In immature dendritic cells (DCs), CD1a is almost exclusively expressed at the cell surface and its membrane organization is poorly understood. In this study, we report that MHC class II, invariant chain (Ii), and CD9 molecules are coimmunoprecipitated with CD1a in immature DCs, and that CD1a/Ii colocalization is dependent on lipid raft integrity. In HeLa-CIITA cells CD1a expression leads to increased Ii trafficking to the cell surface, confirming the relevance of this association. Furthermore, silencing of Ii in DCs induces significant CD1a accumulation on the plasma membrane whereas the total CD1a expression remains similar to that of control cells. These data suggest that CD1a recycling is facilitated by the association with the Ii. The CD1a localization in lipid rafts has functional relevance as demonstrated by inhibition of CD1a-restricted presentation following raft disruption. Overall, these findings identify Ii and lipid rafts as key regulators of CD1a organization on the surface of immature DCs and of its immunological function as Ag-presenting molecule.
Three types of molecules are able to present Ags to T cells: MHC class I (MHC I)3and class II (MHC II) molecules encoded by the MHC locus, and the CD1 molecules, encoded by the CD1 genes localized outside the MHC locus. MHC I and MHC II molecules are specialized in peptide Ag presentation, while CD1 molecules are specialized in the presentation of lipids, glycolipids and lipopeptides (1, 2). The CD1 family is composed of five glycosylated proteins showing limited polymorphism (3). The CD1a, CD1b, and CD1c isoforms, are included in the CD1 group 1 based on their sequence homology (4), are detected on activated human monocytes (5, 6), and are highly expressed on monocyte-derived immature dendritic cells (iDCs) (3) and on dermal dendritic calls (DCs) in vivo.
During their maturation, DCs undergo cellular changes which increase peptide loading, cell surface transport of MHC II-peptide complexes, and reduce MHC II endocytosis (7, 8). An important difference between presentation of peptide and lipid Ags is that optimal peptide presentation to MHC II-restricted T cells is dependent on this maturation, whereas both immature and mature DCs optimally present lipid Ags to CD1-restricted T cells (9).
The T cell response to peptide Ags is initiated by recognition of MHC II expressed at the surface of professional APCs. This recognition is improved when MHC II-peptides complexes are concentrated on the APC plasma membrane in small lipid-rich microdomains (10, 11). We, and others, have demonstrated the localization of MHC II in lipid rafts and the resulting functional consequences on T cell activation in mouse and human B cell lines (12, 13), monocytic cell lines (14), human monocytes (15, 16), dendritic cells (17, 18), and tumor cells (19).
Although MHC II organization in lipid rafts of various cell types is now well documented, this is not the case for CD1 molecules. Two studies have demonstrated the constitutive and restricted presence of the CD1d isoform in the lipid rafts of mouse cells (20, 21), whereas the localization of human CD1 molecules in specialized surface microdomains of professional APCs has not been explored.
We have focused our attention on CD1a localization and organization in iDCs. This study investigates which proteins associate with CD1a, the possible intervention of tetraspanins, proteins acting as molecular adaptors, the CD1a distribution in specialized microdomains at the cell surface of iDCs, and the functional consequences of this organization on lipid Ag presentation.
The results described in this study demonstrate that CD1a molecules are associated with invariant chain (Ii) in iDCs, that CD1a and Ii are recruited to cell surface lipid rafts, and that colocalization of CD1a and Ii is dependent on cholesterol-dependent lipid rafts. Moreover, we show that CD1a-induced expression in HeLa-CIITA cell line led to an increased trafficking of Ii to the cell surface, and that silencing of Ii in iDCs induced significant CD1a accumulation at the cell surface. Finally, we show the role of iDCs membrane lipid rafts in T cell response to glycolipidic Ag presented by CD1a. The overall results of the present study suggest that Ii and lipid rafts have an impact on CD1a surface expression and CD1a-restricted T cell response.
