig. 4A), thus indicating that integral activation of both ERK and p38, but not PI3K, is required for the cytokine-induced B7-H1 upregulation in freshly isolated mDC.

Figure 4.

B7-H1 expression in mDC is ERK/p38 MAPK dependent. (A) FACS analysis of B7-H1 expression on mDC cells activated for 24 h with cytokine cocktail or poly I:C with or without 1-h preincubation with 20 μM UO126 (MEK1/2), 10 μM SB203580 (p38MAPK) or 25 μM LY294002 (PI3K). FACS histogram data represent one typical result out of three independent experiments. Data in graphs show the mean of the MFI±SD of three experiments. (B) Western blot analysis of ERK 42/44 and p38 phosphorylation (pERK42/44, pp38 for phosphorylated form) in mDC before and after activation with cytokine cocktail (cyt) or poly I:C. The numbers represent the average of gray level intensity recalculated by MetaVue software for p38 and ERK44 bands.

Since we previously observed the difference in the regulation of cytokine- versus poly I:C-induced B7-H1 expression in MoDC, we subsequently compared these two stimuli also in mDC isolated from blood. Indeed, the B7-H1 expression induced by poly I:C was accompanied by the activation of ERK42/44 and p38 phosphorylation (Fig. 4B) and was strongly abrogated by the p38 inhibitor SB203580 (Fig. 4A). The MEK/ERK and the PI3K inhibitors UO126 and LY294002, respectively, only partially reduced B7-H1 expression (Fig. 4A). Moreover, the specific inhibitor for NF-κB SN50 also did not reduce cytokine- or poly I:C-induced upregulation of B7-H1 in total blood DC and in mDC (data not shown). Therefore, these data indicate that B7-H1 expression in blood DC induced by poly I:C mainly relies on p38 signaling, whereas MEK/ERK and PI3K play an inferior role.

pDC have a comparable pattern of B7-H1 regulation as mDC

To check whether regulation of B7-H1 expression in pDC requires activation of pathways similar to those activated in mDC, we isolated pDC and treated them with MEK/ERK, PI3K, p38 or NF-κB inhibitors before the activation with CpG and assessed the expression of B7-H1 by flow cytometry.

Here, CpG could activate the surface expression of B7-H1 in pDC and this effect was almost completely abrogated by the p38-specific inhibitor SB203580 and reduced to approximately 50–70% by MEK/ERK (UO126) and PI3K (LY294002) inhibitors, respectively (Fig. 5). However, we did not observe any downregulation of B7-H1 expression by the NF-κB inhibitor SN50. These data show that although pDC differ from mDC in their origin and function, TLR ligands can induce B7-H1 expression in both cell types. Moreover, pDC also use, to various extents, the same signaling pathways for TLR-induced B7-H1 upregulation as mDC, i.e. p38, ERK and PI3K. Additionally, similar to the data obtained for TLR-induced B7-H1 upregulation in mDC, we found that CpG-induced B7-H1 expression by pDC is more substantially affected by the p38 pathway as compared to that by the ERK pathway.

Figure 5.

Regulation of B7-H1 expression in pDC. FACS analysis of B7-H1 expression on pDC cells activated for 24 h with CpG with or without 1-hour pre-incubation with 20 μM UO126 (MEK1/2), 10 μM SB203580 (p38MAPK), 25 μM LY294002 (PI3K) or 20 μM SN50 (NF-κB). FACS histogram data represent one typical result out of three independent experiments. Data in graphs show the mean of the MFI±SD of three experiments.

Blockage of B7-H1 affects the DC-induced cytokine secretion from T cells

To assess the functional consequences of B7-H1 expression on different DC populations in DC–T-cell cocultures, we blocked B7-H1 molecules by means of antibodies and investigated the cytokine production. As shown in Fig. 6A, blockage of B7-H1 on immature MoDC in DC–T-cell cocultures strongly increased the release of IFN-γ and IL-2 (data not shown) by T cells. In contrast, anti-B7-H1 antibodies did not have any effect on cytokine production when added to CD4+ cells in the absence of DC. Moreover, we did not observe a stimulating effect of anti-B7-H1 antibodies in cultures where T cells were stimulated with mature MoDC, as here vigorous cytokine production is readily induced by mature DC. However, when mature MoDC were treated with anti-CD86 antibodies, cytokine secretion was abrogated and subsequent B7-H1 blockage in these cultures restored cytokine production (Fig. 6A).

Figure 6.

