IFN-β-Mediated Up-Regulation of CD1d in Bacteria-Infected APCs¹

The CD1 family of proteins comprises a third lineage of Ag presentation molecules that present lipid, glycolipid, or lipopeptide Ags to T cells. CD1d is expressed in both mice and humans, in particular in cells of the hemopoietic lineage, including B cells, T cells, macrophages, and dendritic cells (DCs)³ (1, 2, 3). CD1d molecules are structurally homologous to MHC class I proteins, having similar domain structure as well as associating with β₂-microglobulin (β₂m) (4). The Ag-binding groove of these molecules differs, however, in that the CD1d groove is narrow and highly hydrophobic. CD1d can present either microbial or self-lipids/glycolipids to a unique subset of T cells, NKT cells (1, 5). The lipid Ags are loaded onto CD1d in a number of cellular compartments, including late endosomes, lysosomes, and the MHC class II containing compartments (6). Recent studies have identified several lipid transfer proteins that may regulate this process. Microsomal triglyceride transfer protein has been shown to function as an endoplasmic reticulum chaperone by loading endogenous lipids onto nascent CD1d (7, 8), while lipid transfer proteins including saposins and GM2 activators may facilitate exchange and loading of antigenic lipids onto CD1d in these endocytic compartments (9, 10).

Most CD1d-restricted NKT cells express an invariant Vα14-Jα18 TCR (invariant NKT (iNKT)) and can be activated by a marine sponge-derived glycolipid α-galactosylceramide (α-GalCer) (5, 11, 12, 13). A similar population of iNKT cells, expressing a Vα24-Jα18 TCR, is also present in humans (14). Upon activation, iNKT cells promptly produce large amounts of IL-4 and IFN-γ, which can subsequently affect a variety of cell types and influence both innate and adaptive immune responses (15, 16, 17). Variable CD1d-restricted NKT cells, in contrast, exhibit greater TCR diversity and do not respond to α-GalCer (18, 19, 20, 21). Recent studies reveal that these two subsets of NKT cells may have distinct roles in tumor immunity and infection (22). Although it is not clear how variable CD1d-restricted NKT cells are activated during infection, it has been demonstrated that iNKT cells respond to microbial stimuli by two different mechanisms (23, 24). First, the TCR of iNKT cells can directly recognize microbial glycolipids presented by CD1d. These microbial Ags include mycobacterial phosphatidylinositol tetramannoside and Sphingomonas-derived glycolipids (25, 26, 27, 28). Second, as has been shown to be the case in Salmonella infection, the activation of NKT cells can be mediated by the recognition of self-lipid Ags presented by CD1d in combination with secondary signals. These secondary signals may include cytokines like IL-12 produced by APCs activated by microbial products as a result of TLR signaling (24, 28). This indirect mechanism allows NKT cells to respond to a wide variety of microbes. It is thus possible that changes in CD1d expression levels during inflammation may influence the extent of NKT cell responses, irrespective of whether self or microbial Ags are presented by CD1d.

Indeed, recent studies have shown that CD1d levels are altered during infections and inflammatory conditions. Oral infection with Salmonella enterica leads to increased CD1d expression on DC in vitro (29), whereas infection with Mycobacterium tuberculosis synergizes with inflammatory cytokines like IFN-γ and TNF-α, leading to increased CD1d levels on macrophages (30). Furthermore, CD1d expression is markedly elevated on hepatocytes during hepatitis C virus infection (31). In contrast, Kaposi sarcoma-associated HSV infection has been shown to cause the loss of CD1d surface expression, which may provide a strategy for viral evasion of immune response (32). Taken together, these studies support the notion that modulation of CD1d expression during infection may have functional consequences on NKT cell response. Yet, the molecular and cellular mechanisms that regulate CD1d expression during infection are not clearly defined.

