|Home | About | Journals | Submit | Contact Us | Français|
Microbial molecules or cytokines can stimulate dendritic cell (DC) maturation, which involves DC migration to lymph nodes and enhanced presentation of Ag to launch T cell responses. Microbial Toll-like receptor (TLR) agonists are the most studied inducers of DC maturation, but type I interferon (IFN-I) also promotes DC maturation. In response to TLR stimulation, DC maturation involves a burst of Ag processing with enhanced expression of peptide-MHC-II complexes and co-stimulator molecules. Subsequently, MHC-II synthesis and expression in intracellular vacuolar compartments is inhibited, decreasing Ag processing function. This limits presentation to a cohort of Ags kinetically associated with the maturation stimulus and excludes presentation of Ags subsequently experienced by the DC. In contrast, our studies show that IFN-I enhances DC expression of MHC-II and co-stimulatory molecules without a concomitant inhibition of subsequent MHC-II synthesis and Ag processing. Expression of mRNA for MHC-II and the transcription factor CIITA is inhibited in DCs treated with TLR agonists but maintained in cells treated with IFN-I. Following stimulation with IFN-I, MHC-II expression is increased on the plasma membrane but is also maintained in intracellular vacuolar compartments, consistent with sustained Ag processing function. These findings suggest that IFN-I drives a distinctive DC maturation program that enhances Ag presentation to T cells without a shutdown of Ag processing, allowing continued sampling of Ags for presentation.
The functions of dendritic cells (DCs) at the nexus of innate and adaptive immunity (1–5) are regulated by their maturation in response to “PAMPs” (pathogen-associated molecular patterns), e.g. agonists of Toll-like receptors (TLRs) (4, 6–11). When activated by TLR agonists at a site of infection, DCs mature, upregulate expression of lymph node homing molecules, and stimulate Ag-specific T cell responses. TLR-induced DC maturation is characterized functionally by increased potency of DCs as Ag presenting cells (APCs), resulting from increased expression of peptide-class II MHC (MHC-II) complexes (3, 4, 12) and co-stimulator molecules (e.g. CD80, CD86) (13, 14). The induction of maturation produces a brief upregulation of Ag endocytosis and processing (15, 16) and a prolongation of the half-life of peptide-MHC-II complexes (3, 6) due to decreased ubiquitination, endocytosis and degradation (17–24). Decreased MHC-II ubiquitination is caused by decreased expression of MARCH1, an E3 ubiquitin ligase expressed primarily in professional APCs (20–25). This results in prolonged expression of a cohort of pathogen-associated peptide-MHC-II complexes, driving activation of pathogen-specific T cells in lymph nodes. Following the initial TLR-induced burst of Ag processing, however, Ag endocytosis decreases (26), intracellular localization of MHC-II molecules in Ag processing compartments decreases (27, 28), and the synthesis of MHC-II molecules and formation of peptide-MHC-II complexes drop dramatically (29, 30). Decreased synthesis of MHC-II is due to decreased expression of class II transactivator (CIITA), which regulates MHC-II gene transcription (6, 29, 31). This mechanism focuses DC Ag presentation on a cohort of peptide-MHC-II complexes formed at the time of exposure to PAMPs, resulting in presentation of peptides that are enriched in pathogen-derived Ag, and it prevents processing and presentation of Ags later encountered by DCs.
Although the TLR pathways, often involving myeloid differentiation primary response gene factor 88 (MyD88), are its most studied inducers, DC maturation is also driven by type I IFN (IFN-I), e.g. IFN-α subtypes or IFN-β, which signal through a single IFN-I receptor (IFN-IR). IFN-I can induce maturation of DCs characterized by upregulation of MHC-II and CD86 (32–39) and drive migration of Langerhans cells (33). IFN-I and TLR/MyD88 signaling are two semi-redundant pathways that drive DC maturation (40, 41). IFN-I is required for the enhancement of T cell responses by TLR agonists such as poly(I:C) that do not signal through MyD88 (42, 43). Regulation of DCs by IFN-I contributes to induction of MHC-I cross processing and CD8+ responses as well as the induction of CD4+ responses (44–48).
Although IFN-I is important for regulation of DCs and T cell responses, its regulation of DC MHC-II Ag processing and presentation has not been studied in depth; important differences between DC maturation induced by IFN-I vs. TLR agonists are not understood. In contrast to TLR-induced DC maturation, our studies show that IFN-I enhances DC expression of MHC-II and co-stimulatory molecules without inducing subsequent inhibition of MHC-II synthesis and Ag processing. Thus, DCs that undergo maturation in response to IFN-I may be less kinetically restricted in the acquisition and processing of Ag for presentation to T cells. This may result in increased opportunity to present Ags that are encountered at different times and locations.