B7-H1 controls IFN-γ release in MoDC- and mDC-stimulated CD4+ T-cells. (A) Allogeneic CD4+ cells were isolated and cocultured with immature (DC) and cytokine cocktail matured MoDC (DC cyt) in the presence or the absence of IgG isotype control, anti-B7-H1 mAb, anti-CD86 mAb or medium only (con). The level of IFN-γ in supernatants was determined by ELISA. Data are representative of three experiments and expressed as mean±SD; *, p<0.05, **, p<0.01. (B) Allogeneic CD4+ cells were isolated and cocultured with freshly isolated mDC (immature or matured for 24 h with cytokine cocktail) in the presence or the absence of IgG isotype control, anti-B7-H1 mAb, anti-CD86 mAb or medium only (con). The level of IFN-γ in supernatants was determined by ELISA. Data are representative of three experiments and expressed as mean±SD; *p<0.05, **p<0.01.

To confirm these results for blood DC, we cocultured immature or mature mDC with allogeneic CD4+ T cells with or without anti-B7-H1 and anti-CD86 antibodies, respectively, and assessed the cytokine production in the tissue culture supernatants. Similar to MoDC, we demonstrated that blockage of B7-H1 increased the amount of T-cells-derived IFN-γ in those cocultures (Fig. 6B). In cultures with immature mDC, an effect of B7-H1 blockade on IFN-γ is as well recorded, despite low level of B7-H1 expression on freshly isolated mDC. As we demonstrated in Supporting Information Fig. 2, this is due to the fact that immature DC upregulated B7-H1 after coculture with T cells. Importantly, in contrast to MoDC, we did not need concomitant blocking of CD86 on mature mDC, to observe increased IFN-γ production in mDC–T-cell cocultures.

Discussion

Immune responses are controlled by a network of positive and negative regulatory mechanisms. Within this network, important stimulating and suppressing signals are conveyed by regulatory molecules expressed on the surface of APC. The aim of this study was, therefore, to investigate the signaling events leading to upregulation of B7-H1 surface expression during the maturation of different DC subsets.

It has been shown that B7-H1 acts as a potent immune suppressor. For instance, the expression of B7-H1 on DC or epithelial cells can protect mice from EAE and cure them from lupus-like syndrome 22–26. In addition, in human studies, it has been shown that B7-H1 conveys immunosuppressive signals in autoimmune disorders 27–30 and during chronic inflammation 31. Although these reports point toward an immune suppressive function of B7-H1, it recently became clear that B7-H1 can also play a role as T-cell costimulatory molecule, depending on the cell type expressing B7-H1 and the surrounding milieu of other regulatory molecules 32–34.

Despite its positive effects on preventing autoimmunity and on amelioration of inflammation, B7-H1 expression can be counterproductive in the course of cancer treatment, as different types of tumor cells can obviously escape immune surveillance by expressing B7-H1 14. Moreover, the expression of B7-H1 is often strongly correlated with an unfavorable prognosis in different types of human cancer 10, 35, 36. Thus, B7-H1 seems to play an important role in the modulation of immune response, but the extent to which expression of B7-H1 on APC tune the immune response under steady state and pathological conditions is still not entirely clear.

Here, we show that augmented cell surface expression of B7-H1 parallels maturation of MoDC, which was initiated by a proinflammatory cytokine cocktail or TLR ligands. Together with increased B7-H1 expression, matured MoDC expressed elevated levels of the classical costimulatory molecules such as CD86 and were more potent in the activation of T cells. Moreover, the same results were observed with freshly isolated blood DC. Although freshly isolated blood DC express only low levels of B7-H1, they upregulate the expression of this molecule once matured by cytokines and/or the TLR ligands poly I:C and LPS. That TLR triggering is indeed able to regulate B7-H1 expression in vivo has been suggested by studies showing upregulation of B7-H1 expression in mDC derived from HIV- and hepatitis-B-infected individuals 7, 37.

We further show that mDC can upregulate surface expression of B7-H1 directly in response to cytokines, poly I:C, LPS, whereas pDC only react to CpG with B7-H1 upregulation. The level of B7-H1 expression induced by maturation was significantly stronger in mDC as compared to pDC. This finding is partially supported by in vivo data obtained with cancer and infectious disease models, where mDC have been shown to be the preferential target cell for B7-H1 induction 7, 25, 38–40.