Listeria monocytogenes (LM), a Gram-positive facultative intracellular bacteria, provides an excellent model for probing cell-mediated immunity to infection (33). It has been shown that iNKT cells can contribute to a Th1 response in LM infection (34), but the role of CD1d expression in this response has not been explored. In this study, we analyzed the regulation of CD1d expression during LM infection. We found that the CD1d surface expression is up-regulated on DCs and macrophages during the early phase of LM infection. This effect is not directly mediated by LM, as conditioned medium from LM-infected DCs (LM-CM) also induced uninfected APCs to express higher levels of CD1d. Neutralization of LM-CM with Abs to IFN-β significantly blocked CD1d up-regulation, suggesting IFN-β is the major mediator for CD1d induction in this model. Indeed, administration of rIFN-β enhanced CD1d surface expression on DCs in a dose-dependent manner. The IFN-β treatment did not affect the transcription or the intracellular trafficking of CD1d, but instead, resulted in CD1d protein synthesis. Furthermore, IFN-β-treated DCs are more efficient in activating iNKT than untreated controls. These data suggest that increased expression of CD1d on DCs is capable of promoting the immune response mediated by iNKT cells. Taken together, our results suggest that IFN-β secreted early during infection and the resultant increase in CD1d expression may play a role in NKT cell activation and the subsequent LM-induced immune responses.

Wild-type (WT) C57BL/6 were purchased from The Jackson Laboratory. Vα14Tg mice bred in a C57BL/6 background were provided by Dr. A. Bendelac (University of Chicago, Chicago, IL). The generation of K^b-CD1dTg mice has been described (35). All animal work was approved by the University of Chicago Institutional Animal Care and Use Committee. Bone marrow-derived DCs (BMDCs) were generated as described previously by culturing in the presence of GM-CSF and IL-4 for 6 days (35). The DC cell line DC2.4 was provided by Dr. K. Rock (University of Massachusetts, Worcester, MA) (36). All the cells were grown in RPMI 1640 medium containing 2 mM l-glutamine, 100 U/ml penicillin, 50 μg/ml streptomycin sulfate, 50 μM 2-ME, and 10% FBS (RPMI 10).

The following Abs used were purchased from BD Pharmingen: FITC-conjugated anti-hamster IgG, PE-conjugated anti-F4/80, allophycocyanin-conjugated anti-CD11c, and biotinylated-anti-K^b. Staining with biotinylated mAb was revealed using allophycocyanin- or PE-conjugated streptavidin. The CD1d-specific mAb 5C6 and generation of CD1d/α-GalCer tetramer have been described previously (3, 37). Purified 5C6 and M5114 (anti-I-A^b) were conjugated to FITC or PE and used for staining. Murine IFN-β, IFN-γ, and the neutralizing anti-IFN-β were obtained from PBL Biomedical Laboratories. Anti-TNF-α was obtained from the National Institutes of Health, whereas anti-IFN-γ was purified from culture supernatants of hybridoma (XMG1.2) using protein A-Sepharose.

The WT LM strain (1043S) and listeriolysin O-deficient (LLO⁻ LM) strain was provided by Dr. H. Shen (University of Pennsylvania, Philadelphia, PA) and were grown in brain-heart infusion broth (Difco Laboratories). Mice were infected i.v. with 5 × 10³ CFU in 100 μl of sterile PBS. The bacterial dose was verified by plating dilutions of the inoculum on brain-heart infusion agar plates. At day 1 postinfection, mice were sacrificed and peritoneal lavage, spleen, and liver cells were harvested and stained for flow cytometric analysis.

BMDCs or DC2.4 cells were infected with either WT LM or LLO⁻ LM for 1 h at a multiplicity of infection of 5:1 or 30:1, respectively. Cells were washed three times and cultured in RPMI 1640 medium containing 50 μg/ml gentamicin. After 16 h, culture supernatants and cells were collected. LM-infected or uninfected cells were washed and stained for relevant markers for flow cytometry analysis. For cytokine treatment, BMDCs or DC2.4 were incubated with indicated concentrations of recombinant IFN-β or IFN-γ. 16 h later, cells were washed and stained for flow cytometry.