C57BL/6J and C57BL/10ScnJ mice were from Jackson Laboratories (Bar Harbor, ME). MyD88−/− mice were from Shizuo Akira. All mice were housed under specific pathogen-free conditions. CpG-A oligodeoxynucleotide (ODN) 2336 (5’-ggG GAC GAC GTC GTG ggg ggG-3’) was from Sigma-Aldrich (St. Louis, MO). Synthetic triacylated lipopeptide N-palmitoyl-S-[2,3-bis(palmitoyloxy)-(2RS)-propyl]-[R]-cysteinyl-[S]-seryl-[S]-lysyl-[S]-lysyl-[S]-lysyl-[S]-lysine (Pam3CSK4) and ultrapure LPS from E. coli 0111:B4-were from Invivogen (San Diego, CA). Recombinant human IFN-β was from Peprotech (Rocky Hill, NJ). Recombinant human IFN-α2a and murine IFN-α4 and IFN-β were from PBL InterferonSource (Piscataway, NJ).
Incubations were carried out at 37°C with 5% CO2. DCs were grown in complete RPMI medium consisting of RPMI with L-glutamine and glucose, supplemented with 10% heat-inactivated FCS, 50 µM 2-mercaptoethanol, 1 mM sodium pyruvate, and 1% penicillin/streptomycin (all from Hyclone, Logan, UT). Ag processing assays were done in complete DMEM medium consisting of DMEM with L-glutamine and glucose (Hyclone), supplemented with 10% heat-inactivated FCS, 50 µM 2-mercaptoethanol, 1 mM sodium pyruvate, 10 mM HEPES (Hyclone) and 1% penicillin/streptomycin.
To make murine DCs, bone marrow cells were isolated from mouse femurs and tibias. Red blood cells were lysed with ACK lysis buffer (Lonza, Walkersville, MD). To make DCs derived from Fms-like tyrosine kinase ligand (Flt3L)-stimulated murine bone marrow (“Flt3L DCs”), marrow cells were cultured with Flt3L-Ig fusion protein (1 µg/ml) (Bioexpress, West Lebanon, NH) to produce a DC culture containing myeloid DCs (mDCs) and plasmacytoid DCs (pDCs). On day 8, non-adherent cells were collected. Alternatively, marrow cells were cultured with J558L cell-conditioned medium (containing GM-CSF) diluted in complete RPMI medium to produce “GM-CSF DCs”, which are mDCs (49). On day 7, non-adherent cells were collected.
Studies with cells from human donors were approved by the University Hospitals Case Medical Center Institutional Review Board. Blood was harvested in heparinized syringes. PBMCs were collected as reported (50). Human mDCs were purified by positive selection for CD1c (BDCA-1) with a kit (Miltenyi, Auburn, CA) as reported (51). DCs were cultured in IMDM (Lonza) supplemented with 5% pooled human serum (Gemini Bioproducts, West Sacramento, CA) and GM-CSF (Berlex, Montville, NJ, 200 U/ml).
The CD4OVA.1 and CD4OVA.2 T hybridoma cell lines were generated in these studies by incubating splenocytes from OT-II TCR transgenic mice in vitro with OVA(323–339) peptide (100 nM) for 4 d and immortalizing proliferating T cells by fusion with BW1100 cells. CD4OVA.1 and CD4OVA.2 T hybridoma cells were used to detect OVA(323–339):I-Ab complexes in Ag processing assays with murine DCs. F9A6 T hybridoma cells (52) were used to detect Ag85B(97–112):HLA-DR1 complexes presented by human DCs.
Murine DCs (2 – 4 × 105/well) were incubated with medium, Pam3CSK4, LPS or IFN-β for 48 h. Cells were washed with PBS with 0.1% BSA (Sigma-Aldrich). DCs were incubated for 15 min on ice in Fc block (anti-CD16, anti-CD32) (BD Biosciences, San Jose, CA) and stained for an additional 30 min on ice with the following antibodies: phycoerythrin-anti-I-A/I-E (BD Biosciences), fluorescein-anti-CD80 (eBioscience), phycoerythrin-anti-CD86 (eBioscience), phycoerythrin-cyanin7-anti-CD11b (eBioscience), and allophycocyanin-anti-CD11c (eBioscience). Human DCs (2–5 × 104) were incubated in polypropylene tubes for 16 – 18 h in the presence of GM-CSF (200 U/ml) with medium, IFN-α, IFN-β, or Pam3CSK4, washed, and stained with phycoerythrin-anti-CD80 (Biolegend, San Diego, CA), phycoerythrin-Texas Red-anti-HLA-DR (Invitrogen), phycoerythrin-cyanin7-anti-CCR7 (BD Biosciences), and Pacific Blue-anti-CD86 (Biolegend). Cells were washed with PBS with 0.1% BSA, fixed in 1% paraformaldehyde (Polysciences, Warrington, PA), and analyzed on a BD LSR-II flow cytometer (BD Biosciences). Data were analyzed using Winlist (Topsham, ME) or Flowjo (Tree Star, Ashland, OR). Mean fluorescence intensity (MFI) with isotype control Ab was subtracted from MFI with specific Ab to determine “specific MFI”.