The repertoire of TLR ligands, providing direct stimulation of B7-H1 in mDC (poly I:C stimulates TLR3, LPS stimulates TLR4) and pDC (CpG stimulates TLR9), correlates with the previously published expression of the respective TLR in human DC subsets (summarized in 41), suggesting a direct correlation of TLR activation and B7-H1 upregulation. In aggregate, our data show that in nonpathological steady-state conditions human blood circulating mDC and pDC posses a rather immature phenotype and low B7-H1 expression, while upon engagement of defined TLR, the expression of B7-H1 is upregulated and accompanied by expression of classical T-cell costimulatory molecules. DC-maturing agents, i.e., cytokines and TLR ligands, used in our study were shown previously to induce several defined signaling cascades. For instance, TNF-α, IL-1β and IL-6 activate, among others, NF-κB, PI3K, MAPK and JAK/STAT pathways 42, 43. Likewise, PGE2 can not only activate PI3K but also stimulate ERK in DC through EP2/EP4 receptors 44, 45. Stimulation of TLR induces signaling cascades through MyD88 (myeloid differentiation factor 88) dependent and/or MyD88 independent (Trif-dependent) pathways, leading eventually to the activation of NF-κB, AP-1, and IRF-3 transcription factors. TLR3 and TLR9 are well-known examples for MyD88-independent and MyD88-dependent activation, respectively, while TLR4 can use both pathways 46. The association of TLR with the adaptor molecule MyD88 recruits members of the IRAK (interleukin-1 receptor kinase) family, leading to the activation of the TRAF6 (TNF-receptor-associated factor) – TAK1 axis. TAK1, in turn, activates the canonical IκB kinase complex and consequently NF-κB. Concomitantly, TAK1 also provides downstream signals for MAPK activation. In MyD88-independent pathway, TLR associates with the adaptor molecule Trif (TIR-domain-containing adaptor-inducing interferon β) and activates TBK1 and noncanonical IKK (IκB kinase), resulting in IRF3 activation. Moreover, Trif can also interact with TRAF6 to provide activation of NF-κB and MAPK. PI3K involvement has also been implicated in the TLR signaling. However, its precise role in these pathways remains elusive 47. The well-balanced TLR signaling plays a crucial role in immune response to pathogens. Disregulated activation of TLR results in the severe systemic inflammatory disorders such as septic shock as well as in chronic inflammatory diseases, which might contribute to the development of cancer or autoimmunity. One possible reason for those pathological processes may be the disregulated expression of regulatory molecules such as B7-H1 on DC, as a result of imbalanced TLR signaling. In this study, we demonstrate that PI3K, ERK and p38 pathways differentially regulate B7-H1 on the subsets of DC. In MoDC, the cytokine cocktail led to phosphorylation of ERK. Likewise in mDC, here the cytokine-driven B7-H1 expression was also accompanied by activation of ERK1/2. Accordingly, the inhibition of MEK, which acts upstream of ERK, strongly downregulated the B7-H1 expression induced by cytokines in the respective DC subsets. Corroborative to our results, ERK activation was recently noted to be important for IFN-γ-induced B7-H1 expression in multiple myeloma and dermal fibroblast cells. Thus, our results underline the important role of the MEK/ERK pathway in influencing B7-H1 expression and show, for the first time, this mode of action in human DC, which is similar to previously described mechanisms of ERK-dependent regulation of B7-H1 in myeloma, dermal fibroblasts and bladder cancer cells 48–50.

In contrast to the cytokine-triggered B7-H1 expression in MoDC and mDC, where both depend on activation of ERK, TLR ligands predominantly utilized the p38 MAPK pathway. Likewise in pDC, the CpG-induced B7-H1 expression was strongly dependent on p38 but less dependent on MEK/ERK. Moreover, pDC also utilize the PI3K to regulate B7-H1 expression. To further dissect the contribution of a PI3K-dependent pathway in these series of events, we performed experiments in the presence of a PI3K-specific inhibitor LY294002 and show that the cytokine-induced expression of B7-H1 in MoDC, as well as the CpG-induced expression of B7-H1 in pDC, is downregulated. This is in line with data from Lee et al., demonstrating that cytokines, such as IFN-γ, induced activation of B7-H1 in dermal fibroblasts via PI3K- and ERK-dependent pathways 48. Moreover, the impact of PI3K on B7-H1 regulation has also been noted in vivo in glioma cells, as Parsa et al. showed that B7-H1 expression is positively correlated with the loss of PTEN (Phosphatase and Tensin homologue) expression in tumors. Consequently, the blockade of PI3K activation and subsequent B7-H1 expression induced by PTEN is abrogated by PTEN-loss in tumors, suggesting a novel link between the tumor suppressor PTEN, PI3K and B7-H1 expression 51.

