During their final differentiation or maturation, dendritic cells (DCs) redistribute their major histocompatibility complex (MHC) class II products from intracellular compartments to the plasma membrane. Using cells arrested in the immature state, we now find that DCs also regulate the initial intracellular formation of immunogenic MHC class II–peptide complexes. Immature DCs internalize the protein antigen, hen egg lysozyme (HEL), into late endosomes and lysosomes rich in MHC class II molecules. There, despite extensive colocalization of HEL protein and MHC class II products, MHC class II–peptide complexes do not form unless the DCs are exposed to inflammatory mediators such as tumor necrosis factor α, CD40 ligand, or lipoplolysaccharide. The control of T cell receptor (TCR) ligand formation was observed using the C4H3 monoclonal antibody to detect MHC class II–HEL peptide complexes by flow cytometry and confocal microscopy, and with HEL-specific 3A9 transgenic T cells to detect downregulation of the TCR upon MHC–peptide encounter. Even the binding of preprocessed HEL peptide to MHC class II is blocked in immature DCs, including the formation of C4H3 epitope in MHC class II compartments, suggesting an arrest to antigen presentation at the peptide-loading step, rather than an enhanced degradation of MHC class II–peptide complexes at the cell surface, as described in previous work. Therefore, the capacity of late endosomes and lysosomes to produce MHC class II–peptide complexes can be strictly controlled during DC differentiation, helping to coordinate antigen acquisition and inflammatory stimuli with formation of TCR ligands. The increased ability of maturing DCs to load MHC class II molecules with antigenic cargo contributes to the >100-fold enhancement of the subsequent primary immune response observed when immature and mature DCs are compared as immune adjuvants in culture and in mice.
Dendritic cells (DCs), perhaps the most potent and versatile of APCs, capture foreign antigens encountered in peripheral tissues, process the antigens into peptides bound to MHC molecules, and migrate to lymphoid organs for presentation of the MHC–peptide complexes to T lymphocytes 1,2. In keeping with this critical role in stimulating antigenically naive T cells, the immunostimulatory activity of DCs is carefully regulated, with the cells exhibiting at least two functionally distinct states. Resting or immature DCs, corresponding to those found in the periphery, generally have a high capacity for endocytosis and thus for antigen uptake 3,4,5,6, but a low capacity for binding and stimulating T cells. Mature DCs, on the other hand, are generated from immature cells by exposure to proinflammatory agents (TNF, LPS), physical trauma, or T cell surface molecules (CD40 ligand [CD40L]; references 7–10). Mature DCs exhibit a limited capacity for endocytosis, but they are exceptional at stimulating T cells 3,4,5,6,7,8,9,10, typically presenting immunogenic peptides derived from previously encountered antigens. The capacity of mature DCs to activate naive T cells reflects several developmentally regulated changes. These include upregulation of CD40, CD80, and CD86 costimulatory molecules 11,12, release of cytokines like IL-12 13,14, resistance to IL-10–mediated immunosuppression 15, and expression of new chemokine receptors like CCR7 that guide DCs to the T cell areas 16,17,18,19. Importantly, mature DCs upregulate surface MHC class II and functional MHC class II–peptide complexes that engage the TCR 20,21.
This increased expression of MHC class II molecules is known to reflect at least two alterations 20,21. First, newly synthesized MHC class II molecules become rerouted intracellularly. Instead of being transported to lysosomes for sequestration or degradation, the MHC class II is targeted to the plasma membrane. One underlying mechanism involves the activation of the APC-restricted protease cathepsin S, which in turn mediates the post–Golgi complex cleavage of the MHC class II–associated invariant (Ii) chain. These events facilitate the early release of Ii chain from MHC class II α/β dimers, allowing surface transport rather than intracellular retention due to lysosomal targeting/endocytosis signals found in the Ii chain cytoplasmic domain 22. The cleavage of Ii could also free the MHC class II binding site for antigen binding 23,24, but this level of regulation has not yet been demonstrated in maturing DCs. Second, because of the downregulation of endocytosis that occurs concomitant with maturation, there is an increase in the surface residence time, half-life, and thus overall expression of any MHC class II–peptide complexes that do reach the plasma membrane 20,21.