After 7 d in culture, murine DCs were sorted with CD11c microbeads (Miltenyi MACS) and incubated with agonists for 24 h. DCs were washed twice in PBS with 0.1% BSA, incubated on ice for 15 min in Fc block, incubated 30 min with unconjugated anti-I-Ab or isotype control (BD Biosciences), washed 4 times in ice-cold PBS with 0.1% BSA, and resuspended in complete medium. Duplicate samples were cultured at 37°C for 0, 45 or 90 min, chilled on ice, and stained with AlexaFluor488 rabbit anti-mouse Ab (Invitrogen). Samples were washed, fixed, and analyzed by flow cytometry as above. Specific MFI was normalized to the level at time 0.
Murine DCs (5–10 × 104/well) were incubated for 20–24 h in complete DMEM medium with IFN-β, Pam3CSK4, LPS, or CpG-A ODN with or without OVA. Cells were washed and fixed with 1% paraformaldehyde or washed and incubated 4 h with OVA prior to fixation. DCs were washed, and CD4OVA T hybridoma cells (105/well) were added and incubated for 24 h. Alternatively, human mDCs (104 per well) were incubated for 16 – 18 h with Pam3CSK4 or IFN-α in the continued presence of GM-CSF (200 U/ml), washed and incubated with Mycobacterium tuberculosis Ag85B (purified as described (53)) and F9A6 T hybridoma cells (50) (105/well) for 24 h. Supernatants (100 µl) were harvested, frozen, thawed and used in a CTLL-2 bioassay for IL-2 (54). Alamar blue (15 µl, Invitrogen, Carlsbad, CA) was added for the last 24 h. Alamar blue reduction was analyzed by reading the difference between OD at 570 nm and OD at 595 nm on a Biorad Model 680 plate reader.
DCs (3 – 5 × 106/well) were incubated for 18–24 h with medium, IFN-α, IFN-β, Pam3CSK4, LPS, or CpG-A ODN. RNA was purified using the Qiashredder kit and RNEasy Plus kit (Qiagen). RNA yield was quantified by OD and 1 µg of each sample was used to synthesize cDNA using the Quantitect Reverse Transcription Kit (Qiagen); 4% of the total cDNA from each reaction was used in a quantitative PCR reaction with 500 nM of 5’ and 3’ primer for each gene and SYBR green detection (Bio-Rad, Hercules, CA) using the Bio-Rad CFX96 Real Time fluorescence detection system. All conditions were tested in triplicate. Primer sequences are as follows. GAPDH: sense 5’-AACGACCCCTTCATTGAC-3’, antisense 5’-TCCACGACATACTCAGCAC-3’. Total CIITA mRNA: sense 5’-ACGCTTTCTGGCTGGATTAGT-3’, antisense 5’-TCAACGCCAGTCTGACGAAGG-3’. The β chain of MHC-II I-Ab (I-Aβb): sense 5’-AAGATGTTGAGCGGCATCGG-3’, antisense 5’-GTCAGGAATTCGGAGCAGAG-3’. MARCH1: sense 5’-ATGCACGGACAAAGCAATGG-3’, antisense 5’-GTGTGAAGTCACGGGCAATC-3'. Primers were previously described (31, 55) or designed using Clone Manager Suite v7.11 and Primers Designers v5.11 (Scientific & Educational Software, Cary, NC). A BLAST search was performed to verify specificity.
DCs were incubated with medium, IFN-β, LPS or Pam3CSK4 for 20 h at 37°C. Subsequent steps for purification of nuclei were performed at 4° C. Cells were washed in PBS with protease inhibitors (protease inhibitor cocktail (P8340, Sigma-Aldrich) plus 1 mM NaF, 1 mM PMSF and 10 nM calyculin A), pelleted and incubated in 1 ml Nuclei EZ Prep Lysis Buffer (Sigma-Aldrich) with protease inhibitors for 10 min. After centrifugation at 500 × g for 5 min, supernatants were removed and pellets were resuspended in 1 ml Nuclei EZ Prep Lysis Buffer with protease inhibitors for 10 min to remove residual cytosolic material. Nuclei were collected by centrifugation at 500 × g for 5 min, solubilized in 10% glycerol, 2% SDS, 63 mM Tris-Cl, then sonicated and boiled. Protein concentrations were determined using the BCA assay (Thermo Scientific, Rockford, IL) to standardize loading for analysis by 12% SDS-PAGE and Western blotting.
