Plasmacytoid dendritic cells (pDCs) secrete large amounts of IFN-α upon exposure to virus, subsequently promoting and regulating innate and adaptive immune responses. However, little is known about the functional regulation of virus-activated pDCs after they exert functions in secondary lymph organs. Our previous studies show that splenic stromal microenvironment can down-regulate the T cell response by inducing generation of regulatory myeloid dendritic cells; therefore, we wondered whether the splenic stromal microenvironment can regulate the function of virus-activated pDCs. In this study, we provide evidences that the splenic stromal microenvironment can chemoattract vesicular stomatitis virus (VSV)-activated pDCs via stromal cell-derived dactor 1 (SDF-1), inhibit the secretion of IFN-α, IL-12, TNF-α, and expression of I-Ab, CD86, CD80, and CD40 by VSV-activated pDCs, and subsequently inhibit VSV-infected pDCs to activate NK cell IFN-γ production and cytotoxicity. Stroma-derived TGF-β participates in the negative regulation of VSV-activated pDCs. Therefore, we demonstrate that splenic stromal microenvironment negatively regulates the virus-activated pDCs through TGF-β, outlining an additional mechanistic explanation for maintenance of immune homeostasis.
Plasmacytoid dendritic cells (pDCs)4 act at the first line of defense in antiviral immunity, which is characterized by rapid and robust secretion of type I IFN in response to TLR7 or TLR9 agonists (1, 2). In addition to Th1 responses induced by activated pDCs, pDCs and pDC-derived IFN-α have been shown to promote NK cell activation and Ab production (3, 4), contributing to rapid elimination of invading pathogens. However, continuously activated pDCs and their overproduced IFN-α also play pathogenic roles in several autoimmune diseases such as lupus and psoriasis (5, 6). Therefore, to maintain immune homeostasis, there may exist some mechanisms in hosts for the control of activated pDCs. Therefore, it is important to investigate the mechanisms underlying negative regulation of activated pDCs. As we know, immune microenvironments, such as bone marrow, thymus, and peripheral lymphatic tissues, have been confirmed to play important regulatory roles in immune responses via inhibiting immune cell activation or inducing regulatory T cells and regulatory DCs (7, 8, 9, 10, 11). Recent evidence has revealed that pDC precursors are derived from bone marrow and then released into peripheral blood. Similar to T cell migration, pDCs are believed to migrate into secondary lymphatic tissues via high endothelial venules from circulation (12). These pDCs constitutively express CXCR4, CXCR3, and are induced to express CCR7 after stimulation, which endow pDCs with the ability to sense chemotactic signals from lymphatic tissues, such as SDF-1 and MIP-3β, and then migrate into lymphatic tissues where they interact with immune cells (13, 14). Up to now, it remains unclear whether immune microenvironments can regulate the functions of activated pDCs.
Spleen is one of the important secondary lymphoid organs where DCs present Ags to and activate T lymphocytes. We previously used mouse endothelial splenic stromal cells (ESSCs) to mimic the splenic immune microenvironment and found that ESSCs were involved in immune homeostasis by driving mature myeloid DCs and hemopoietic stem cells to differentiate into regulatory DCs (8, 15). Considering that pDCs, especially virus-activated pDCs, migrate into spleen, we wondered whether the splenic stromal microenvironment can negatively regulate the function of activated pDCs? In this study, we demonstrate that the splenic stromal microenvironment can negatively regulate VSV-activated pDCs via TGF-β, outlining an additional mechanistic explanation for maintenance of immune homeostasis.
Materials and Methods
Mice and cell culture
Five- to 6-wk-old C57BL/6 mice were obtained from Joint Ventures Sipper BK Experimental Animal and Smad3−/− mice were established as described previously (16), which were maintained under specific pathogen-free conditions. ESSCs derived from newborn mice were prepared and maintained in long-term culture in RPMI 1640 and 20% FCS by passaging to new plates each week as described previously (8).