To further delineate the mechanism underlying the critical conversion of DCs from immature sentinels to mature immunostimulatory cells, we made use of a monoclonal antibody, C4H3, that allows the direct visualization of a peptide–MHC class II complex formed from a hen egg lysozyme (HEL) peptide and an I-Ak α/β dimer 25,26. We also identified conditions that allow the in vitro maintenance and rapid manipulation of immature bone marrow–derived DCs. Together, these approaches revealed an additional and more fundamental level at which the development of antigen presentation capacity is controlled during DC maturation, the actual formation of MHC class II–peptide complexes within MHC class II–rich lysosomal compartments. We show that this maturation is a major control for T cell priming by DCs in vivo, a new finding that impacts upon their optimal use for immunotherapy in humans 27,28,29,30.
Materials And Methods
CBA/J and lipid A–unresponsive C3H/HeJ female mice were purchased from The Jackson Laboratory or from Japan SLC, and were used at 6–8 wk of age.
DCs were grown in RPMI 1640 containing 5% FCS from bone marrow progenitors with mouse rGM-CSF 31 at 10 ng/ml (gift of Dr. Y. Yamaguchi, Kirin, Maebashi, Japan), or the supernatant (3% vol/vol) from J558L cells transduced with murine (m)GM-CSF (from Dr. A. Lanzavecchia, Basel Institute, Basel, Switzerland). At day 6, the cultures contained many aggregates of immature DCs loosely attached to the monolayer. To mature the DCs, we added either TNF (100–500 U/ml; Dainippon Pharmaceutical Co.), CD40L (a baculovirus preparation provided by Drs. M. Nussenzweig and Y. Choi, The Rockefeller University, New York, NY), or LPS (1–10 ng/ml, Escherichia coli type 0111.B4; Sigma Chemical Co.), or we simply transferred the cells to a fresh vessel at <5 × 106 cells/ml 31.
HEL (Sigma Chemical Co.) was added to immature bone marrow DCs at 30–3,000 μg/ml for 0.5–24 h. Also, explants of ear skin (epidermis and dermis) were bathed in 3,000 μg/ml, after which the DCs were examined within epidermal sheets or as cells that had emigrated from the explants 32,33. The dominant HEL 46-61 peptide for I-Ak was synthesized at Yale University Medical School. OVA was used as control protein. HEL protein uptake was visualized with 1B12 monoclonal IgG2b anti-HEL antibody, provided by Dr. P. Allen (Washington University, St. Louis, MO), and MHC class II–HEL peptide complexes visualized with C4H3 rat IgG2b monoclonal antibody 25,26. Antibody staining, including that with isotype controls (PharMingen), was assessed on a FACScan™ (Becton Dickinson) or by immunofluorescence confocal microscopy. DCs were identified by labeling for the I-E MHC II product (14-4-4S antibody), and mature DCs by high expression of CD86 (GL-1; PharMingen).
Removal of Endotoxin Activity.
To deplete much of the endotoxin activity (limulus amebocyte assay; BioWhittaker) in HEL preparations, HEL at 10 mg/ml was adsorbed with tachyplesin III–conjugated Sepharose CL (34; Kuttsuclean™; Maruha Corp.) according to the manufacturer's instructions.
Antigen Presentation Assays.
Presentation of HEL to T cells was monitored using purified CD4+ T cells from 3A9 TCR transgenic mice 35 provided by Dr. M. Davis (Stanford University, Palo Alto, CA). These T cells are specific for the same MHC class II–peptide complex recognized by the C4H3 antibody. Graded doses of DCs that were exposed to HEL minus or plus a maturation stimulus were applied to 250,000 CD4+ transgenic T cells in 96-well flat-bottomed microtest plates in RPMI 1640 containing 5% FCS. The DCs were fixed beforehand in 0.75% paraformaldehyde for 30 min on ice. CD4+ T cells were enriched by negative selection from spleen and lymph node suspensions by coating other cells with antibodies (TIB 120 anti–MHC class II, TIB 207 anti-CD8, HB198 F4/80 anti-macrophage, 6B2 anti-B220, and NK1.1) and depleting them with sheep anti–rat Ig Dynabeads® M-450 (No. 110.08; Dynal). T cell responses were monitored at 5 h by a decrease in TCR (Vβ8) or increase in CD69 (PharMingen antibodies), or at 30–42 h by [3H]thymidine (3H-TdR) uptake at 1 μCi/ml. Data are from triplicate cultures with SE <10% of the mean. For presentation studies in vivo with adoptively transferred DCs, immature cells were cultured with graded doses of HEL overnight with or without CD40L or LPS as a maturation stimulus. The DCs were harvested, washed, and injected subcutaneously at a dose of 200,000 DCs per paw of nontransgenic mice. 5 d later, the draining lymph nodes were removed, dissociated into single cell suspensions, and cultured at 300,000 cells per flat-bottomed microtest well in Click's medium with 0.75% mouse serum and graded doses of HEL. 3H-TdR uptake was measured at 52–64 h to document the extent of CD4+ T cell priming.