DCs (106 per sample) were incubated for 24 h with medium, IFN-β, LPS or Pam3CSK4, washed once with PBS, and then lysed in 1% Triton X-100 with 150 mM NaCl, 25 mM Tris-Cl, pH 7.4, 25 mM NEM and protease inhibitor cocktail. Samples were centrifuged for 15 min at 14,000 × g. Supernatants were pre-cleared for 1 hr with 50 µL of rec-Protein G Sepharose 4B (Invitrogen). Immunopreciptiations were performed for 90 min at 4° C using Protein G Sepharose and 4 µg of control antibody (mouse IgG) or Y3P, specific for I-Ab (56). The beads were washed three times with ice-cold 0.1% Triton X-100 in 150 mM NaCl, 25 mM Tris-Cl, pH 7.4 and then resuspended in non-reducing sample buffer for analysis by 10% SDS-PAGE and Western blotting.
Samples were analyzed by SDS-PAGE and transferred to Immobilon membranes (Millipore, Billerica, MA). Membranes were blocked using PBS-T with 5% dry milk with the exception of anti-ubiquitin blots, which were performed using 3% BSA in PBS-T. Primary Abs were specific for actin (clone I-19, Santa Cruz Biotechnology, Santa Cruz, CA), C/EBPβ (clone C-19, Santa Cruz Biotechnology), ubiquitin (biotinylated-P4D1, Covance, Princeton, NJ), or I-Aβb (KL295, ATCC). Detection was performed using HRP-conjugated neutravidin (Invitrogen) or HRP-conjugated secondary antibodies (eBioscience) followed by ECL Western Blotting substrate (Thermo Scientific).
DCs (3 – 5 × 106/well) were incubated for 24 h with medium, IFN-α, IFN-β, Pam3CSK4, LPS, or CpG-A ODN. Cells were fixed and permeabilized using a “Fix and Perm” kit (Invitrogen). After permeabilization, cells were stained with biotinylated anti-I-A/I-E (eBioscience) for 30–60 min, washed with PBS, 0.1% BSA, and incubated with streptavidin-Alexafluor488 (Invitrogen) for 30 min. Cells were washed with PBS, 0.1% BSA and mixed with Prolong Gold mounting medium with DAPI (Invitrogen). The resulting cell suspension was placed on glass slides and covered with a poly-lysine coated coverslip (BD Biosciences). Images were captured using a Leica TCS-SP microscope. Visual counts were performed to determine the number of cells with cell-surface MHC-II expression, and at least 100 cells were counted per condition.
DCs (3 – 5 × 106/well) were incubated for 24 h with medium, IFN-α, IFN-β, Pam3CSK4, LPS, or CpG-A ODN. Cells were surface-stained for 30 min on ice with a combination of phycoerythrin-cyanin7-anti-CD11b and phycoerythrin-cyanin7-anti-CD11c (eBioscience). Cells were washed, fixed, permeabilized as for confocal microscopy, and stained for 30 min with phycoerythrin-cyanin5-anti-IA/I-E (eBioscience) and a combination of Alexafluor488-anti-LAMP1 and Alexafluor488-anti-LAMP2 (eBioscience). DAPI (10 µg/ml) was added and cells were analyzed by imaging flow cytometry on an Imagestream cytometer (Amnis, Seattle, WA). Data were analyzed using the IDEA software (Amnis).
Although TLR agonists and IFN-I stimulate distinct signaling pathways, these agents share the capacity to induce DC maturation. However, it is unclear whether mature DCs induced by these different agents differ in phenotype and function. Moreover, the Ag processing and presentation functions of DCs induced by these different agents have not been studied in depth. We confirmed that TLR agonists and IFN-β increased expression of CD80, CD86 and MHC-II on murine Flt3L DCs, although IFN-β produced a weaker induction of CD80 and CD86 (Fig. 1A–C). We also determined that IFN-β, Pam3CSK4 and LPS all enhanced the ability of DCs to present OVA(323–339) peptide to MHC-II-restricted T hybridoma cells (Fig. 1D). When DCs were exposed to OVA protein with simultaneous stimulation by Pam3CSK4 or IFN-β, we observed an increase in OVA processing and presentation to MHC-II-restricted T cells (Fig. 1E). C57BL/10J Scn mice that are deficient in TLR4 signaling were used to avoid maturation of DCs by possible LPS contamination of the OVA Ag preparations (Fig. 1E); similar results were seen with C57BL/6J DCs, but with a higher level of background DC maturation (data not shown). While the DC maturation phenotypes induced by TLR agonists and IFN-β may differ to some degree, they share the common features of increased expression of MHC-II and co-stimulator molecules as well as increased MHC-II Ag presentation.