For surface staining, 2 × 10⁵ cells were washed with FACS buffer (HBSS containing 2% FBS and 0.1% sodium azide) and incubated with 2.4G2 FcR blocking Ab for 5 min, followed by staining with saturating amounts of the mAbs for 45 min at 4°C. For intracellular staining, cells were fixed with 1% paraformaldehyde following surface staining for 1 h at room temperature. Cells were washed to remove fixative, followed by permeabilization in 0.3% saponin in PBS containing 10% FBS for 15 min at room temperature. Cells were then stained for 45 min at 4°C with specific mAbs in FACS buffer containing 0.3% saponin. After staining, cells were washed, resuspended in FACS buffer and examined in a FACSCalibur flow cytometer (BD Biosciences) and analyzed using FlowJo software (Tree Star).

DC2.4 cells with or without IFN-β treatment were collected. Total RNA was isolated using TRIzol reagent (Invitrogen Life Technologies) following the instructions provided by the manufacturer. Total RNA (1.0 μg) from each sample was reverse transcribed to cDNA using Superscript reverse transcriptase (Invitrogen Life Technologies). Quantitative PCR was performed with the GeneAmp 7700 Sequence Detection System using SYBR Green PCR reagents according to the manufacturer’s instruction. All PCR were done in duplicate and the level of CD1d1 expression was normalized to GAPDH using Sequence Detector software (Applied Biosystems). PCR was performed with the following primer pairs: CD1d1 forward primer, 5′-GACACCTGCCCCCTATTTGT-3′; CD1d1 reverse primer, 5′-TGGCTTCTCTTGCTTCTCTAGGTC-3′; GAPDH forward primer, 5′-TTCACCACCATGGAGAAGGC-3′; GADPH reverse primer, 5′-GGCATGGACTGTGGTCATGA-3′; H2-K^b forward primer, 5′-GGTGGCTTTTGTGATGAAGATGAGAAGGAG-3′; and H2-K^b reverse primer, 5′-GTGCAGGGACAGGGTCCTGG-3′.

IFN-β-treated or untreated DC2.4 cells were allowed to adhere to the glass slides at 37°C. Cells were fixed in 3.7% paraformaldehyde for 15 min, quenched in 10 mM glycine for an additional 15 min and then permeabilized in PBS containing 0.3% saponin. For colocalization experiments, cells were labeled sequentially with FITC-conjugated anti-CD1d (5C6) and biotinylated anti-LAMP-1 on ice for 1 h. Cells were washed in buffer containing 0.3% saponin followed by staining with Texas Red-conjugated donkey anti-mouse IgG Ab. Confocal microscopy was performed on the labeled cells with Zeiss Axiovert 200. Images were collected using OpenLab (Improvision) at the indicated wavelengths. Stacks of optical sections were obtained, deconvoluted, and examined as single sections.

The rate of internalization and recycling of CD1d were measured using flow cytometry based assays (38). IFN-β-treated or untreated DC2.4 cells were incubated with 1 μg/ml 5C6-PE at 37°C. At the indicated times, aliquots were removed and divided into two parts. One part was left untreated on ice, while the other was incubated at 4°C for 45 s in PBS acidified to pH 2.0 with HCl and supplemented with 0.03 M sucrose and 10% FBS. Subsequently, samples were washed in a large excess of RPMI 1640 supplemented with 10% FBS and 100 mM HEPES buffer and analyzed by flow cytometry. Untreated samples account for total cell-associated fluorescence, while acid-stripped aliquots account for fluorescence in acid-resistant compartments. Results are expressed as the percent of internalization which is the ratio of acid-resistant (internal) to total PE fluorescence. For recycling, cells were incubated at 37°C with 5C6-PE for 40 min and acid stripped to remove all surface-bound Ab. Cells were resuspended in 37°C prewarmed culture medium in the presence or absence of IFN-β and incubated at 37°C for the indicated time periods. Aliquots were either left untreated or acid stripped at the time points and results were expressed according to the equation: percentage of recycled = 1 − (acid resistant fluorescence/total cell-associated fluorescence).