RPMI 1640 medium (PAA Laboratories) supplemented with 10% FCS (PAA Laboratories), 2 mM l-glutamine, 1% sodium pyruvate, 2 × 10−5 M 2-ME (Sigma-Aldrich) was used throughout the experiments. Recombinant mouse FMS-like tyrosine kinase 3 ligand (Flt-3L) and TGF-β, neutralizing anti-SDF-1 and anti-TGF-β Abs, and isotype Ig were purchased from R&D Systems. Fluorescence-conjugated Abs specific for CD3, CD19, CD11b, DX5, B220, CD11c, I-Ab, CD40, CD80, and CD86 were obtained from BD Pharmingen. Fluorescence-conjugated anti-PDCA1 Ab and a mouse pDC isolation kit were from purchased from Miltenyi Biotec. A CFSE staining kit was obtained from Molecular Probes. The PKH26 staining kit and 7-AAD were obtained form Sigma-Aldrich.
Preparation of pDCs
pDCs were isolated from bone marrow cells using the two-step MACS: first pre-enrichment of pDCs by depletion of T cells, B cells, NK cells, and myeloid cells and then positive selection of B220+ pDCs. Routinely, >95% of cells were Lin−B220+CD11c+PDCA1+ cells. To obtain VSV-infected pDCs, purified vesicular stomatitis virus (VSV) was added to 1 × 105/ml pDCs at 128 PFU/ml for 4 h and then pDCs were extensively washed using PBS.
Coculture of ESSCs and pDCs
In brief, 1 × 105/ml pDCs infected with or without VSV were cocultured with 2 × 105/ml ESSCs or ESSC-derived supernatants (ESSC-SN), which were also supplemented with Flt-3L at 50 ng/ml for 24 h. ESSC-derived supernatant (ESSCs-SN) was obtained by culturing 2 × 105/ml ESSCs for 48 h. In some experiments, neutralizing anti-TGF-β Ab or its isotype Ig at 5 μg/ml was added into the culture system of ESSCs before adding pDCs. In other experiments, VSV-infected pDCs were incubated with rTGF-β for 24 h. Then, the culture supernatants and phenotype of pDCs were measured with ELISA and FACS, respectively. The production of IFN-α, IL-12p70, and TNF-α by pDCs was detected by ELISA (R&D Systems). The expression of CD40, CD80, CD86, and I-Ab on Lin−B220+ pDCs was analyzed by FACS (LSRII; BD Biosciences).
Chemotaxis of pDCs
pDCs or VSV-infected pDCs were cultured in the presence of 50 ng/ml Flt-3L for 8 h to obtain immature pDCs or activated pDCs, respectively. ESSC-mediated chemotaxis of pDCs and VSV-activated pDCs was detected using a 3-μm Transwell system (BD Biosciences) as described previously (17). Briefly, pDCs or VSV-activated pDCs were suspended in RPMI 1640 medium with 0.5% FCS and 10% ESSC-SN at a concentration of 5 × 105cells/ml. A total of 1 × 105 cells was loaded into the upper chambers and serial dilution of ESSC-SN into the lower wells. For neutralizing experiments, anti-SDF-1 Ab or its isotype Ig at 10 μg/ml was applied to ESSC-SN for 1 h at 37°C before adding to the lower wells. After incubation at 37°C for 2 h, pDCs in the upper wells were extensively removed and plates were centrifuged at 200 × g for 3 min. pDCs in lower wells were collected, stained with PE-anti-Lin and allophycocyanin-anti-B220, and then suspended in equal amounts of PBS. Each sample was counted for 106 s by FACSCalibur and analyzed by CellQuest software (BD Bisosciences). Each experiment was performed in triplicate.
Activation of NK cells by VSV-activated pDCs
After incubation with ESSC-SN for 24 h, 1 × 105/ml VSV-activated pDCs were cocultured with 5 × 105/ml freshly isolated splenic DX5+ NK cells for 12 h. Percentage of IFN-γ+NK1.1+ NK cells was detected by intracellular staining and FACS analysis. YAC-1 cells were labeled with CFSE according to the manufacturer’s instruction. The cytotoxicity of NK cells to CFSE-labeled YAC-1 cells was performed with 7-AAD staining of dead cells as previously described (18). The NK cytotoxicity (percent) was calculated using the following formula: (percent dead YAC-1 cells by NK cells − percent dead YAC-1 cells alone)/(1 − percent dead YAC-1 cells alone).