When DC maturation is induced by TLR signaling, MHC-II Ag processing is transiently boosted; but as maturation proceeds, MHC-II Ag processing is diminished and efficient Ag presentation is focused on previously-created peptide-MHC-II complexes. Thus, after 20 h of maturation induced by Pam3CSK4 or LPS prior to the addition of antigen, we observed substantial inhibition of MHC-II Ag processing activity by Flt3L DCs (Fig. 2A). In contrast, when DCs were incubated with IFN-β for 20 h prior to the addition of antigen, active MHC-II Ag processing was maintained (Fig. 2A). Similar results were observed with several concentrations of IFN-β (10–1000 U/ml), Pam3CSK4 (1 – 100 nM), and LPS (10 – 100 ng/ml) (data not shown). We conclude that IFN-I induces a distinct pattern of DC maturation that allows continued MHC-II Ag processing, in contrast to the inhibition of MHC-II Ag processing after DC maturation induced by TLR agonists. We confirmed these results in murine GM-CSF DCs and found that incubation with Pam3CSK4 and LPS inhibited subsequent MHC-II Ag processing (e.g. after 20 h), but MHC-II Ag processing was maintained after 20 h of incubation with IFN-β (Fig. 2B). MHC-II Ag processing was also maintained in Flt3L DCs after maturation with IFN-α (Fig. 2C). CpG-A ODN activates TLR9 and drives production of IFN-I, but this TLR agonist also inhibited MHC-II Ag processing (Fig. 2C). Furthermore, the addition of IFN-β did not prevent Pam3CSK4 from inhibiting MHC-II Ag processing (Fig. 2C), indicating that inhibition by TLR agonists is a dominant effect. DCs treated with TLR agonists expressed higher surface CD80, CD86 and MHC-II (Fig. 1), indicating that the inhibition is not due to toxicity and is specific for MHC-II Ag processing. In these studies we found that MHC-II Ag processing was inhibited after maturation driven by IFN-inducing and non-IFN-inducing TLR agonists, but not after maturation induced by IFN-I alone.
To expand the relevance of our findings, we confirmed these results in human DCs. In primary human mDCs from peripheral blood, incubation with IFN-α, IFN-β and Pam3CSK4 induced CD86 (Fig. 3A), CCR7 (Fig. 3B), CD80 (data not shown), and HLA-DR (data not shown). Furthermore, human mDCs that were pre-incubated with Pam3CSK4 for 16–18 h prior to antigen exposure had decreased subsequent MHC-II Ag processing, but MHC-II Ag processing was maintained after 16–18 h of pre-incubation with IFN-α (Fig. 3C). The DCs were viable and responsive to stimulation, as indicated by upregulation of CCR7 induced by all of these agents (Fig. 3B). We conclude that IFN-I induces a distinct DC maturation program that maintains MHC-II Ag processing activity in multiple DC models, while maturation driven by a TLR agonist leads to inhibition of subsequent MHC-II Ag processing.
Since IFN-I-induced DC maturation allowed continued MHC-II Ag processing activity, in contrast to DC maturation induced by TLR agonists, we hypothesized that DC maturation states induced by these two stimuli would differ in regulation of MHC-II synthesis. Incubation of DCs with Pam3CSK4, LPS or CpG-A ODN strongly inhibited MHC-II mRNA (e.g. mRNA for the β chain of I-Ab, I-Aβb), whereas IFN-β decreased MHC-II mRNA expression only slightly (Fig. 4A,B). The difference in inhibition of MHC-II mRNA was particularly striking at the 24-hour time point (Fig. 4B), which corresponds kinetically with substantial progression to a mature DC phenotype. Addition of IFN-β did not prevent the inhibition of MHC-II mRNA by Pam3CSK4 (Fig. 4B), again indicating that the TLR-induced DC maturation response is dominant over the effects of IFN-I. Transcription of MHC-II mRNA is dependent on the transcription factor CIITA, which exists in three different isoforms (types I, III and IV; expressed from a single gene with three promoters, pI, pIII and pIV); type I is the dominant CIITA form in DCs (31). Our qRT-PCR analysis targeted a sequence common to all CIITA isoforms, yielding a measure of total CIITA expression. We found that TLR agonists Pam3CSK4, LPS, and CpG-A ODN inhibited DC expression of CIITA mRNA, whereas CIITA expression remained high after treatment with IFN-α or IFN-β (Fig. 4C). Treatment with IFN-β did not prevent CIITA mRNA inhibition by Pam3CSK4 (Fig. 4C). As a control, we found that all of these reagents induced higher expression of CD86 (data not shown), suggesting that the cells were viable and responsive to the different stimuli. We conclude that expression of CIITA and MHC-II is inhibited after TLR-induced DC maturation but not IFN-I-induced DC maturation.