DC2.4 cells were incubated in the presence or absence of IFN-β for 16 h. After starving for 2 h in methionine/cysteine-free medium, cells were labeled with 0.4 mCi of [³⁵S]translabel (MP Biomedicals) for 2 h. After washes, cells were chased in the presence of excess methionine and cysteine (1 mM) in complete RPMI 1640 medium for the indicated time periods. Cells were harvested and lysed in TBS containing 1% Nonidet P-40 and immunoprecipitated with anti-CD1d, anti-MHC class I, or control Ab-conjugated protein-A beads. Immunoprecipitates were separated by 12% SDS-PAGE and visualized by autoradiography.

Day 6 BMDCs were cultured in the presence or absence of 1000 U/ml IFN-β for 16 h and used as stimulators. After washes, these stimulators were seeded (1 × 10⁵/well) in U-bottom microtiter plates and incubated with anti-CD1d (5 μg) or a control Ab for 1 h at 37°C. A total of 1 × 10⁵ hepatic lymphocytes isolated from Vα14Tg mice were then added to the cultures. After 36 h, the culture supernatants were harvested and the levels of IL-4 and IFN-γ were quantitated by sandwich ELISA (BD Pharmingen).

Mean values were compared using the unpaired Student’s t test. All statistical analyses were performed with the Prism program (GraphPad). Statistically significant differences p < 0.05 and p < 0.01 are noted with ∗ and ∗∗, respectively.

iNKT cells can become activated as early as 1 day after LM infection as measured by CD69 up-regulation (Fig. 1,A). It is possible that changes in CD1d expression levels may play a role in this response. To see whether LM infection indeed leads to CD1d up-regulation during early phase of infection, we examined CD1d expression levels in peritoneal lavage, where significant numbers of macrophages and DCs reside. As with iNKT cell activation, on day 1 after LM infection, CD1d up-regulation (3- to 4-fold) was readily detectable on both macrophages (F4/80⁺) and DCs (CD11c⁺) (Fig. 1 B). To address the kinetics of CD1d induction during LM infection, we also examined CD1d expression on DCs and macrophages at different time points postinfection. We observed 5- to 6-fold higher levels of CD1d on DCs and macrophages at 3 days postinfection consistent with a previous report (39). These increased levels were sustained for up to 7 days postinfection. Interestingly, the level of CD1d expression on B and T cells is not altered in LM-infected mice (data not shown). These results indicate that LM infection results in preferential induction of CD1d on APCs early in infection.

To examine whether up-regulation of CD1d surface expression is a direct result of LM infection of APCs, we performed in vitro LM infection studies on BMDC and analyzed CD1d expression by flow cytometry. We found that infection of BMDCs with LM results in a ∼3-fold increase of CD1d surface expression (Fig. 2,A). To rule out the possibility that CD1d up-regulation on BMDCs could occur indirectly through the effect of LM on some minor cell populations present in DC cultures, we performed similar in vitro infection studies using a DC line, DC2.4. As shown in Fig. 2,A, surface CD1d expression is increased in the DC2.4 cell line 16 h after LM infection, indicating that infection of DCs with LM is sufficient to induce CD1d up-regulation. As both live and heat-killed LM (HKLM) can be internalized by DCs and up-regulate the expression of costimulatory molecules and MHC class II on DCs, we also examined the effect of HKLM on CD1d up-regulation on DCs. Unlike LM infection, incubating BMDCs or DC2.4 cells with HKLM does not lead to increased CD1d surface expression, suggesting that up-regulation of CD1d on LM-infected DCs is not due to the effects of microbial Ags on DC maturation (Fig. 2 B).

