Absence of IFN-β Impairs Antigen Presentation Capacity of Splenic Dendritic Cells via Down-Regulation of Heat Shock Protein 70¹

Dendritic cells (DCs)⁵ are essential for the induction of specific immune responses and represent the most important cellular link between the innate and the adaptive immune system (1, 2, 3, 4, 5). DCs are found in most tissues where they capture and transport Ag to draining lymph nodes. During this migration, DCs mature and become highly stimulatory for T as well as B cells (6). In addition to DCs that emigrate from peripheral tissues, resident DCs can also be found in lymphoid organs such as spleen. These DCs are crucial for the sampling of blood-borne Ags or pathogens (2).

DCs are generally considered as “professional APCs,” in which two principal Ag presentation pathways can be distinguished. Endogenous Ags like self or viral components are presented via MHC class I (MHC I) molecules to CD8⁺ T cells, while exogenous Ags are presented via MHC class II (MHC II) to CD4⁺ T cells. In addition, DCs have the unique capacity to deliver exogenous Ags into the MHC I presentation pathway, a process known as cross-presentation. This enables CD8⁺ T cells to respond against Ags that are not directly expressed within DCs (7, 8, 9)_. The development, migration, maturation, and function of DCs are critically influenced by cytokines produced in their surroundings (10, 11), including type I IFNs.

IFNs encompass a large family of closely related cytokines comprising at least 13 IFN-α isotypes and a single IFN-β. Both IFN-α and IFN-β exert their activity through a common receptor IFNAR (12). IFN-β is thought to be the master regulator in that it is rapidly induced and can, in turn, induce the other IFN isotypes (13, 14). Furthermore, even in the absence of infection, spontaneous IFN-β production, albeit at a low level, is known to occur (12, 15, 16). These spontaneously produced IFNs contribute to host defense and cell growth in a manner similar to those induced by pathogens. In addition, constitutive production of IFNs is crucial for maintaining cells in a “primed” state and thus enabling them to mount a rapid and robust response upon encounter of external stimuli. It has thus been proposed that the absence or dysregulation of the basal constitutive IFN signaling could be the reason for development of certain diseases (12, 15, 16, 17, 18).

In the present study, we addressed the question whether spontaneously produced IFNs play a role in the development of cell-mediated immunity. Comparing the function of splenic conventional DCs (cDCs) from wild-type (WT) mice and mice deficient in either IFN-β or IFNAR, we found that IFN-β serves as a crucial factor for maturation of the T cell stimulatory capacity of cDCs via MHC I and MHC II. In its absence, we detected a lower number of specific MHC-peptide complexes at the surface of splenic cDCs. We also found that the diminished T cell stimulatory capacity of splenic cDCs occasioned by the in vivo absence of IFN-β is due to low expression of heat shock protein 70 (Hsp70), which is required for efficient generation of stable MHC-peptide complexes expressed on the cell surface of cDCs. Consistent with these findings, cDCs from Hsp70-deficient mice (Hsp70.1/3^−/−) were impaired in their capacity to present soluble Ags to naive T cells.

Female IFN-β^−/− (14), IFN-β^+/+, IFNAR^−/− (19) C57BL/6, OT I, and OT II mice (20, 21) were bred at the animal facility of the Helmholtz Centre for Infection Research (HZI). Female C57BL/6 mice were obtained from the Harlan-Winkelmann. The initial generation of Hsp70.1/3 knockout mice has been previously described (22). The C57BL/6 Hsp70.1/3^−/− mice were derived by transfer from a 129 background into the C57BL/6 background and were raised at Washington University School of Medicine (St. Louis, MO). All mice were used between 8 and 12 wk of age. Mice were bred and maintained in specific pathogen-free conditions. Mouse care and experimental procedures were performed under the approval of local authority Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit.