Assay for the phenotype and intracellular cytokine expression of in vivo-transferred VSV-activated pDCs
VSV-activated pDCs were stained with PKH26 and then i.v. administrated into congenic C57BL/6 mice (1 × 106/mouse). After 24 h, splenocytes of the recipient mice were collected, and then stained with FITC-anti-I-Ab, CD40, CD86, or IFN-α, respectively, combined with allophycocyanin-anti-B220 for pDCs. We analyzed PKH26+B220+ pDCs for the expression of I-Ab, CD40, CD86, and IFN-α, which represented the in vivo-transferred VSV-activated pDCs.
Statistical significance of differences was determined by the paired or unpaired two-tailed Student t test. Differences were considered statistically significant for p < 0.05.
Splenic stromal cells mediate the chemotaxis of pDCs and VSV-activated pDCs
It has been reported that pDCs and activated pDCs migrate to secondary lymph nodes from blood. To investigate whether ESSCs we prepared were functionally mimicking the spleen microenvironment, we first observed ESSC-mediated chemotaxis of pDCs and VSV-activated pDCs. As shown in (Fig. 1,A), ESSC-SN could chemoattract pDCs and VSV-activated pDCs in a dose-dependent manner. Moreover, ESSCs chemoattracted VSV-activated pDCs more efficiently than immature pDCs. We also found that CXCR4, which was constitutively expressed on pDCs at low levels, was significantly up-regulated after VSV infection, indicating that viral infection could up-regulate expression of some chemokine receptors on pDCs in favor of their migration into lymph tissues (Fig. 1,B). Our previous data showed that ESSCs expressed SDF-1 (8), one ligand of CXCR4; therefore, we blocked soluble SDF-1 in ESSC-SN with neutralizing anti-SDF-1 Ab and then observed the chemotaxis of pDCs by ESSC-SN. As shown in Fig. 1,C, the chemotaxis of pDCs to ESSCs was significantly reduced after blockade of SDF-1, whereas the chemotaxis of VSV-activated pDCs to ESSCs was partially reduced, suggesting that SDF-1 derived from splenic stromal cells was a major chemotactic signal to resting pDCs. Although SDF-1 also contributed to chemoattraction of activated pDCs by ESSCs, other known or unknown chemotactic signals from splenic stromal cells may play roles in chemoattracting the activated pDCs. The results indicated that ESSCs could exhibit splenic microenvironment characteristics, such as chemotaxis of pDCs. In addition, we found that freshly isolated pDCs or VSV-activated pDCs tended to be apoptotic even cultured in the presence of Flt-3L, whereas pDCs cultured in the presence of both ESSCs and Flt-3L can survive for a longer time (Fig. 1 D), suggesting that ESSCs we used supported the survival of pDCs and could be used to functionally mimic the spleen microenvironment.
Splenic stromal cells inhibit the secretion of IFN-α, TNF-α, and IL-12p70 from pDCs by VSV
Upon entering into spleen, pDCs will be subjected to the regulation by the splenic microenvironment. Since one of characteristics of pDCs is their rapid and robust secretion of IFN-α after viral infection, we analyzed the effects of ESSCs on cytokine secretion of pDCs after being infected by the virus. As predicted, VSV-activated pDCs could release high levels of IFN-α, IL-12p70, and TNF-α. However, the secretion of IFN-α, IL-12p70, and TNF-α from VSV-activated pDCs decreased significantly when cocultured with ESSCs; moreover, the decrease was also observed when VSV-activated pDCs were incubated with ESSC-SN, suggesting that ESSC-derived soluble factors contributed to the inhibition of cytokine secretion of VSV-activated pDCs by ESSCs (Fig. 2). We found that ESSCs almost did not secret IFN-α, IL-12p70, and TNF-α even after infection with VSV (Fig. 2, right panel), suggesting that these cytokines in the supernatants of coculture were from VSV-activated pDCs, but not from ESSCs.
Splenic stromal cells inhibit the expression of I-Ab, CD40, CD80, and CD86 on pDCs by VSV
Next, we observed the effect of ESSCs on the phenotype of VSV-activated pDCs. As shown in (Fig. 3), VSV infection up-regulated the expression of I-Ab and CD40, CD80, and CD86 on pDCs; however, after VSV-infected pDCs were cocultured with ESSCs for 24 h, the expression of I-Ab and CD40, CD80, and CD86 was markedly reduced as compared with those on VSV-activated pDCs without ESSCs.