To explore the possibility that a soluble factor secreted by LM-infected DCs is responsible for up-regulation of CD1d, we tested whether LM-CM could induce CD1d up-regulation. Filtered LM-CM was fully able to induce up-regulation of CD1d on BMDCs and DC2.4 cells as compared with direct LM infection (Fig. 2,A). Supernatants derived from bacterial cultures without DCs or from HKLM-treated DCs had little capacity to induce CD1d up-regulation (Fig. 2 B), suggesting that the molecules that induce CD1d up-regulation were not produced by bacteria, but were secreted by DCs after LM infection.

To identify the secreted molecules produced by LM-infected DCs that mediated CD1d up-regulation, we looked for differential production of proinflammatory cytokines by treated DCs and examined the effect of these cytokines using specific neutralizing Abs. Consistent with a previous report (40), we detected comparable levels of IL-12 and TNF-α in conditioned medium from LM-infected DCs and HKLM-treated DCs (data not shown). In addition, no detectable amount of IFN-γ was produced by DCs in response to HKLM treatment and LM infection. Because the production of IL-12, TNF-α, and IFN-γ was similar after HKLM treatment or LM infection, they are unlikely to be responsible for the differences in CD1d up-regulation observed. Indeed, neutralizing Abs against TNF-α, IL-12, and IFN-γ had very little or no effect on the activity of LM-CM in up-regulating the CD1d expression.

Recent studies have shown that infection with live but not HKLM induces DCs to secrete large amounts of IFN-β (40). To determine whether IFN-β is responsible for LM-CM mediated CD1d up-regulation, Abs to IFN-β were used in blocking experiments. As shown in Fig. 3, neutralization of LM-CM with Abs to IFN-β significantly blocked CD1d up-regulation. We also infected DC2.4 cells with LLO⁻ LM, which lacks the ability to escape into the cytosol and is unable to induce IFN-β in infected DCs (41). We found that supernatants derived from LLO⁻ LM-infected DCs were unable to induce the up-regulation of CD1d, which further supports the role of IFN-β in mediating the activity of LM-CM to induce CD1d up-regulation (Fig. 2 B).

To verify the direct effect of IFN-β on surface expression of CD1d, DC2.4 cells were treated with titrating amounts of rIFN-β for 16 h and CD1d surface expression was analyzed by flow cytometry. Fig. 4,A shows a dose-dependent increase in CD1d expression on DC2.4 cells. The dose-response curve of IFN-β plateaued at 1000 U/ml and resulted in an ∼3-fold increase of CD1d surface expression on DC2.4 cells, comparable to that induced by LM-CM. The induction of CD1d was observed as early as 2 h after IFN-β treatment and reached a plateau after 16 h of treatment, consistent with early in vivo up-regulation (Fig. 4,A). Increased CD1d surface expression can also be detected in IFN-β-treated BMDCs, peritoneal macrophages, and a macrophage cell line, P388 (Fig. 4,C). Consistent with previous reports, the surface expression of MHC class Ia, H2-K^b, is also up-regulated in IFN-β-treated DCs (data not shown) (42). As IFN-γ has been shown to regulate CD1d expression on macrophages during mycobacteria infection, we compared the effect of IFN-β and IFN-γ on the surface expression of CD1d in BMDCs, DC2.4, and P388 cells. As shown in Fig. 4 C, the extent of CD1d up-regulation induced by IFN-γ treatment is much lower than that by IFN-β in the time frame (16 h) examined.

To explore the mechanisms underlying the increased surface expression of CD1d by IFN-β, we first examined whether the increase in surface CD1d expression is due to an increase in total protein levels of CD1d. To measure the steady state total CD1d levels, we performed immunofluorescence staining on permeabilized DC2.4 cells, followed by flow cytometric analysis. As shown in Fig. 5 A, total CD1d levels are increased in the IFN-β-treated cells as compared with the untreated cells. The extent of CD1d up-regulation is comparable to changes observed at the cell surface.