The B3Z T cell hybridoma (23) specific for the H-2K^b-SIINFEKL complex was maintained in IMDM supplemented with 10% FCS, 100 U/ml penicillin, and 100 μg/ml streptomycin. The hybridoma 25-D1.16 secreting an IgG1κ mAb specific for the pOV8 · H-2K^b (24) was provided by Dr. R. Germain (National Institute of Allergy and Infectious Diseases, Bethesda, MD). Ab was purified and conjugated with FITC according to standard procedures.

Spleen cells were prepared by gentle flushing out the splenocytes with IMDM supplemented with antibiotics (100 U/ml penicillin and 100 μg/ml streptomycin) and 10% FCS, 50 μM 2-ME, and 2 mM l-glutamine. Erythrocytes were lysed for 2 min in ACK buffer (0.15 M NH₄Cl, 10 mM KHCO₃, and 0.1 mM EDTA) and washed two or three times in PBS. Cell clumps were removed by passage through a 50-μm nylon filter. Splenocyte preparation was conducted strictly on ice. Splenic cells were than stained with appropriate Abs and cDCs were sorted (see below).

Single-cell suspensions were treated with anti-mouse CD16/CD32 BD Fc Block (2.4G2; BD Biosciences) for 10 min followed by staining with appropriate mAbs for 20 min on ice. Abs used in this work included: CD11c (clone N418) conjugated with al1ophycocyanin or PE-Cy7, CD11b (M1/70) PE-Cy7, CD8α (53-6.7) Pacific blue, FITC, or PE, B220 (RA3-6B2) allophycocyanin-Alexa Fluor 750, CD4 (GK1.5) PE, CD86 (GL1) FITC, CD40 (HM40-3) FITC, CD80 (16-10A1) PE, CD1d (1B1) PE, ICAM-1 (YN1/1.7.4) PE, B7-H1 (MIH5) PE, B7-H4 (MIH29) PE, and B7-RP1 (HK5.3) PE (all purchased from eBioscience). H-2K^b (Y-3) FITC, I-A^b (M5/114.15.2) FITC. Anti H-2K^b-SIINFEKL complex Ab (clone 25-D1.16) FITC was purified and conjugated with FITC in our laboratory. cDCs were sorted as cells exhibiting low side scatter (SSC^low) in addition to being CD11c^highCD8α⁺CD11b⁻, CD11c^highCD8α⁻CD11b⁺. Flow cytometric analysis and sorting were performed using FACSCanto, LSRII, and FACSAria (BD Biosciences), respectively. Final purity of cDCs was always >97%. All samples during the sorting procedure were kept at 4°C. The data were analyzed using FACSDiva (BD Biosciences) software.

OT I (OVA-specific CD8⁺ T cells) and OT II (OVA-specific CD4⁺ T cells) were isolated from lymph nodes (s.c. and mesenteric) and spleen. Single-cell suspensions were further purified using the CD8- or CD4-negative isolation kits (Dynal) containing mAb against B220, CD11b, Ter-119, CD16/32, and CD4 (for OT I isolation) or CD8 (for OT II isolation) following the protocol provided by the manufacturer. Cell preparations contained >90% of the desired cell population and were essentially free of CD11c^high cells as determined by flow cytometry using Abs specific for CD4 or CD8 and CD11c, respectively. For Ag presentation assays, OT I or OT II cells were stained with 1 μM CFDA(SE) (Molecular Probes) for 10 min at 37°C according to the manufacturer’s protocol.