Synergistic Effects between the Exposure to Antigen and a Maturation Stimulus in the Formation of MHC Class II–Peptide Complexes.
Immature DCs from day 6 GM-CSF–stimulated marrow cultures 31 were exposed for various periods of time to increasing concentrations of HEL. The formation of the C4H3 epitope, i.e., MHC class II–HEL peptide complexes, was monitored by FACS®. Strong signals were not seen at the DC surface after a 3-h exposure to HEL, but were present after 18–24 h (pulse data; Fig. 1 A). The formation of C4H3 was proportional to the added dose of HEL; signals became detectable at 30 μg/ml (not shown). Immature DCs only had to be pulsed a short time with HEL to develop strong C4H3 signals 1 d later (pulse–chase data; Fig. 1 A). For example, a 3-h pulse of 3,000 μg/ml HEL followed by an 18-h chase in its absence gave comparable C4H3 staining to continuous culture in HEL.
However, the addition of HEL to immature DCs not only led to the appearance of C4H3 epitope, but also to maturation. Most MHC class II–positive DCs began to express high levels of CD40 and CD86 and higher levels of surface MHC class II compared with cultures not incubated with HEL (Fig. 1 B). Although initially surprising, HEL-induced maturation could be explained by the fact that 1 mg/ml HEL contained high endotoxin activity, 1–10 ng/ml in the limulus endotoxin assay, as reported 36. Likewise, simple transfer of the immature DCs to a fresh culture dish, a procedure that is used to enhance the expansion of residual proliferating cells in the culture, enhanced maturation (Fig. 1 B; references 21 and 31). Therefore, in all subsequent experiments, a new invertebrate adsorbent termed Kuttsuclean™ was used to remove most endotoxin 34 without decreasing HEL content, and we compared immature and maturing cells in cultures that were not subcultured from their initial wells.
The addition of Kuttsuclean™-treated HEL did not induce DC maturation, as indicated by a lack of generation of CD86+ cells or the increased expression of surface MHC class II (I-E and I-A). Importantly, only small amounts of C4H3 epitope appeared at the plasma membrane (Fig. 2 A). DCs incubated under these conditions could be kept in the immature state for at least 3–4 d. In prior work 20,21, immature DCs exhibited a slow but spontaneous rate of maturation, and some formation of the SDS stable dimers suggestive of MHC–peptide complex formation. The maturation stimuli (LPS, TNF, CD40L, and transfer to a new culture vessel) induced an increase in CD86+ cells, and some increase in background C4H3 staining (Fig. 2 A), presumably because the higher levels of I-A on mature DCs presented self-peptides that cross-reacted with HEL peptide 25,26. Nevertheless, the combination of HEL and a maturation stimulus was markedly synergistic, with very high C4H3 levels being observed on most DCs in >10 similar experiments (Fig. 2 A).
Mechanisms Underlying the Synergism of Antigen and a Maturation Stimulus in Controlling MHC Class II–Peptide Complex Formation.
To establish that the maturation stimulus was not merely increasing uptake of HEL, we applied the antigen as a pulse for 3 h, and then washed the cells before the addition of LPS or CD40L. The synergistic effect of the maturation stimulus remained apparent in all three experiments (Fig. 2 B), indicating that maturation functioned to enhance HEL processing and/or formation of C4H3-reactive MHC class II–HEL peptide complexes. In a companion study 37, we found that the HEL-pulsed DCs can be cultured for at least 48 h before adding a maturation stimulus, but the cells still display C4H3 at levels >50% of that seen when the addition of LPS or CD40L is not delayed. Therefore, maturation stimuli increase the formation of TCR ligands, and do not simply reduce their degradation.