Because TLR agonists exerted a dominant inhibitory effect on MHC-II mRNA expression, we tested the hypothesis that this inhibition is downstream of MyD88 signaling. Poly(I:C) activates MyD88-independent signaling through TLR3. Treatment with poly(I:C) resulted in downregulation of MHC-II, suggesting that the TRIF pathway can also mediate inhibition of MHC-II mRNA transcription (Fig. 4D). Furthermore, Pam3CSK4 and CpG-A ODN 2336 did not inhibit MHC-II synthesis in MyD88−/− DCs, while poly(I:C) did (Fig. 4D). Interestingly, LPS inhibition of MHC-II mRNA was largely reversed in MyD88−/− DCs, suggesting that LPS drives this inhibition primarily through MyD88 (Fig. 4D).
We have previously demonstrated that IFN-γ-induced CIITA expression is inhibited by TLR2 agonists in macrophages due to the induction of C/EBP transcription factors (57). To explore whether a similar mechanism may be employed during DC maturation, we treated DCs with agonists for 20 h, prepared nuclear extracts, and performed Western blots for C/EBP-β. LPS and Pam3CSK4 strongly induced the expression of both C/EBP-β isoforms, LAP and LIP, whereas IFN-β induced only low levels of C/EBP-β (Fig. 4E). Taken together, these results suggest that TLR signaling by the MyD88 and TRIF pathways in DCs induces C/EBP and thereby inhibits CIITA and MHC-II synthesis. In contrast, IFN-I induces only low levels of C/EBP-β, thereby allowing continued synthesis of CIITA and MHC-II.
Maturation of DCs by TLR signaling involves decreased MHC-II synthesis and subsequent Ag processing, but IFN-I does not inhibit MHC-II synthesis and Ag processing (Figs. 2–4). We considered whether differences in the regulation of Ag processing function by IFN-I and TLRs might involve functions other than MHC-II synthesis, e.g. endocytosis of Ag or MHC-II, or effects on the stability of peptide-MHC-II complexes. We did not observe significant differences in Ag uptake or degradation after treatment of DCs with IFN-I or TLR agonists within the time frame and conditions of these experiments (data not shown), so we focused our attention on MHC-II endocytosis and stability. MHC-II targeting for degradation is driven in part by ubiquitination of MHC-II by MARCH1, an E3 ubiquitin ligase that is expressed in immature DCs and downregulated during DC maturation in response to TLR agonists such as LPS (20, 22, 23, 25). Since endogenous MARCH1 expression is difficult to detect by Western blotting with available Abs (21), we performed qRT-PCR for MARCH1 mRNA. Relative to immature DCs (incubated in normal medium), expression of MARCH1 mRNA was inhibited in DCs that were matured with IFN-β, LPS or Pam3CSK4 for 24 h, although the degree of inhibition was less with IFN-β than with Pam3CSK4 or LPS (Fig. 5A). Coincident with the changes in MARCH1 expression, ubiquitination of MHC-II molecules was drastically reduced in DCs matured with IFN-β, Pam3CSK4 or LPS (Fig. 5B). Furthermore, MHC-II degradation products were reduced in Western blots of total cell lysates from all of the mature DCs compared to immature DCs (data not shown).
To evaluate the stability of peptide-MHC-II complexes expressed on the plasma membrane, DCs were treated with medium, IFN-β, LPS or Pam3CSK4 for 24 h, pulsed with OVA(323–339) peptide, washed, and returned to culture for 0–31 h. Using CD4OVA T hybridoma cells and IL-2 production as a readout, we observed that peptide-MHC-II complexes were highly stable in DCs after maturation with LPS or Pam3CSK4 (apparent half-life greater than 31 h), whereas these complexes declined rapidly in immature DCs or DCs matured with IFN-β (apparent half-life of less than 7 h) (Fig. 5C). In parallel with the changes in stability of peptide-MHC-II complexes, MHC-II endocytosis was inhibited by LPS but not IFN-β (Fig. 5D), suggesting that some MHC-II endocytosis proceeds independent of MHC-II ubiquitination in DCs matured with IFN-I. Taken together, these data indicate that MARCH1 expression and MHC-II ubiquitination are decreased by all of these agonists, whereas endocytosis of MHC-II and the short half-life of peptide-MHC-II complexes are maintained after treatment with IFN-β, distinct from the effects of the TLR agonists. While MARCH1 expression and MHC-II ubiquitination are associated with MHC-II turnover in immature DCs, they may not fully predict MHC-II stability and Ag presentation phenotypes in DCs matured by different stimuli. Overall, these results indicate that peptide-MHC-II complexes in IFN-I-matured DCs have a relatively short half-life, similar to that of immature DCs and distinct from the much longer half-life seen in TLR-matured DCs.