Materials and Methods
Abs and immunoflurescence staining
CD1a mAbs were: OKT6 (American Type Culture Collection), WM35 (BD Pharmingen), BL6 (Immunotech/Beckman), and C19, CD1 polyclonal goat serum (Santa Cruz Biotechnology). The anti-MHC II mAbs locally produced were D1.12 specific for HLA-DR α-chain and DA6.147 recognizing intracytoplasmic epitope of HLA-DRα. The Ii, CD74 mAbs was: LN2 (BD Pharmingen). The CD9 mAbs were Syb.1 (gift from Dr. E. Rubinstein, Institut National de la Santé et de la Recherche Médicale, Villefuif, France), and PHN 200 (provided during the V International Workshop) (22). FITC-conjugated mAbs were ALB6 (CD9) from Immunotech/Beckman; G46-6 (HLA-DR), HI149 (CD1a), MB741 (Ii), and HB15e (CD83) from BD Pharmingen; LN2 from Santa Cruz Biotechnology; and AD5–8E7 (CD1c) from Miltenyi Biotec. PE-conjugated mAbs were G46-6 (HLA-DR) and HI149 (CD1a). Isotype control mAbs were from BD Pharmingen. Goat-anti-mouse F(ab′)2 IgG FITC- and PE-conjugated mAbs, FITC-conjugated cholera toxin B subunit (CT) and HRP-conjugated CT were from Sigma Aldrich. Texas-red-conjugated goat anti-mouse IgG was from Immunotech/Beckman. Goat anti-mouse IgG2b Alexa594-conjugated mAb used to detect intracellular CD1a by confocal microscopy was purchased from Invitrogen. Mouse mAb anti-transferrin receptor was from Zymed. Flotillin-1 mAb was from BD Transduction Laboratories. The CD41 mAb was from Diaclone.
Generation of monocyte-derived immature DCs
iDCs were prepared from PBMC isolated by Ficoll-Paque (Pharmacia Biotech) density centrifugation of buffy coats obtained from the local blood bank and depleted of platelets. Monocytes were purified by positive selection using CD14 conjugated microbeads (Miltenyi Biotec) and then incubated for 7 days at 37°C in 10% FCS-RPMI 1640 complete medium with 1000 UI/ml human rIL-4 and 800 UI/ml human recombinant GM-CSF (R&D Systems) to induce monocyte differentiation into iDC. Unless otherwise indicated, cells were collected at day 7 for their use in subsequent experiments.
Immunoprecipitations and Western blotting
When indicated, iDCs were first surface biotin-labeled by incubation for 1 h on ice with 5 mM sulfo-NHS-LC-biotin (Pierce) and unbound biotin was removed by washing with cold PBS containing 1 mM glycine. After biotinylation, integrity of plasma membranes was checked by blue trypan exclusion and >85–90% of cells were still alive after this treatment. Detergent lysates of unlabeled and biotin-labeled iDCs were made by incubating cells for 30 min on ice in a buffer containing either 1% Nonidet P-40 or 1% CHAPS with 50 mM Tris, 150 mM NaCl, 5 mM EDTA, 20 μg/ml aprotinin, 10 μg/ml leupeptin, 1 mM PMSF, 10 mM sodium pyrophosphate, and 10 mM Na3VO4. Remaining particulate material was removed by centrifugation at 15,000 × g for 30 min at 4°C.
Lysates were precleared with CD41 mAb and protein A-Sepharose (Pharmacia Biotech) and subsequently incubated with specific mAbs overnight at 4°C. Immune complexes were recovered with protein A-, or protein G-Sepharose beads (Pharmacia Biotech), according to the isotype of mAbs used and washed with detergent-containig lysis buffer. Immunoprecipitated proteins were eluted from beads by boiling in SDS nonreducing sample buffer, subjected to SDS-PAGE on a 10% polyacrylamide gel, and transferred to a nitrocellulose membrane. After blocking in PBS containing 5% milk and 0.1% Tween 20, filters were probed with either biotinylated mAbs followed by streptavidin-conjugated HRP (Amersham Biosciences) or, when immunoprecipitations were conducted on lysates from biotinylated cells, by streptavidin-conjugated HRP only. Enzymatic reaction was detected with enhanced chemiluminescent reagents (ECL-kit) (Amersham Biosciences).
iDCs were biotin-labeled, lysed, and precleared as described above, and immunoprecipitation with LN2 or IgG1-isotype control mAbs was performed. Proteins that had been immunoabsorbed on beads were then eluted by boiling at 95°C in PBS containing 2% SDS. Eluted material was then diluted in 1% Triton X-100 lysis buffer. CD1a and Ii were then immunoprecipitated from eluted proteins with specific mAbs and protein A- and G-Sepharose beads, eluted in nonreducing conditions, and analyzed as described above. Nonreducing conditions were used to avoid interference between molecules of interest and Ig heavy and light chains.