To determine if maturation was acting only at the level of antigen processing/proteolysis, we pulsed DCs for 3 h with preprocessed HEL peptide (that is able to bind directly to I-Ak), washed the cells, and then cultured without or with a maturation stimulus. The latter was again required for MHC–peptide complex formation (Fig. 2 C), indicating that peptide access to the MHC class II binding site was controlled by the maturation process.
Conceivably, maturation might control C4H3 epitope expression by controlling the delivery of HEL to the MHC class II compartments (MIICs) that are thought to be critical for MHC–peptide complex formation. Therefore, we performed immunofluorescence confocal microscopy to localize C4H3 in the absence or presence of a maturation stimulus. As shown in Fig. 3 (left), immature DCs accumulated endotoxin-free HEL protein in intracellular compartments, as detected with 1B12 antibody to intact HEL. These structures were positive for MHC class II, Ii chain, and the peptide editing H-2M product, and were judged to be late endosomes and lysosomes by staining for the lysosomal membrane glycoprotein, lgp-B (or lysosomal-associated membrane protein [LAMP]-2; Fig. 3). Despite the fact that HEL was delivered to MIICs, no C4H3 staining was observed even after incubations of up to 24 h in HEL (Fig. 3) or preprocessed HEL peptide (not shown). Also, no C4H3 staining was found when only a maturation stimulus was given (LPS or CD40L; not shown), in spite of the weak signals on the FACS® (Fig. 2).
Quite a different result was obtained in cells exposed to HEL together with a maturation stimulus (Fig. 3, right). As before, HEL protein was found in MIICs, but in all the DCs, abundant C4H3 stain could also be observed within MIICs. After 3 h or less in HEL, C4H3 staining was only observed within MIICs (Fig. 3). At 24 h, most of the MHC–peptide complexes were on the cell surface (Fig. 3), as described previously for the distribution of MHC class II products in maturing DCs 20,21. Similar findings were made in pulse–chase experiments, carried out as in Fig. 2 B; when a maturation stimulus was applied after the pulse with HEL antigen, MHC–peptide complexes began to form in MIICs within 3 h, and were later found on the cell surface.
MIICs as a Site for MHC–Peptide Complex Formation in the Epidermis.
To visualize the intracellular sites for MHC class II–peptide formation in DCs in intact tissue, we studied epidermal DCs that mature when they migrate from skin 32,33 via afferent lymphatics 38,39. Endotoxin-containing HEL was injected into the ear dermis or added to skin explants in vitro. Epidermal sheets and the emigrated DCs were then examined. 5 h after injection, clear cut colabeling for C4H3 and the lysosomal marker H-2M was observed within all DCs in the epidermal sheets (Fig. 4 A), but not in DCs given control OVA protein. At 22 h, the C4H3 was on the surface of most DCs (Fig. 4 A). DCs that emigrated from explants into the medium during 2 d of culture had a mature CD86++ phenotype (Fig. 4 B) and stained strongly for C4H3 if HEL had been added to the explant. If HEL was added to mature DCs after emigration, C4H3 staining was not seen (Fig. 4 B). Thus, immature endocytic epidermal DCs form MHC–peptide complexes in MIICs in vivo, but mature DCs are no longer able to do so.
Functional Consequences of the Synergism between Antigen and a Maturation Stimulus: In Vitro Studies of Antigen Presentation to Naive TCR Transgenic T Cells.
In vitro antigen presentation assays were used to assess the functional consequences of DC maturation and increased amounts of MHC–peptide complexes. We used naive 3A9 TCR transgenic T cells 35 specific for the same I-Ak–HEL peptide complex recognized by C4H3 antibody. Immature, marrow-derived DCs were exposed for 1 d to HEL in graded doses, to CD40L or LPS, or to both HEL and the maturation stimulus. The cells were washed, fixed in paraformaldehyde, and added in graded doses to TCR transgenic T cells (Fig. 5 A). Fixation ensured that the DCs did not mature during the culture with T cells. HEL-pulsed immature DCs only weakly stimulated the TCR transgenic T cells, but when the DCs had been exposed to both HEL and CD40L or LPS, strong T cell stimulation ensued (Fig. 5 A). With mature DCs, the DNA synthesis (3H-TdR uptake) that we observed early in the response (day 2) with low DC to T cell ratios (1:60–1:2,000) was >2 × 105 cpm per well under standard assay conditions, and it was proportional to the level of MHC class II–peptide on the DCs (Fig. 5 A).