During MHC-II Ag processing, MHC-II molecules are loaded with peptides in endocytic vacuolar compartments and then translocated to the cell surface for MHC-II Ag presentation. Immature DCs express MHC-II molecules in intracellular compartments, but MHC-II expression is shifted to the cell surface following DC maturation (28). Given the different MHC-II regulation seen during DC maturation driven by TLR stimulation vs. IFN-I, we assessed whether these stimuli differ in regulation of MHC-II localization. Immature DCs (unstimulated, medium control) primarily expressed MHC-II in intracellular vacuoles, which appeared as punctate structures by fluorescence microscopy (Fig. 6A). After incubation with IFN-β (Fig. 6B,C) or IFN-α (data not shown), many DCs continued to express MHC-II in intracellular vacuoles, although more DCs expressed MHC-II at the cell surface, and some DCs expressed MHC-II simultaneously at the cell surface and in vacuolar structures (Fig. 6C). In contrast, incubation with Pam3CSK4 (Fig. 6D) or LPS (Fig. 6E) resulted in DCs that expressed MHC-II predominantly on the cell surface with few noticeable punctate structures. Quantification of the percentage of cells with punctate MHC-II staining showed that intracellular MHC-II localization was maintained more in DCs matured with IFN-I than in DCs matured with TLR agonists (Fig. 6F).
In order to further characterize the differences between populations of DCs after treatment with different stimuli, we used imaging flow cytometry to quantitatively determine colocalization of MHC-II with lysosomal markers LAMP1 and LAMP2 (LAMP1/2). Colocalization of MHC-II with LAMP1/2 was highest in untreated cells (Fig. 7), persisted at slightly lower levels in cells treated with IFN-β (Fig. 7) or IFN-α (data not shown), and was substantially decreased by Pam3CSK4 (Fig. 7), LPS (Fig. 7) or CpG-A ODN (data not shown). Interestingly, DC maturation driven by TLR agonists also resulted in a coalescence of LAMP1/2 staining (Fig. 7). These observations suggest that DCs induced to mature with TLR agonists no longer express MHC-II in the LAMP1/2-positive vacuolar MHC-II Ag processing compartments, whereas DCs treated with IFN-I continue to express MHC-II in these compartments, albeit at reduced levels in some cells.
To determine the extent of translocation of MHC-II from the vacuole to the cell surface, we assessed MHC-II colocalization with DC surface markers CD11b and CD11c (CD11b/c). Colocalization of MHC-II with CD11b/c was lowest in untreated cells (Fig. 7), increased in cells treated with IFN-β (Fig. 7) or IFN-α (data not shown), and was highest in cells treated with Pam3CSK4 (Fig. 7), LPS (Fig. 7) or CpG-A ODN (data not shown). We conclude that DC maturation by IFN-I is characterized by continued expression of intracellular MHC-II in the setting of continued synthesis and endocytosis of MHC-II, as well as higher MHC-II surface levels in parallel with decreased degradation. These unique features of MHC-II expression in IFN-I-matured DCs may allow for continued access of MHC-II to newly processed Ag in endocytic compartments in conjunction with high levels of cell surface MHC-II and Ag presentation function.
IFN-I shares with TLR agonists the ability to initiate DC maturation as defined by enhanced expression of MHC-II, CD80, CD86, and enhanced Ag presentation function in T cell assays. In our studies, however, IFN-I and TLR agonists differed in their regulation of Ag processing function subsequent to maturation due to differences in MHC-II expression and localization in DCs. Expression of CIITA and MHC-II mRNA was maintained in IFN-I-matured DCs but not TLR-matured DCs. Similarly, MHC-II Ag processing function was maintained after stimulation with IFN-I but was inhibited 20 h after TLR stimulation. Intracellular MHC-II, required for active intracellular MHC-II Ag processing, was maintained after DC maturation in response to IFN-I but not after stimulation of DCs with TLR agonists. Fig. 8 summarizes the unique phenotype of IFN-I-matured DCs in comparison to immature DCs and TLR agonist-matured DCs. Together, these data establish that DC maturation by TLR agonists produces a subsequent shutdown of MHC-II Ag processing, whereas DC maturation induced by IFN-I allows continued MHC-II Ag processing function.