Biochemical isolation of rafts and immunoprecipitations
Rafts of unlabeled iDCs were isolated by sucrose gradient equilibrium centrifugation after nonionic detergent lysis as previously described (15, 23). After centrifugation, ten fractions of 400 μl each were collected from the top to the bottom of the tube. Lipid rafts (fractions 3–4) and soluble proteins (fractions 8–10) were then collected, lysed for 30 min on ice with buffer containing 1% CHAPS, and immunoprecipitated as described above. AlphaEaseFC (Fluochem 8800 from Alpha Innotech Corporation) software was used to protein quantification.
Capping, patching, and microscopy immunofluorescence
For cocapping experiments, iDC were incubated with specific mAbs (10 μg/ml) for 30 min on ice, washed twice in cold PBS, and subsequently incubated with Texas-red-conjugated goat anti-mouse Ig for 25 min at 37°C to allow complete capping. Cells were then washed twice and stained with the indicated FITC-conjugated mAbs for 30 min on ice. After two washes, cells were fixed in PBS-1% paraformaldehyde for 30 min at 4°C. Fixed cells were kept in PBS supplemented with DABCO (Sigma, Lyon, France) to prevent photobleaching and 4′,6′-diamidino-2-phenylindole (Sigma-Aldrich) to visualize nuclei. Cells were analyzed with a MRC 1024 confocal microscope (Bio-Rad). Images were obtained using LSM510 version 4.0 (Carl Zeiss).
Copatching experiments were performed as previously described (24).
For single intracellular stainings, cells were fixed in 2% paraformaldehyde in PBS for 10 min at room temperature. After an extensive wash in PBS, cells were incubated in fresh glycine 0.1 M in PBS and subsequently permeabilized 45 min in saponin buffer (0.05% saponin, 0.2% BSA in PBS). For direct labeling, the conjugated Ab were then added for 30 min at room temperature. After extensive washes in PBS, cells were fixed in 1% paraformaldehyde and immediately analyzed by flow cytometry. For single indirect labeling, cells were extensively washed and then incubated with goat-anti-mouse F(ab′)2 IgG FITC- or PE-conjugated mAbs and finally treated as described above.
For double intracellular stainings of iDC analyzed by confocal microscopy, iDCs were permeabilized as above and CD1a protein was detected using WM35 (IgG2b) mAb followed by goat anti-mouse IgG2b Alexa594-conjugated mAb from Invitrogen Life Technologies. HLA-DR and the invariant chain were respectively revealed using FITC-conjugated mAb G46-6 (IgG2a) from BD and LN2 (IgG1) from Santa Cruz Biotechnology. No reactivity has been observed between the goat anti-mouse IgG2b Alexa 594-conjugated mAb and those FITC-conjugated mAbs.
Cell lines, plasmids, and siRNA transfection
CD1A and CD1B coding sequences were amplified with Pfu Turbo DNA polymerase (Invitrogen Life Technologies) from the plasmid construct pCDM8 kindly provided by B. Seed. The following primers were used to add a 5′XhoI and a 3′BamHI restriction site on CD1A coding sequence: forward, 5′-CCGCTCGAGCGGCAAATGATATGCTGTTTTTGCTAC-3′ and reverse, 5′-CGGGATCCACACAGAAACAGCGTTTCCTGAACC-3′. The following primers were used to add a 5′NheI and a 3′BamHI restriction site on CD1B coding sequence: forward, 5′-CTAGCTAGCTAGATCTCCCAGTGAAATGCTGCTG-3′ and reverse, 5′-CGGGATCCACTGGGATATTCTGATATGACCGGC-3′. The two fragments were sequenced and cloned in pEGFP-N1 plasmid (Clontech, Invitrogen Life Technologies). The vectors without target sequences were used as controls.