For the experiments using soluble OVA or peptides, individual APC populations were plated in 96-well plates (Nunc) at 2 × 10⁴ cells/well with the indicated amount of soluble EndoGrade OVA (Profos), OVA_257–264 (SIINFEKL; Ana Spec), or OVA_323–339 (provided by Dr. W. Tegge, HZI) for 1 h at 37°C in complete medium. The cells were further washed three times and resuspended in complete medium containing 2 × 10⁵ CFSE- labeled OT I or OT II cells. Proliferation of T cells was analyzed by flow cytometry after 1.5 (OT I peptide) or 2.5 days of culture. Cells were stained with anti-CD4 or anti-CD8 mAbs for 20 min, washed, and resuspended in 200 μl of PBS containing Cy5-labeled 0.6-μm latex beads. Samples were analyzed until 2 × 10⁴ beads were collected. The number of divided cells (propidium iodide^low, CFSE^low, CD4⁺, or CD8⁺) was determined as previously described (25). For ex vivo experiments, mice were injected i.v. with 1 mg of OVA or 1 mg of OVA along with 200 μg of polyinosinic:polycytidylic acid (poly(I:C); Fluka). Twenty-four hours later, mice were sacrificed, APCs were sorted, and incubated with OT I or OT II cells for 2.5 days. In some experiments, murine recombinant IFN-β (R&D Systems) was added to the cultures of T cells and APCs.

Sorted DCs were incubated with 62.5 μg/ml DQ-OVA (Molecular Probes) for 45 min at 37°C or on ice. DCs were then washed and analyzed by flow cytometry.

Sorted DCs (10⁴ cells/well) were pulsed for 1 h with various concentrations of SIINFEKL peptide, washed twice, and resuspended in phenol red-free RPMI 1640 (Life Technologies) containing 100 U/ml penicillin and 100 μg/ml streptomycin, 1% FCS, and 2 mM l-glutamine. DCs were then cocultured in a 96-well U-bottom plate with 5 × 10⁴ B3Z cells/well overnight at 37°C. The next day 150 μl of supernatant was taken from each well and replaced with 150 μl of PBS containing 5 mM ONPG (Sigma-Aldrich) and 0.5% IGEPAL-20 (Sigma-Aldrich). The plate was than incubated at 37°C for 2 h and OD was measured at 450 nm with wavelength correction set at 650 nm.

RNA isolation, cDNA preparation, and DNA microarray analysis of gene expression were performed at the gene array facility of the HZI. Fluorescent images of hybridized microarrays (MOE-430 version 2.0; Affymetrix) were obtained using an Affymetrix Genechip Scanner. Microarray data were analyzed using BioConductor Suite 2.1 software. All samples were repeated two times with individually sorted cells and averaged. Data discussed here have been deposited in National Center for Biotechnology Information’s Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/) and are accessible through GEO series accession number GSE12392.

Total RNA was extracted from sorted APCs using a RNeasy mini kit (Qiagen) according to the manufacturer’s instructions. DNA contamination in the total RNA preparation was eliminated using DNase I (Qiagen). Oligo(dT)₁₈ primers and a RevertAid First Strand cDNA Synthesis Kit (Fermentas) were used for reverse transcription of purified RNA. All gene transcripts were quantified by quantitative PCR with Power SYBR Green qPCR Master Mix (Applied Biosystems) and a Light Cycler apparatus (Applied Biosystems PRISM Cycler). Primers specific for Hsp70.1 were synthesized as described before (26).

Splenic cDCs were sorted out and stained intracellularly using a Cytofix/Cytoperm kit (BD Biosciences) according to the manufacturer’s protocol with anti-Hsp70 (C92F3A-5) PE-conjugated Ab (Stressgen).

DSG was a generous gift from Nippon Kayaku. Animals were injected i.p. daily with 10 mg/kg DSG or PBS for 6 days before splenic cDCs were sorted out and tested for their Ag presentation capacity with OT I or OT II cells in a CFSE dilution assay.

IFNs are known to be constitutively produced at low levels under noninflammatory conditions (12). To study the influence of IFNs on Ag presentation under physiological conditions, we decided to focus on freshly isolated splenic cDCs. These cells are representative of typical nonmigratory DCs found in vivo at steady state (1, 2, 3). Analyzing IFN-β- and IFNAR-deficient mice, we detected no differences with regard to the percentage of various splenic cDCs subpopulations in mice with and without either IFN-β or IFNAR (Fig. 1,A). Furthermore, we determined the overall number of leukocytes in several lymph nodes and spleen. Consistently, there was no significant difference observed in comparison to WT mice (Fig. 1 B).