Activation of the aforementioned pathways resulted in dermal fibroblasts in the activation of NF-κB 48. In contrast, in our study, using the NF-κB-specific inhibitor SN50 we could not confirm the involvement of NF-κB activation in the regulation of B7-H1 in MoDC upon maturation with the cytokine cocktail or with TLR ligands. Similarly in mDC, activation with cytokines only required synergistic activation of p38 and ERK pathways, whereas activation of NF-κB was dispensable for induction of B7-H1. In addition, the CpG-induced B7-H1 expression in pDC did not depend on NF-κB. These data are corroborated by findings in a murine system, where the expression of the B7-H1- related PD-L2 molecule was exclusively controlled by NF-κB activation, as opposed to B7-H1 itself, whose expression was not abrogated in NF-κB p50−/−p65+/− DC 52. Thus, our data indicate that B7-H1 regulation in DC occurs independent of NF-κB and it remains to be investigated which transcription factors directly regulate expression of B7-H1 within the different DC subsets. Moreover, the upregulation of B7-H1 seems not to rely on MyD88- or Trif-dependent pathway exclusively.

Many members of the B7 family, including B7-H1, were described as costimulatory molecules for T cells; however, several reports also defined them as immunosuppressive ligands 53. In line with other reports, we show, for the first time, for blood DC that blocking of B7-H1 strongly increased IFN-γ and IL-2 production in T cells 38, 54–56. Thus, these data indicate that the expression of B7-H1 by human DC is involved in controlling the release of IL-2 and IFN-γ by T cells and, therefore, can play a role in the downregulation of ongoing immune responses 57. This is in agreement with recently published data showing that anti-B7-H1 antibodies enhance T-cell proliferation in MLR 58, 59. IFN-γ was shown to induce B7-H1 expression in DC per se7, 60. Therefore, this B7-H1–IFN-γ interplay could provide a regulatory feedback loop, where the production of IFN-γ from T cells stimulated by DC during immune responses increases the expression of B7-H1 on the surface of DC, which in turn downregulates the IFN-γ release by T cells. Thus, immune responses would be dampened.

In aggregate, this study demonstrates that surface expression of B7-H1 is directly induced in different DC subsets during maturation by a classical cytokine cocktail and TLR3, TLR4 and TLR9 signaling. Signal transduction events leading to B7-H1 upregulation involved PI3K, ERK and p38 pathways in the different DC subsets. As a functional consequence, we show that B7-H1 expression by DC is involved in the maintenance of low levels of IFN-γ secretion by T cells and, thus, may help to downregulate immune responses.

Our findings in combination with other published reports on the involvement of B7-H1 expression in different disease models may lead to the development of new therapeutic strategies using specific signal transduction inhibitors or agonists as a part of combinatory therapy in order to precisely modulate B7-H1 expression in DC during cancer and disease.

Materials and methods

Antibodies and reagents

Following antibodies and reagents were used in this study.

Antibodies: anti-CD86-PE (Beckman Coulter, distributor Greiner Bio-one GmbH Frickenhausen, Germany); anti-BDCA1-APC, anti-BDCA2-APC, anti-BDCA3-APC (all Miltenyi, Bergisch-Gladbach, Germany); anti-B7-H1-PE, purified anti-CD86, anti-B7-H1, mouse IgG1 (all NatuTec, Frankfurt, Germany); anti-phospho-ERK42/44 (Thr202/Tyr204), anti-phospho-p38 (Thr180/Tyr182), anti-ERK, anti-p38 (all Cell Signaling Technology). Cytokines: IL-4, GM-CSF, TNF-α, IL-6, IL-1β, PGE2 (Promokine, Germany; Immunotools, Germany; Strathmann, Germany). TLR ligands: Poly I:C (Amersham, Bioscience, USA), LPS (Sigma), CpG ODN (Eurofins, MWG, Germany). Signal transduction inhibitors: UO126 (MEK1/2 inhibitor, Cell Signalling Technology), SB203580 (p38 MAPK inhibitor, Cell Signalling Technology), SN50 (NF-κB inhibitor; Calbiochem, Germany), LY294002 (PI3K inhibitor; Calbiochem, Germany).