HeLa-CIITA cells (provided by P. Benaroch, Institut Curie, Paris, France), expressing HLA-DRα and β, and Ii, were grown in DMEM (Biochrom KG) supplemented with 2 mM l-glutamine, 10 u/ml penicillin, 10 μg/ml streptomycin (Invitrogen Life Technologies) and 10% heat-inactivated FCS (Biochrom), 1 mM sodium pyruvate (Biochrom) and 0.33 mg/ml hygromycin. (Invitrogen Life Technologies). HeLa-CIITA cells were transfected with the above recombinant plasmids, using the calcium-phosphate-DNA precipitation method (25). Forty-eight hours after transfection, 0.5 mg/ml geneticin was introduced in cell culture medium and maintained several weeks before flow cytometry analysis.
Control siRNA and two different siRNA (with the equivalent GC%) specific for human Ii were obtained from Invitrogen Life Technologies. Each siRNA Ii tested led to identical results. iDCs were transfected at day 5 with siRNA specific for Ii using Amaxa technology and the transfection procedure was realized according to the manufacturer’s recommendations.
Ag presentation assay
Sulfatide (Fluka) was prepared as previously described (26). iDCs (1 × 106 cells/ml) were loaded for 2 h at 37°C with indicated concentrations of sulfatide in serum-free medium. After washing, iDCs were treated with 10 mM Mβ-cyclodextrin (MβCD) in RPMI 1640-containing 2% FCS for 30 min at 37°C. Cells were then fixed or not with 0.5% paraformaldehyde for 15 min at room temperature and blocked with 100 mM glycine. iDCs (1 × 105 cells/well) were plated in triplicate wells of 96-well flat-bottom plates and CD1a-restricted and sulfatide-specific K34B9.1 T cells (27) (1 × 105 cells/well) were added. After overnight stimulation, TNF-α production was quantified by TNF-α-capture ELISA according to the manufacturer’s instructions (R&D Systems).
CD1a and Ii are associated in iDCs
The expression of CD1a, Ii, and HLA-DR molecules on the plasma membrane of iDCs was analyzed on day 5 and 7 of their differentiation from monocytes. Analysis of cell surface (Fig. 1,A) and intracellular expression (Fig. 1 B) showed that CD1a was highly expressed at the cell surface and poorly detected in intracellular compartments of iDCs. There was a low level of Ii chain expression and it was predominantly intracellular. The HLA-DR molecules were clearly detected both at the cell surface and in intracellular compartments.
To find the partner molecules associated to CD1a at the surface of iDCs, we performed immunoprecipitation experiments with anti-CD1 mAbs. Lysates were made from surface-biotin-labeled iDCs using either Nonidet P-40, a nonionic detergent known to dissociate many molecular complexes found in membranes, or CHAPS, a milder detergent. Integrity of plasma membranes was checked by blue trypan exclusion and >85–90% of cells were still alive after this treatment. We thus assume that the majority of biotin-labeled proteins were surface molecules, but cannot exclude that a minor part of intracellular proteins could have been biotin labeled. A single biotinylated 49 kDa species that was characteristic of CD1a was immunoprecipitated from Nonidet P-40 lysates with anti-CD1 mAbs (Fig. 2,A). With CHAPS lysates, three major proteins around 33–35 and 24 kDa were coprecipitated specifically with CD1 (Fig. 2,B). The additional bands detected were nonspecific because they were also observed in immunoprecipitations conducted with isotype control mAb. (Fig. 2,B). With regard to previous studies (28, 29, 30, 31), and bands specifically coprecipitated with HLA-DR and the Ii in CHAPS lysates (Fig. 2,B), we speculated that the p33–35 proteins detected in the CD1a precipitates may be the 33 and 35 kDa isoforms of the Ii and/or the 35 kDa DRα-chain. The CD1a precipitates were analyzed by Western blot with Ii- or HLA-DR-specific mAbs and relatively high levels of Ii were detected in CD1a precipitates as compared with the relatively low levels of HLA-DR molecules (Fig. 2,D). These results demonstrate the association of CD1a and Ii in iDCs. To confirm this association at the cell surface, we have performed sequential immunoprecipitation experiments. Ii was immunoprecipitated from surface biotinylated iDCs and eluted material was reimmunoprecipitated with CD1a- or Ii-specific mAbs. Ii isoform at ∼33 kDa and specific Ii dimers at ∼66–80 kDa (32) were visualized in Ii precipitates. Reciprocally, surface CD1a was precipitated from Ii-immunoprecipitates (Fig. 2 E), thus demonstrating that CD1a and Ii are indeed associated at the cell surface of iDCs. This observation does not preclude that a third molecular partner could be involved in this association.