Thus, we analyzed the ability of cDCs exhibiting the markers CD11c^high, CD11b^+/−, CD8α^+/−, and B220⁻ from spleens of WT, IFN-β^−/−, and IFNAR^−/− mice to present OVA protein to CFSE-labeled OT I or OT II T cells. When compared with WT, cDCs from IFN-β^−/− and IFNAR^−/− mice were severely impaired in their ability to activate such CD8⁺ and CD4⁺ T cells (Fig. 2, A and C).

To test whether the anomaly associated with IFN-β or IFNAR deficiency affects also the presentation of preprocessed Ag, i.e., peptides, cDCs from WT, IFN-β^−/−, and IFNAR^−/− mice were loaded with MHC I- or MHC II- specific peptides (OVA_257–264 (SIINFEKL) and OVA_323–339, respectively) and then incubated with CFSE-labeled OT I and OT II T cells. As shown in Fig. 2, B and D, T cell stimulation was also highly impaired when peptide-loaded cDCs from IFN-β^−/− and IFNAR^−/− mice were used.

Throughout most of the experiments, we used bulk-sorted splenic cDCs because during in vitro cocultures with T cells the two distinct populations, CD8α⁺ DCs and CD8α⁻ DCs (myeloid DCs), from IFN-β^−/− and IFNAR^−/− mice were similarly impaired in their T cell stimulatory capacity compared with WT DCs (supplemental Fig. S1⁶).

Since IFNs provide cellular survival signals under certain conditions (27, 28), we first wanted to test whether the reduced ability to stimulate T cells might be due to lower survival of cDCs from IFN-β^−/−or IFNAR^−/− mice during the in vitro T cell stimulation assay. Splenic cDCs sensitized with OVA protein were incubated with OT I or OT II cells. After 16 and 32 h, the percentage of live cDCs was assessed by propidium iodide exclusion. WT, IFN-β^−/−, and IFNAR^−/− cDCs were equally viable under these conditions (supplemental Fig. S2, A and B). Thus, the reduced ability to stimulate T cells in vitro was not due to lower survival of IFN-β^−/− cDCs.

Next, we asked whether exogenous administration of IFNs could restore the impaired T cell stimulatory capacity of IFN-β^−/− cDCs in vitro. Titration of murine rIFN-β into cocultures of IFN-β^−/− cDCs and T cells showed that low amounts (0.1 U/ml) could completely restore the impaired T cell stimulatory function (Fig. 3,A). However, probably due to activation of negative feedback mechanisms, addition of higher concentrations of rIFN-β (5–500 U/ml) to the cocultures failed to restore the T cell stimulatory ability of cDCs (Fig. 3 A and data not shown). These results support the argument that the low levels of IFN-β produced at steady state are well optimized for maintaining cDCs in the Ag presentation-competent state.

Nevertheless, in such a situation it is difficult to exclude that exogenous rIFN-β influenced T cell proliferation. A direct effect of IFNs on T cells has been well documented (28, 29), although it only partially could explain our results. We could show that IFNAR^−/− cDCs, which are able to produce IFN-β (data not shown), are still inefficient in activating a T cell response (Fig. 2). In addition, when such cocultures are complemented with rIFN-β, the inefficiency of T cell activation remained (Fig. 3 B). This clearly demonstrates that steady-state production of IFN-β is required for maintenance of proper cDC function.