The poor T cell–stimulatory activity of immature DCs could reflect a lack of either TCR ligand (“signal one”) or costimulatory molecules such as CD86 (“signal two”). The use of the C4H3 antibody for the detection of signal one was likely to be insensitive relative to that of the TCR. For example, an exposure of DCs to 30 μg/ml of HEL was necessary to detect MHC–peptide complex formation with C4H3 antibody, whereas an exposure to 10 μg/ml was sufficient to saturate T cell stimulation at a DC/T cell ratio of 1:60 (Fig. 5 A, left). To assess MHC class II–peptide complexes at the level of the TCR, we used the criterion of TCR downregulation. Prior studies have shown that the TCR is cleared from the cell surface upon encounter of MHC–peptide complexes; this downregulation appears not to require a special type of APC or the presence of costimulatory molecules on that APC 40,41. Using staining with anti-Vβ8 antibody, we monitored expression of the TCR on 3A9 cells cultured for 5 h with mature or immature HEL-pulsed DCs, prepared exactly as described in the legend to Fig. 5 A. At all doses of HEL studied (10–1,000 μg/ml), mature DCs quickly activated a sizable fraction of the T cells, as indicated by de novo expression of the CD69 T cell activation marker. TCR downregulation was evident on all CD69+ cells. In contrast, immature DCs induced little increase in CD69 or decrease in TCR (Fig. 5a). We conclude that a maturation stimulus is needed for DCs to form MHC class II–peptide complexes, as assessed with B cell (C4H3 antibody) or T cell (the 3A9 TCR) receptors.
Functional Consequences of the Synergism between Antigen and a Maturation Stimulus: In Vivo Studies of Antigen Presentation to Naive Mice.
The above data establish a relationship between maturation and the formation of immunogenic MHC–peptide complexes in vitro, but the relevance of DC maturation to immunogenicity in vivo has yet to be tested. To this end, immature DCs were incubated in graded doses of endotoxin-free HEL with or without a maturation stimulus, CD40L. Unfixed DCs were then injected subcutaneously into naive syngeneic mice. The mature cells were far more powerful immunogens compared with immature cells, and priming was proportional to the dose of HEL and MHC peptide on the DCs (Fig. 6). When we tested T cells isolated from lymph nodes of mice injected with mature HEL-loaded DCs, the proliferation was >100-fold more sensitive to a given dose of antigen. If 25 μg of HEL was injected without DCs, T cell priming was not seen (not shown). Therefore, maturation markedly increases the adjuvant capacity of adoptively transferred DCs that have been previously pulsed with antigens ex vivo.
As introduced above, maturation changes DCs in many ways that explain their potency in initiating immunity. Maturing cells express surface molecules like CD40 and CD86 required for T cell expansion and cytokine synthesis 11,12, produce high levels of IL-12 13,14, resist suppression by IL-10 15, and upregulate chemokine receptors 16,17,18,19 that recognize chemokines produced in lymphoid organs 42,43. Here, a new feature of maturation has emerged, an intensified capacity to convert endocytosed antigens to MHC class II–peptide complexes.
There is evidence with human monocyte–derived DCs that maturation can prolong the life span of MHC class II–peptide complexes on the cell surface 20. However, direct visualization of these complexes now shows that the maturation process can regulate an earlier and more critical step, the formation of TCR ligands in MIICs. The results in a companion study 37, in which antigen-pulsed DCs can be cultured for 48 h before forming MHC–peptide complexes in response to a maturation stimulus, nullifies the possibility that immature DCs are simply degrading these complexes.