Splenic stromal cells inhibit VSV-activated pDC-mediated activation of NK cells
It has been shown that pDC-derived IFN-α promotes the activation of NK cells that have enhanced IFN-γ secretion and cytotoxicity (3). Thus, we investigated whether ESSC-mediated inhibition of VSV-activated pDCs could affect the ability of pDCs to activate NK cells. As shown in (Fig. 4), VSV-activated pDCs could significantly activate NK cells, leading to the enhanced cytotoxicity of YAC-1 cells and the increased number of IFN-γ+NK1.1+ NK cells. However, the cytotoxicity of NK cells to YAC-1 cells (Fig. 4,A) and the number of IFN-γ+NK1.1+ NK cells (Fig. 4 B) were significantly reduced when VSV-activated pDCs were pretreated with ESSC-derived supernatant. Therefore, the results demonstrated that the splenic stromal microenvironment can negatively regulate NK cell activation via silencing VSV-activated pDCs.
Splenic stromal microenvironment inhibits VSV-activated pDCs through TGF-β
Above data suggested that ESSC-derived soluble factors played inhibitory roles in the activation of pDCs by VSV infection. We previously demonstrated that ESSCs expressed several immunosuppressive factors, including TGF-β (8). Since TGF-β is abundant in the ESSC-SN and a well-known negative regulator of immune responses (19), we investigated whether ESSC-derived TGF-β participates in inhibiting the functions of VSV-activated pDCs. We found that ESSC-mediated suppression of IFN-α, IL-12p70, and TNF-α secretion from VSV-activated pDCs was partially reversed when TGF-β was blocked by neutralizing Ab (Fig. 5,A). Moreover, we observed that rTGF-β could directly inhibit the secretion of IFN-α, IL-12p70, and TNF-α secretion from VSV-infected pDCs (Fig. 5,B). In contrast to normal pDCs, VSV-activated Smad3−/− pDCs were still able to secrete comparable levels of IFN-α in the presence or absence of ESSCs (Fig. 5,C). Furthermore, VSV-activated Smad3−/− pDCs pretreated with ESSC-SN could effectively activate NK cells to kill YAC-1 cells and secret IFN-γ (Fig. 5, D and E). The results demonstrate that ESSC-derived TGF-β is responsible for the ESSC-mediated negative regulation of functions of VSV-activated pDCs.
Splenic microenvironment inhibits the expression of I-Ab, CD86, CD40, and IFN-α in VSV-activated pDCs transferred in vivo
To further confirm whether the splenic microenvironment negatively regulates VSV-activated pDCs in vivo, we transferred PKH26-labeled VSV-pDCs to congenic mice and then observed the phenotype and IFN-α expression of the transferred VSV-infected pDCs located in spleen 24 h after transfer. As shown in Fig. 6, the expression of I-Ab, CD40, CD86, and IFN-α of the transferred PKH26+B220+ pDCs in recipient spleen was reduced significantly as compared with that of PKH26+B220+ pDCs before transferred. Therefore, these results indicated that the splenic microenvironment can negatively regulate the activation of VSV-infected pDCs in vivo.
In addition to the negative regulation of immune response by driving mature myeloid DCs to proliferate and differentiate into regulatory DCs, here we for the first time demonstrate that the splenic stromal microenvironment can inhibit the virus-activated pDCs to secrete IFN-α, IL-12, and TNF-α, express I-Ab, CD40, CD80, and CD86, and activate NK cells, suggesting that virus-activated pDCs could be silenced by the stromal microenvironment after their migration into secondary lymphoid organs. Our results suggest a new manner for the negative regulation of the virus-activated pDCs after they complete their task to activate an immune response.
It is well known that IFN-α secreted from virus-activated pDCs can activate NK cells efficiently and then the activated NK cells with increased cytotoxicity and IFN-γ secretion contribute to the elimination of invading pathogens (3). pDCs and NK cells are the main effectors of innate immunity. Stromal cell-mediated inhibition of IFN-α production from the VSV-activated pDCs may be responsible for the down-regulation of pDC-mediated NK cell activation. In addition, IFN-α has also been reported as an important regulator of IL-15 expression, a critical cytokine for NK survival and expansion (20). Therefore, in addition to negative regulation of T cell response by generating regulatory myeloid DCs at the late stage of immune responses (8), the splenic stromal microenvironment may directly, at least partially, control the innate immune responses to an appropriate level after viral infection by down-regulating the function of virus-activated pDCs and NK cells.