Furthermore, we tested whether triggering IFN-α in vivo could compensate the impaired development of T cell stimulatory capacity of cDCs from IFN-β^−/− mice. WT, IFN-β^−/−, and IFNAR^−/− mice were i.v. injected with OVA alone or OVA in combination with poly(I:C). After 24 h, splenic cDCs were sorted and tested for their ability to activate the proliferation of OT I or OT II T cells. Data depicted in Fig. 3,C show that IFN-α induction by poly(I:C) compensated for the lack of IFN-β during cDC development in vivo and partially recovered their function. As expected, in vivo administration of poly(I:C) did not improve the stimulatory function of splenic cDCs from IFNAR^−/− mice as they are completely unresponsive to IFN signaling (Fig. 3 D).

We also tested the T cell stimulatory capacity of DCs differentiated in vitro by incubating bone marrow cells with IL-4 and GM-CSF (bone marrow-derived DCs). After 8 days of culture, we obtained ∼80% CD11c-positive cells from WT, IFN-β^−/−, and IFNAR^−/− mice. In this study, WT and IFN-β^−/− or IFNAR^−/− bone marrow-derived DCs were comparable in their ability to stimulate the proliferation of T cells (data not shown). This suggests that the influence of IFNs observed ex vivo is greatly dependent on the overall stimulatory context under which the DCs develop.

A differential ability to acquire and process soluble Ag could account for the diminished stimulatory capacity of cDCs in the absence of the IFN system. Therefore, the efficiency of splenic cDCs from WT, IFN-β^−/−, and IFNAR^−/− mice to take up and degrade soluble OVA was assessed. We used DQ-OVA, which generates fluorescent byproducts upon degradation. As shown in Fig. 4, splenic cDCs from all groups acquired and generated comparable amounts of fluorescent DQ-OVA products. This was true for different DQ-OVA concentrations tested (data not shown). Thus, changes in Ag uptake and degradation could not account for the impaired T cell stimulatory capacity of cDCs from IFN-β^−/− or IFNAR^−/− mice.

The maturation status of DCs is known to be a fundamental factor for proper T cell stimulation. One of the mechanisms by which constitutive IFN signaling could influence T cell stimulation is to enhance expression of MHC or adhesion and costimulatory molecules on the surface of DCs (18). However, analysis of splenic CD8α⁺ and CD8α⁻ cDCs for MHC I and MHC II as well as costimulatory or adhesion molecules like CD86, CD80, CD40, and ICAM-1 indicated no significant differences between WT, IFN-β^−/−, and IFNAR^−/− mice in both cDC populations (Fig. 5 and supplemental Fig. S3). Therefore, the impaired function of splenic cDCs from IFN-β- and IFNAR-deficient mice was not due to lower expression of such surface molecules.

The fact that an impaired T cell stimulatory capacity of cDCs from IFN-β^−/− and IFNAR^−/− mice was also found for the presentation of peptides, not requiring further processing, as well as unimpaired DQ-OVA degradation suggested that the phenotype of IFN-β^−/− and IFNAR^−/− cDCs was most likely due to a defect in peptide presentation rather than Ag processing steps. We therefore decided to study the MHC-peptide complexes on the surface of cDCs. We used B3Z cells, a H-2K^b-restricted T cell hybridoma specific for the OVA epitope SIINFEKL (OVA_257–264), which upon TCR activation expresses β-galactosidase (23). The activation of B3Z cells, being a hybridoma, is independent of costimulation (30); thus, their activation should only be dependent on the concentration of MHC I-peptide complexes recognizable by the TCR. Therefore, splenic cDCs from WT, IFN-β^−/−, and IFNAR^−/− mice were loaded with OVA_257–264 peptide and tested for their ability to activate B3Z cells. As shown in Fig. 6, A and B, cDCs from mice deficient in IFN-β or IFNAR exhibited lower stimulatory capacity. These results confirmed that the impaired function of splenic cDCs from such mice is not due to lower levels of costimulatory molecules, but strongly suggested that the defect was in the process of MHC I-peptide complex formation. By using the 25-D1.16 Ab which recognizes SIINFEKL bound to the H-2K^b molecule (24), we confirmed our interpretation. DCs from either IFN-β^−/− or IFNAR^−/− mice had lower levels of MHC I-SIINFEKL complexes compared with cDCs from WT mice (Fig. 6 C).