To test whether the increase in protein level of CD1d following IFN-β treatment is associated with an increase in RNA level, real-time RT-PCR was performed using primers specific for CD1d1. As shown in Fig. 5,B, CD1d message levels relative to a housekeeping gene (GAPDH) did not change after 16 h of IFN-β treatment. In contrast, the level of H2-K^b mRNA was increased in IFN-β-treated cells compared with untreated cells (Fig. 5 B) These results indicate that increased expression of CD1d by IFN-β is not regulated at the RNA level, unlike MHC class Ia.

To examine whether IFN-β treatment can affect the intracellular trafficking of CD1d, we examined the distribution of CD1d using confocal microscopy. Compared with untreated cells, IFN-β-treated DC2.4 cells showed significantly brighter overall staining with anti-CD1d mAb, consistent with the notion that total pool of CD1d protein is increased in IFN-β-treated cells (Fig. 6). However, the intracellular distribution of CD1d in IFN-β-treated cells was similar to that in untreated cells, with a substantial amount of the intracellular CD1d colocalized with LAMP-1, a lysosomal marker (Fig. 6,A). These results suggest that increased surface expression of CD1d by IFN-β cannot be attributed to redistribution of the intracellular pool of CD1d. Furthermore, IFN-β does not affect the rate of internalization and recycling of CD1d as measured by fluorescence-based assays (Fig. 6, B and C).

To monitor the effect of IFN-β on the biosynthesis and stability of CD1d, we performed pulse-chase experiments. DC2.4 cells were incubated with medium alone or with IFN-β for 12 h, followed by pulse labeling for 2 h with [³⁵S]translabel and chased for 0, 6, or 16 h in the presence or absence of IFN-β. After cell lysis, total CD1d protein was immunoprecipitated with anti-CD1d Ab and analyzed by SDS-PAGE. Compared with untreated cells, the amount of newly synthesized CD1d was considerably greater in the IFN-β-treated samples (t = 0), suggesting that IFN-β enhanced the biosynthesis of CD1d protein. Comparing the ratio of CD1d between IFN-β-treated and untreated cells at different times of chase revealed no clear changes in the half-life of CD1d in the IFN-β-treated cells. Furthermore, we observed no significant difference in the acquisition of Endo H resistance between IFN-β-treated and untreated samples (data not shown), confirming our confocal studies which showed that IFN-β has no obvious effect on intracellular distribution of CD1d molecules. Thus, treatment of DC with IFN-β appears to enhance the de novo synthesis of CD1d, but does not have a significant effect on the stability and intracellular transport of CD1d during the time period that we examined. For comparison, we also examined the effect of IFN-β on the metabolism of H2-K^b. As shown in Fig. 7, IFN-β treatment also increased biosynthesis of H2-K^b.

To test whether IFN-β mediated increase in surface CD1d expression is associated with an altered ability of DCs to stimulate iNKT cells, IFN-β-treated and untreated DCs were cultured with iNKT cells freshly isolated from Vα14Tg mice and cytokine production was measured after 48 h. We found that iNKT cells produce significantly higher amounts of IL-4 and IFN-γ in response to stimulation by IFN-β-treated DCs, while no noticeable amounts of IFN-γ and IL-4 can be detected from iNKT cells stimulated with untreated DCs (Fig. 8). The cytokine response to IFN-β-treated DCs can be blocked by anti-CD1d, indicating this activation process is CD1d dependent. Although IFN-β has been shown to induce the expression of some costimulatory molecules, we did not detect significant changes on the expression of CD80, CD40, and CD86 during the time frame (16 h) used to induce CD1d up-regulation (data not shown). Thus, the observed difference in cytokine production in response to IFN-β-treated DCs compared with untreated DCs is likely due, at least in part, to the increase in surface CD1d expression seen in IFN-β-treated DCs.