Upon recognition of pathogens and activation, pDCs can secrete Th1 cytokine IL-12 and proinflammatory cytokine TNF-α and act as initiator and activator of adaptive immunity (1). In addition to inhibition of IFN-α production, we found that splenic stromal cells could inhibit the production of IL-12 and TNF-α from virus-activated pDCs. Moreover, the expression of I-Ab, CD40, CD80, CD86, and the T cell stimulatory phenotype of pDCs was also down-regulated by coculture with splenic stromal cells. These results suggest that the splenic stromal microenvironment may control the activated pDCs to activate the Th1 response and trigger inflammation. Considering that immature pDCs have been proposed to be involved in induction of Th2 or regulatory T cells (Treg) generation (21, 22) and splenic stromal cells could regulate the activated pDCs to become a low expression of I-Ab, CD40, CD80, and CD86 and a low secretion of Th1 cytokine, just like the immature pDCs, the reversion of virus-activated pDCs into immature pDC-like status by the splenic stromal microenvironment suggests a reduction in Th1 vs Th2 generation or Th1 vs Treg generation. Both Th2 and Treg responses contribute to suppression of Th1 immune responses; thus, the results propose one possibility that the splenic stromal microenvironment may exert a negative regulatory role in adaptive immune responses via affecting the balance of Th1 vs Th2 or Th1 vs Treg. Such a possibility needs to be investigated in the future.
In our study, we found that pDCs were able to up-regulate the expression of CXCR4 after viral infection, and splenic stroma-derived SDF-1 indeed chemoattracted pDCs or VSV-activated pDCs. However, blockade of SDF-1 only partially inhibited the chemoattraction of VSV-activated pDCs by ESSCs, suggesting that, besides CXCR4/SDF-1, other factors may also contribute to the migration of the VSV-activated pDCs into secondary lymphatic tissues. Moreover, our finding that the splenic stromal microenvironment maintains survival of pDCs or virus-activated pDCs, and that the splenic stromal microenvironment can down-regulate the enhanced expression of CXCR4 to the levels of resting pDCs, may reveal one possibility for the recycle of pDCs, i.e., pDCs migrate from blood into secondary lymph nodes and then after staying for a certain time they make ready for the next migration, leaving out of secondary lymph nodes to peripheral blood. Our unpublished data also showed that VSV infection of the cellular mixture of pDCs and splenic stromal cells could make pDCs secret large amounts of IFN-α, suggesting that splenic stromal cells could not impair resting pDCs to recognize viral infection and pDCs resident in spleen maintained their ability to respond as innate immune cells to the virus invasion. However, the hypothesis on the pDCs recycling needs to be investigated in the future.
In summary, here we provide an additional mechanistic explanation for the maintenance of immune homeostasis after virus infection by showing that splenic stromal microenvironment negatively regulates the virus-activated pDCs through TGF-β. Functional regulation of pDCs by the stromal microenvironment will be helpful to the understanding of the immunobiology of pDCs in physiological conditions and several pathological conditions, such as lupus. As we know, overactivation of pDCs and overproduction of type I IFN are regarded as one reason for the pathogenesis of lupus. Future study will be performed to investigate the regulation of the activated pDCs by stromal cells derived from the lupus model, lpr/MRL and NZB/ZWF F1 mice, possibly outlining new pathogenic mechanisms or intervention targets of the autoimmune diseases.
We thank J. Jiang for technical assistance.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by grants from the National Natural Science Foundation of China (30490240, 30121002, 30471573), National Key Basic Research Program of China (2007CB512403), and Shanghai Committee of Science and Technology (05DZ22106).
Abbreviations used in this paper: pDC, plasmacytoid dendritic cell; DC, dendritic cell; ESSC, endothelial splenic stromal cell; VSV, vesicular stomatitis virus; SDF-1, stromal cell-derived factor 1; Flt-3L, FMS-like tyrosine kinase 3 ligand; 7-AAD, 7-aminactinomycin D; ESSC-SN, ESSC-derived supernatant; Treg, regulatory T cell.