To understand the molecular basis for the decreased formation of MHC-peptide complexes, splenic DC RNA from WT, IFN-β^−/−, and IFNAR^−/− mice was analyzed by microarrays for expression of genes known to be involved in Ag processing and presentation, costimulation, or IFN response. Extensive analysis of the microarrays indicated that most of these genes were unaltered in cDCs from the knockout mice (supplemental Fig. S4, A–C). The only dramatic difference found was in the expression of the heat shock protein Hsp70.1 and Hsp70.3 genes. The expression of these two genes was significantly lower in cDCs from IFN-β^−/− (∼15- to 20-fold down-regulated in comparison to WT) and IFNAR^−/− (∼75- to 150-fold down-regulated in comparison to WT) mice (Fig. 7 A).

To verify the above findings, we first stained for intracellular Hsp70 protein. As shown in supplemental Fig. S5, levels of Hsp70 were indeed lower in IFN-β^−/− and IFNAR^−/− cDCs compared with WT cDCs. Most likely due to presence of other highly homologous members of the Hsp70 family and the low sensitivity of the Ab, differences were not very pronounced. However, in confirmation of the microarray data, transcriptional levels of Hsp70.1 were severely decreased in DCs from IFN-β^−/− and IFNAR^−/− mice in comparison to WT (Fig. 7,B). Moreover, treatment with low amounts of rIFN-β (0.1 U/ml) increased Hsp70 levels in both WT and IFN-β^−/− DCs, whereas treatment with 5 U/ml rIFN-β did not markedly change the Hsp70 levels (Fig. 7,C). This correlates well with the functional restoration of IFN-β^−/− cDCs at low, but not at high concentrations of rIFN-β. Consistent with this finding, 24 h of poly(I:C) administration up-regulated Hsp70.1 levels in WT as well as in IFN-β^−/− cDCs (Fig. 7 D).

To test for a possible causative link between Hsp70 expression and Ag presentation, we used DSG, a pharmacological inhibitor of Hsp70. DSG is a synthetic derivative of spergualin from Bacillus laterosporus and binds to Hsp70 and Hsp90 (31, 32, 33). Therefore, we treated mice for 6 days with DSG and then tested the splenic cDCs of such mice for Hsp70 expression. Intracellular staining revealed that DSG treatment led to partial reduction of the Hsp70 level in WT cDCs (Fig. 8 A). In contrast, expression of surface molecules involved in T cell stimulation was not affected by this treatment (supplemental Fig. S6).

We then tested cDCs isolated from DSG-treated mice for their capacity to stimulate CD4⁺ and CD8⁺ T cells. Clearly, cDCs from DSG-treated mice exhibited a reduced ability to stimulate OT I or OT II T cells compared with cDCs from untreated mice, independent of whether protein or peptides were used as an Ag (Fig. 8 B). This was consistent with the claims that DSG abrogates the ability to present Ag in the context of both MHC I and MHC II (31, 32, 33, 34).

In addition, compared with untreated control, cDCs from DSG-treated mice revealed lower surface levels of MHC-peptide complexes (Fig. 8 C), suggesting that Hsp70 is necessary for efficient formation of MHC-peptide complexes. Thus, by supporting the expression of Hsp70, constitutive IFN-β expression in vivo helps to maintain cDCs in a primed and competent state for Ag presentation.