DC maturation is often considered an all-or-nothing program in which DCs are either immature or become mature, but these studies and others show that different stimuli result in different DC maturation states. One issue is whether some different stimuli induce differentiation toward qualitatively different phenotypes. Another issue is whether DC maturation must be all-or-nothing (the “binary model”) or whether some stimuli (e.g. IFN-I) may induce a partially mature phenotype that reflects incomplete progression along a standard maturation pathway. This partially mature phenotype may exhibit some but not all of the features induced by agents (e.g. TLR agonists) that drive more complete DC maturation. Even if maturation is all-or-nothing at the single cell level, some stimuli may induce DC populations with an intermediate phenotype that are composed of a mixture of immature and fully mature cells. On the other hand, our results suggest that, for some stimuli, maturation is a continuum rather than a binary state, e.g. IFN-I provides a partial maturation signal for DCs that results in increased Ag processing and expression of MHC-II and co-stimulatory molecules without the subsequent shut-down that occurs with TLR stimulation. It appears that this reflects at least in part an intermediate state of maturation of some cells, since IFN-I induced DCs to express MHC-II simultaneously in vacuolar compartments and on the cell surface, distinct from both immature and TLR-matured phenotypes. Moreover, in dose response studies, IFN-I failed to induce the full TLR-driven maturation phenotype even at high concentrations at which other IFN-I-induced changes (e.g. co-stimulator expression, MHC-II expression) were at plateau, further suggesting that IFN-I and TLR agonists drive qualitatively different states of DC maturation. Alternatively, this difference in maturation phenotypes may represent a kinetic difference in the induction of maturation of DCs; MyD88 signaling may rapidly drive maturation, resulting in downregulation of MHC-II Ag processing at early time points, while IFN-I induces maturation over a longer time course. While some of these distinctions remain to be determined, we propose that IFN-I induces a state of partial maturation that is qualitatively distinct from the maturation state driven by TLR stimulation.
Young and colleagues demonstrated that pDCs and mDCs differ in the regulation of MHC-II following TLR9 stimulation (22). In some ways the maturation response of mDCs to IFN-I stimulation (this study) is similar to the maturation response of pDCs to TLR9 stimulation (22). Unlike mDCs stimulated with TLR ligands, both mDCs treated with IFN-I and pDCs treated with CpG DNA express higher levels of MHC-II at the surface without a concomitant shutdown of MHC-II synthesis (Fig. 4 and (22)) or MHC-II antigen processing (Fig. 3 and (22)). In contrast to the observations with pDCs, however, we noted a more significant reduction in MARCH1 expression and MHC-II ubiquitination following maturation of mDCs with IFN-I (Fig. 5). Thus, there are both parallels and differences between these systems.
The unique DC maturation state induced by IFN-I may be important in pathological situations that are driven primarily by IFN-I, e.g. certain viral infections, or where there is sequential exposure to IFN-I and then TLR agonists (Fig. 8). In the course of infection, some DCs may be exposed to IFN-I before they encounter PAMPs (e.g. TLR agonists) and Ags expressed by pathogens, and it may be important to avoid premature shutdown of Ag processing before DCs are exposed to pathogens. DCs generally encounter PAMPs at the time they encounter pathogen Ags, as both are contained within the pathogen, so PAMP signaling can be coordinated with production of a final cohort of peptide-MHC-II complexes that will provide presentation of pathogen Ags, followed by shutdown of Ag processing. In contrast, pathogen-induced IFN-I is a host-derived molecule that may reach and stimulate DCs that have not been exposed to pathogen Ags; if Ag processing shutdown were to occur in this scenario, it would focus Ag presentation on an antigenic repertoire missing pathogen-derived Ags, indicating the need for broader kinetic sampling relative to the time of IFN-I exposure. Accordingly, our results demonstrate that IFN-I initiates DC maturation without inhibiting Ag processing, allowing DCs to both increase MHC-II expression and continue to process Ags at sites of infection. The MyD88-driven maturation pathway is dominant over the IFN-I-induced maturation program, since we observed that simultaneous signaling by IFN-I and a MyD88-dependent TLR agonist drives shutdown of MHC-II Ag processing. Thus, if DC maturation is initiated by exposure to IFN-I, subsequent exposure to PAMPs may drive further maturation to produce a final cohort of pathogen-associated peptide-MHC-II complexes, followed by shutdown of Ag processing, focusing presentation on Ags encountered near the time of TLR stimulation.
We thank Leola Jones for technical assistance. Paul Roche kindly provided protocols and advice. J558L cells were a kind gift from Ira Mellman (Genentech), courtesy of David Gray (University of Edinburgh).
This work was supported by NIH grants AI034343, AI035726 and AI069085 to C.V.H.; AI027243 and HL055967 to W.H.B.; AI073217 and AI077056 to D.H.C; the Case Comprehensive Cancer Center (P30 CA43703) and the Case Western Reserve University Center for AIDS Research (NIH grant AI036219). D.P.S. received partial support from NIH grants HL083823 and GM007250.