Generation of human DC

Human MoDC were generated as previously described 61. Briefly, PBMC were prepared from whole blood of healthy donors or from Buffy Coats by Biocoll gradient centrifugation (Biochrom AG, Germany). Plastic adherent monocytes obtained from freshly isolated PBMC were cultured in a medium in the presence of GM-CSF and IL-4 cytokines for 5–6 days in X-VIVO 15 medium (Bio-Whittaker, Vervies, Belgium) supplemented with 1.5% human plasma and 1 ml of the medium was substituted with a fresh medium on day 3. The resulting DC population exhibited morphologic and phenotypic features consistent with DC, as determined by FACS. Peripheral blood DC, mDC and pDC were isolated from freshly prepared PBMC using MACS isolation kits: Blood Dendritic cell II isolation kit, CD1c mDC isolation kit and pDC isolation kit (all Miltenyi) according to the manufacturer's protocol.

FACS analysis

For FACS analysis the cells were collected, washed and stained for 15 min at 4C° in FACS-buffer (PBS; 1 mmol EDTA; 0,5% HSA; Xandoglobulin 10 μg/ml) with FITC-, PE- or APC-labeled antibodies. The fluorescence was measured on a FACS Canto flow cytometer and data were analyzed with Diva-Software (both from BD Bioscience). All isotype controls remained below a threshold of 1×103. In all figures displaying histograms, the vertical line indicates the threshold for positive cells in comparison to respective isotype controls. That is, marker-positive cells are indicated right from the line.

MLR assay

For MLR assay, CD4+ cells were isolated from freshly prepared PBMC using MACS beads (Miltenyi) according to the manufacturer's protocol. This procedure yielded a 96–98% pure CD4+ T-cell population as determined by FACS. For proliferation assays 5×104 CD4+ cells were cultured either alone or together with 5×103 MoDC or total blood DC in round bottom 96-well plates. After 72 h T cells were pulsed with 3H-Thymidine and proliferation was quantified 20 h later by determination of 3H-Thymidine incorporation in triplicates.

Signal transduction inhibitors assay

To study signaling pathways involved in B7-H1 regulation the following stimuli and inhibitors were used: standart maturating cytokine cocktail (10 ng/mL TNF-α, 1000 U/mL IL-6, 10 ng/mL IL-1β, 1 μg/mL PGE2), poly I:C (50 μg/mL), LPS (100 ng/mL), CpG ODN (50 ng/mL), 20 μM UO126 (MEK1/2), 10 μM SB203580 (p38 MAPK), 20 μM SN50 (NF-κB), 25 μM LY294002 (PI3K). For the inhibition assays the DC were pretreated with inhibitors 1 h before maturation stimuli were added. After 24 h cells were collected and subjected to FACS analyses. For the visualisation of phosphorylated and nonphosphorylated forms of ERK42/44 and p38 Western blot analysis was performed.

Western blot analysis

After induction with different stimuli for indicated periods of time DC were harvested, lysed in 2× sample buffer (100 mM Tris. pH 6.8; 4% SDS; 16% glycerol; 0.57 2-mercapto-ethanol; 0.01% Bromphenolblue), boiled at 95°C for 5 min and stored at −20°C.

The proteins were loaded on 10% SDS-PAGE, separated and transferred to PVDF membrane. The blots were pretreated for 2 h in 5% milk TBST solution (10 mM Tris, pH 8.0; 150 mM NaCL; 0.1% Tween-20), incubated over night with primary Ab in 5% BSA TBST, washed three times in TBST and incubated with HRP-conjugated secondary Ab with subsequent ECL (Lumineg TMA6, UK) detection. The bands were quantified using MetaVue software by recalculating the average Gray Level Intensity.

ELISA

For ELISA CD4+ cells were isolated using MACS beads (Miltenyi). Aliquots of 5×104 CD4+ cells and 5×103 DC were cultured alone or together in round bottom 96-well plates in a medium with or without the addition of anti-B7-H1, anti-CD86 or IgG1 isotype control antibodies. After 5 days, the production of cytokines in supernatants was determined in triplicates using IFN-γ and IL-2 ELISA kits (both R&D Systems, Wiesbaden, Germany) according to the manufacturer's instructions.

Statistical analysis

The results are expressed as mean±SD of data obtained in three and more experiments. Statistical significance was analyzed by Student's t-test. p<0.05 was considered significant. *p<0.05; **p<0.01.

Acknowledgements

The authors thank Theron Johnson for critical reading of the article. This work was supported by grants from the Deutsche Forschungsgemeinschaft (SFB 405 and Ma1924), the Helmholtz Alliance against Cancer, the European Union (LSHC-2005-518178) and the Tumor Center Heidelberg-Mannheim to A. H. Enk and K. Mahnke.

Conflict of interest: The authors declare no financial or commercial conflict of interest.

Ancillary