To explicitly demonstrate the involvement of the Hsp70.1 and Hsp70.3 proteins in Ag presentation and thus confirm that in the absence of IFN-β or IFNs signaling down-regulation of Hsp70 results in impaired Ag presentation, we used Hsp70.1/3 double knockout mice (22). The surface phenotype of cDCs from Hsp70.1/3^−/− mice appeared to be very similar to the surface phenotype of cDCs from WT (data not shown). To test the Ag presentation capacity of Hsp70.1/3^−/− cDCs, cells were sorted and loaded with appropriate OVA peptides or whole protein and incubated with OT I or OT II-transgenic T cells. The results clearly show that cDCs from Hsp70.1/3^−/− are impaired in their ability to present OVA-derived peptides as well as whole protein to naive T cells when compared with WT cDCs (Fig. 8 D). This further substantiates our finding that down-regulation of Hsp70 in the absence of IFN-β or IFNs can alter Ag presentation. The discovery that IFN signaling regulates MHC-peptide complex formation by Hsp70 proteins highlights a hitherto unrecognized mechanism via which IFNs might regulate presentation of self-Ags in the steady state and has, therefore, important consequences for our understanding of how regular homeostatic conditions are maintained in the immune system.

IFNs, found in high amounts in cells exposed to viruses, were first characterized and named as such on the basis of their antiviral activity. It is now well known that IFNs have widely overlapping, pleiotropic and immunomodulatory effects and their production is not the sole preserve of viral infections, but they are also induced in response to bacterial and parasitic infections (12, 15, 16). IFNs represent important immunomodulators for the innate as well as the adaptive arm of the immune system (16, 18). They exert broad regulatory effects and various subtypes of DCs are affected by these cytokines (12, 15, 16, 17, 18). For instance, IFN-α can promote Ag cross-presentation by enhancing endosomal processing, up-regulating the expression of costimulatory molecules, and augmenting DC viability in settings of viral infection (18, 27, 29, 35, 36, 37). Additionally, direct stimulation of T cells by IFN-α has been shown to be essential for efficient induction of cross-priming (29). Furthermore, IFN-α/β were described as crucial survival factors for activated T cells (28). Importantly, apart from that, even in the absence of infection, spontaneous low level production of IFN-β has been shown to occur (12).

The host response elicited by IFNs is largely dependent on signal strength. Most of the studies conducted to date have focused on the cellular effects induced by the high levels of IFNs elicited under inflammatory conditions. However, whether the low levels of IFNs produced under noninflammatory conditions have an important housekeeping immune function is not known. We now show that the low but constitutive production of IFN-β is necessary for maintaining DCs in a state competent for Ag presentation. Compared with those from WT mice, DCs freshly isolated from spleens of IFN-β^−/− and IFNAR^−/− mice were found to be highly impaired in Ag presentation to CD4⁺ and CD8⁺ T cells. This defect could in part be rectified with exogenous rIFN-β or through in vivo induction of IFN-α using synthetic dsRNA. Interestingly, restoration of function was possible with extremely low amounts of rIFN-β, probably mimicking amounts produced constitutively. Failure of IFN-β^−/− DCs cultured with high rIFN-β levels to activate T cell proliferation might be attributed to negative feedback mechanisms activated by a strong IFN-β signal.

The function of DCs is not only influenced by cytokines present in their environment, but also by other cells of the immune system, particularly T cells. Therefore, splenic cDCs isolated from IFN-β-deficient mice theoretically could be altered due to an effect of IFN-β deficiency on T cells. It was shown before that IFN-α and IFN-β have a direct effect on activated T cells and prevent their death during inflammatory conditions (28). Activated T cells can provide feedback signals to DCs and induce their maturation. This can be mediated by cytokines produced by T cells as well as by cell-cell interactions, including CD40L-CD40 interactions (38). However, this phenomenon is unlikely to play a significant role in steady-state conditions because of the lack of activated T cells. T cell-mediated conditioning of DCs via B7-H1 in homeostasis has recently been shown to play an important role in inducing DC maturation (39). However, since surface expression of B7-H1 was not altered on IFN-β^−/− and IFNAR^−/− cDCs in comparison to WT cDCs, one may assume that this is not the case in our situation. Moreover, we demonstrated that rIFN-β treatment in vitro can regulate cDCs function and Hsp70 levels.