To directly address the question of whether increased CD1d surface expression is sufficient to lead to iNKT cell activation, we used BMDCs from K^b-CD1dTg mice, which express CD1d under the control of the H2-K^b promoter. The cell surface level of CD1d on K^b-CD1dTg⁺ BMDCs is 5- to 6-fold higher that that of WT BMDCs, while the expression of other DC maturation markers is similar (data not shown). As shown in Fig. 8, iNKT cells stimulated with K^b-CD1dTg⁺ BMDCs produced even greater amounts of IFN-γ and IL-4 than IFN-β-treated DCs, possibly due to their higher CD1d expression levels. These data demonstrate that increased CD1d surface expression on DCs is associated with enhanced cytokine secretion by iNKT cells, suggesting that up-regulation of CD1d expression during infection or inflammatory conditions may play a role in regulating the function of iNKT cells.

Up-regulation of MHC class I expression during infection has been shown to be largely mediated by cytokines, in particular IFNs and TNF-α (43, 44). Increased levels of MHC class I can lead to more effective recognition of infected cells by CD8⁺ T cells and promote cytolytic responses against intracellular pathogens. Recent studies have shown that iNKT cells are activated early during infection and may participate in host defense against a wide range of pathogens (1, 22, 30). However, little is known about how expression of CD1d is regulated during infection. In this study, we demonstrated that LM infection induced CD1d up-regulation on DCs and macrophages both in vivo and in vitro. IFN-β secreted by LM-infected APCs was the major factor responsible for CD1d up-regulation during LM infection. Neutralizing Abs against IFN-β could abrogate the effect of LM-mediated CD1d up-regulation; rIFN-β treatment enhanced CD1d expression to the levels similar to those induced by LM infection. This is, to our knowledge, the first study to demonstrate that IFN-β mediates CD1d expression in response to an infectious agent.

IFN-β treatment augments the biosynthesis of CD1d protein, but has no significant effect on the steady state levels of CD1d mRNA, suggesting a posttranscriptional regulatory mechanism that is distinct from the regulation of MHC class I. Our finding that IFN-β-treated DCs and K^b-CD1dTg⁺ DCs, but not untreated WT DCs, induced significant cytokine production from iNKT cells in the absence of exogenous Ags suggests that up-regulated CD1d levels on DCs are sufficient to activate NKT cells in vitro. The preferential induction of CD1d expression on DCs and macrophages in LM-infected mice is noteworthy. CD1d up-regulation can be detected as early as 24 h after LM infection, concomitant with the activation of iNKT cells, suggesting increased CD1d expression during infection may have functional consequences.

Cytokines and microbial signals have been implicated in CD1d induction during infection. TNF-α has been shown to be necessary, but not sufficient, for CD1d induction in coxsackievirus B3-infected endothelial cells (45). A recent study on M. tuberculosis infection demonstrated a synergistic effect of IFN-γ and TNF-α on CD1d induction on macrophages (30). Our Ab-blocking experiment, however, suggests that IFN-γ and TNF-α do not contribute significantly to CD1d up-regulation in the LM infection model. This result is further substantiated by the findings that LM infection of IFN-γ knockout (KO) BMDCs and WT DCs resulted in similar levels of CD1d up-regulation (data not shown). Furthermore, treatment of TNFRKO BMDCs with conditioned medium derived from LM-infected cells resulted in increased CD1d levels, comparable to WT BMDCs (data not shown). Thus, it is possible that different pathogens may use distinct mechanisms to modulate the expression of CD1d.

IFN-β, a type I IFN, is rapidly induced in response to viral infection but can also be produced upon exposure to nonviral pathogens, such as LM (40), M. tuberculosis (46), Chlamydia trachomatis (47), and Schistosoma mansoni (48). The specific induction of CD1d on APCs but not T cells by IFN-β strongly supports a role for NKT cells in infectious response. Thus, it is conceivable that LM-induced CD1d up-regulation by IFN-β can be extended to other pathogens, implying a broad role for CD1d up-regulation in early infectious response.