MHC class II products are designed to capture antigenic peptides from pathogens internalized by endocytosis 44,45. Through the chaperone function of the Ii chain 46,47 or by dileucine motifs within the cytoplasmic domain of the class II β chain 48, MHC class II molecules are targeted to acidic compartments termed MIICs 49,50,51. These lysosomes contain proteases for antigen processing, as well as HLA-DM or H-2M molecules for enhancing peptide loading to MHC class II 52,53. Nevertheless, there has been little direct visualization of newly formed MHC–peptide complexes in physiological APCs. A key control is likely to be the uptake and delivery of antigens into MIICs. DCs are actively endocytic at the immature stage of development 3,4, when they also contain numerous MIICs 21,54,55. Both endocytosis and the number of MIICs are decreased, and even extinguished, upon further differentiation or maturation in response to bacteria and TNF family members. Here, we have studied DCs frozen at the immature stage (Materials and Methods). These DCs internalize antigens into MIICs, but surprisingly, MHC class II–peptide complexes fail to form efficiently until an inflammatory stimulus is applied.
Because cathepsin S controls Ii chain proteolysis 56 and the availability of MHC class II molecules for peptide loading, one of the possible mechanisms explaining these data is that maturation downregulates the cathepsin S inhibitor, cystatin C 22. Consistent with this, we observed a striking inability to load the MHC class II products of immature DCs with peptide, until a maturation stimulus was applied. Similar events could take place in vivo, because when the maturation stimulus LPS in HEL preparations is given to mice, there is an accumulation of DCs with high levels of C4H3 in T cell areas 36. The control of MHC class II–peptide formation could also be exerted at other levels, e.g., the acidity of the endocytic system and the function of additional proteases.
In companion studies, we show that MHC class II–peptide complexes, newly formed in the late endosomes or lysosomes of DCs, selectively accumulate, together with CD86 costimulators, into specialized vesicular carriers before their delivery to the plasma membrane 37. We term these MHC class II–positive lysosome-negative vesicles “CIIVs”, for class II vesicles. Remarkably, the MHC class II–peptide complexes and CD86 costimulators remain in clusters at the cell surface. This juxtaposition may explain the potent naive T cell stimulation of DCs that we now observe, even when the cells are fixed in a low dose of formaldehyde (Fig. 5).
Our findings, especially the in vivo priming results in Fig. 6, are relevant to the optimal use of DCs for immunotherapy in humans 27,28,29,30. It is not only valuable to deliver antigens to DCs, but also to mature the cells to increase successful antigen processing and to provide other valuable physiological features cited above. The data also relate to the potential of DCs to control the alternative outcome of antigen deposition, tolerance rather than immunity. Antigen-bearing immature DCs may induce nonresponsiveness, according to views that tolerance develops if T cells recognize antigen in the absence of costimulation 57,58. We would question whether immature DCs, at least the cells that we have generated here from bone marrow progenitors, can directly tolerize, since requisite MHC class II–peptides do not form efficiently. Instead, the recently described transfer of antigen 59 from immature peripheral DCs to resident lymph node–processing DCs may generate the MHC class II peptide complexes needed by tolerizing APCs that are found in lymph nodes 60. The mature cells that we generate ex vivo and that efficiently prime T cells in vivo (Fig. 6) express high levels of CD86 costimulators. We are intrigued by the fact that in the apparent absence of maturation stimuli, lymph node DCs that are implicated in tolerance 60 can express high levels of MHC–peptide but relatively low levels of CD86 61.
We are grateful to many colleagues who provided vital reagents for this work: Dr. Mark Davis for 3A9 TCR transgenic T cells, Dr. E. Unanue for 1G12 anti-clonotype antibody, and Dr. P. Allen for 1B12 anti-HEL antibody.
This work was supported by grants to R.M. Steinman from the National Institute of Allergy and Infectious Diseases (AI-13013 and AI-39672) and the Yamanouchi Fund, to I. Mellman from the National Institutes of Health (AI-34098 and GM-33904), and to K. Inaba from the Ministry of Education, Science and Culture of Japan (Nos. 08282104, 10153226, and 10044268).
C. Reis e Sousa's present address is Immunobiology Laboratory, Imperial Cancer Research Fund, London WC2A 3PX, United Kingdom.
Abbreviations used in this paper: CD40L, CD40 ligand; DC, dendritic cell; HEL, hen egg lysozyme; 3H-TdR, [3H]thymidine; Ii chain, MHC class II–associated invariant chain; MIICs, MHC class II compartments.