Potential roles of IFNs in Ag presentation have previously been postulated. However, such effects have only been observed for high levels of IFNs reminiscent of inflammation. For instance, induction of IFNs in DCs by dsRNA and LPS in vitro was shown to up-regulate costimulatory and MHC molecules, hence enhancing their ability to activate CD8⁺ T cells (17). Conversely, our results show that the Ag presentation defects in splenic IFN-β^−/− and IFNAR^−/− DCs are neither due to low expression of costimulatory and MHC molecules nor a block in Ag capture and processing. We demonstrate that the defect is due to a blockade in a step downstream of Ag processing: MHC-peptide complex formation. Furthermore, the data presented here strongly suggest that impairment in the MHC-peptide complex formation step caused by the absence of IFN signaling is due to down-regulation of Hsp70.

The 70-kDa Hsp70.1 and Hsp70.3 belong to a larger, highly homologous and conserved gene family whose expression can be significantly induced in response to a number of pathophysiological conditions, including pathogen exposure (40). In addition to a generalized role in protein folding and transport, the proteins also have distinct functions in the promotion of Ag processing and presentation (31, 41, 42, 43). Hsp70 proteins are involved in chaperoning proteins/peptides during degradation and during Ag presentation via MHC I as well as via MHC II. Hsp70 has been shown to physically associate with the TAP, hence enabling efficient loading of chaperoned peptides onto MHC I molecules (31, 41, 44). In addition, a role for Hsp70 in Ag presentation by MHC II molecules has also been described (31, 45).

By microarray and quantitative real-time PCR (qRT-PCR) analysis, we found that in the absence of a functional IFN system in vivo, the expression of Hsp70.1 and Hsp.70.3 in splenic cDCs was significantly down-regulated. In addition, partial blockade of Hsp70 protein in WT cDCs using the pharmacological inhibitor DSG resulted in a diminished ability to activate T cells. This correlated with a decrease in the surface MHC-peptide complexes on WT splenic DCs after treatment with DSG. These results were further substantiated when we analyzed the ability of cDCs from Hsp70.1/3^−/− mice to present soluble OVA to naive CD4⁺ and CD8⁺ T cells. The function of such cDCs was highly impaired when compared with WT cDCs, suggesting that indeed altered levels of Hsp70 in IFN-β^−/− and IFNAR^−/− cDCs are responsible for impaired Ag presentation in the steady state.

From data presented here, one could expect that in the absence of chaperones, like Hsp70, when formation of antigenic MHC-peptide complexes is impaired, total levels of surface MHC molecules should be diminished. Nevertheless, our experiments show no differences between WT, IFN-β^−/−, and IFNAR^−/− cDCs in total surface MHC I and MHC II expression. Likewise, the analysis of the surface phenotype of Hsp70.1/3^−/− cDCs showed no significant difference in comparison to WT cDCs. Possibly, Hsp70 might be required for efficient presentation of particular peptides, like the ones used in our study, but not for others. Therefore, lack of Hsp70 might not reflect in changes of total surface MHC levels. The exact molecular reason for this phenomenon remains to be elucidated.

The discovery that basal IFNs production regulates efficiency of MHC-peptide complex formation via the expression of Hsp70 is a novel and important finding not only in the context of pathogen recognition but also in homeostasis. Not all pathogens are associated with robust IFNs production. Therefore, by sustaining Hsp70 expression, constitutively produced IFNs probably ensure that DCs are kept in a primed state for efficient presentation of Ags from such pathogens. An equally or more important function could be in the maintenance of tolerance to self-Ags. DCs that capture and present Ag under noninflammatory conditions are generally believed to acquire tolerogenic properties and generate regulatory T lymphocytes that potentiate tolerogenic responses.

We thank Dr. Andreas Krueger from Hannover Medical School for critically reading this manuscript and Regina Lesch and Susanne zur Lage for excellent technical assistance.

The authors have no financial conflict of interest.