In addition, early IFN-β production has been shown to promote the subsequent expression of additional type I IFN genes during infection through a positive feedback mechanism (49, 50). The mouse type I IFN locus contains 14 IFN-α genes and single IFN-β, IFN-κ, and IFN-ε genes, which have different binding affinities to the common IFN-α/IFN-β receptor, perhaps leading to differential signaling and gene activation (51, 52). It is thus likely that these other type I IFN subtypes may be involved in CD1d regulation as well.

Binding of IFN-β to IFN-α/IFN-β receptor leads to activation of JAK-STAT signaling pathways and the formation of the heterotrimeric transcription factor complex, IFN-stimulated gene factor 3 (ISGF3), which binds to and transactivates genes containing IFN-stimulated response element (53). IFN-β can also activate the formation of STAT-1 homodimer, which binds to a γ-activated site in the promoters of IFN-responsive genes to mediate an overlapping set of response with IFN-γ. Our finding that rIFN-β has a more profound effect on the up-regulation of CD1d expression than IFN-γ suggests that ISGF3 is involved in this process. An analysis of CD1d1 sequence shows that no IFN-stimulated response element is present in the proximal promoter region of CD1d1 (54), however, suggesting CD1d1 is not a direct target of ISGF3. This may explain the absence of transcriptional changes in CD1d after IFN-β treatment in the cell lines examined. ISGF3 target genes may instead mediate posttranscriptional regulation of CD1d expression.

CD1d, like MHC class II molecules, survey Ags in the endocytic pathway. It has been reported that the MHC class II-invariant chain complex interacts with CD1d and affects the intracellular distribution of CD1d (9, 55). Thus, it has been postulated that the expression and intracellular trafficking of CD1d may be altered by MHC class II molecules and invariant chain induced during inflammation. Although IFN-γ can induce the expression of MHC class II, type I IFNs have not been shown to be involved in the up-regulation of MHC class II (56, 57). Indeed, we did not detect significant difference in MHC class II expression in IFN-β-treated BMDCs (data not shown). Furthermore, the degree of CD1d up-regulation by IFN-β is similar between WT and MHC class II-deficient DCs (data not shown), indicating that the observed effect of IFN-β on CD1d expression is independent of the expression and intracellular trafficking of MHC class II.

Modulation of surface CD1d expression levels has been shown to have a direct effect on the numbers, activation, and function of CD1d-restricted iNKT cells. For example, TGF-β synthesized by human keratinocytes has been shown to have an inhibitory role on the proliferation and differentiation of CD1d-restricted NKT cells (58). This effect correlates to decreased expression of human CD1d on DCs being mediated by TGF-β. Biliary CD1d up-regulation and increased numbers of hepatic iNKT cells have been observed in primary biliary cirrhosis (59). In addition, up-regulation of CD1d on macrophages during M. tuberculosis infection is able to induce NKT cell activation both in vivo and in vitro (30). Therefore, although we cannot exclude the possibility of presentation of specific listerial Ags to NKT cells, IFN-β-mediated increase in CD1d levels on DCs may indeed be one possible mechanism to activate iNKT cells during infection.

Besides being promptly induced in response to microbial pathogens, IFN-β is widely used for clinical treatment of several human pathologies such as infectious diseases, multiple sclerosis, or tumors (60, 61). NKT cells have been shown to have strong functional roles in either pathology or amelioration of disease (22). Up-regulation of CD1d expression by IFN-β may thus inadvertently strengthen or impair immune responses in these patients. Understanding the regulation of CD1d expression as well as the functional consequence of this regulation on iNKT cells by IFN-β may therefore lead to a better understanding of the current role of IFN-β in the clinical setting as well as potential extension of this treatment to other diseases.

We thank Drs. Angela Colmone, Mike Zimmer, Honglin Xu, and Hak-Jong Choi for critical reading of the manuscript, Dr. Yan-Xin Fu (University of Chicago) for providing TNFRKO mice, and Shirley Bond for advice on confocal microscopy.

The authors have no financial conflict of interest.