The tetraspanin CD9 has been shown to be mostly expressed at the plasma membrane of human iDCs (33), and we have previously shown that CD9 can also interact with MHC II at the surface of human monocytes (15). We have thus explored the possibility that the p24 protein detected in CD1a precipitates could be CD9. After solubilization in CHAPS (Fig. 2,C), many surface-biotin-labeled proteins were coprecipitated specifically with CD9 including a 49 and 33 kDa protein (as it has been previously described, Ref. 34). Western blotting of CD1a precipitates with the CD9 mAb Syb.1, identified the p24 protein as CD9 (Fig. 2 D). Identical experiments performed on iDCs lysed in Brij-98 led to the same observations (data not shown).
These results show that CD1a associates with the invariant chain, the tetraspanin CD9 and MHC II molecules, and demonstrate the CD1a/Ii association at the cell surface in human iDCs.
CD1a and Ii colocalize at the surface of living iDCs
To confirm the CD1/Ii association at the cell surface of living cells, we used mAbs to cross-link CD1a molecules at the surface of living iDCs and then analyzed the redistribution of Ii, HLA-DR, Ii, and CD9 by confocal microscopy.
While in steady-state, the four molecules are uniformly distributed at the cell surface (data not shown); cross-linking of CD1a led to its redistribution either in caps or in patches at the cell surface (Fig. 3, A–C, -MβCD, red panels). Ii, which is expressed at low density at the cell surface of iDCs, clearly colocalized with CD1a patches in these conditions (Fig. 3,A, -MβCD, merge panel). The same observation was made with CD9, which is highly expressed on iDCs and is also redistributed in CD1a patches (Fig. 3,C, -MβCD, merged panel), indicating the proximity of these molecules at the surface of living iDCs. The redistribution of HLA-DR molecules after CD1a-cross-linking was more heterogeneous: some minority areas of colocalization with CD1a were observed, and the majority of HLA-DR labeling was also observed outside CD1a patches (Fig. 3 B, -MβCD).
Together, these results demonstrate that the dynamic recruitment of CD1a at the surface of iDCs induces a highly significant redistribution of Ii and CD9 together with CD1a and partial redistribution of HLA-DR, therefore confirming that CD1a associates with Ii at the surface of living iDCs.
CD1a induces enhanced surface expression of Ii
After sorting from the Golgi apparatus, Ii is mainly directed to endo/lysosomal compartments. Nevertheless, a pool of Ii can also escape from this trafficking and directly reach the cell surface without entering endosomal compartments. As we have observed a tight link between CD1a and Ii at the surface, we asked whether CD1a could participate in cell surface transport of Ii. We have established HeLa-CIITA cell lines stably expressing CD1a- or CD1b-GFP molecules, and controlled that GFP does not modify the cell surface or the cellular trafficking of the proteins as compared with wild type ones, as previously demonstrated (35). Fig. 4,A depicts the CD1a and CD1b surface and total expression in both cell lines. We have analyzed the surface and total expression of Ii and HLA-DR on these CD1a+ cells and compared the results with those observed on CD1b+ cells and mock-treated HeLa-CIITA cells. A significant increase in the surface expression level of Ii was observed in CD1a+HeLa-CIITA cells, while the total (surface and intracellular) Ii expression was unchanged in comparison with the non-CD1a expressing HeLa-CIITA cell lines (Fig. 4 B). This indicates that, on CD1a+Hela-CIITA cells, the proportion of Ii molecules transported to the surface is increased. Concerning HLA-DR expression, these cells harbored a slight and equivalent increase in both surface and total expression. Therefore, this indicates a global increase of HLA-DR expression in those cells without any change in surface vs intracellular HLA-DR distribution.
The expression of Ii and MHC II molecules in CD1b-expressing HeLa-CIITA cells was equivalent to that observed on mock cells: a slight decrease of Ii was observed at the cell surface as well as in permeabilized cells, and HLA-DR was decreased at the cell surface. Thus, the increase in Ii surface transport was a specific effect of the CD1a expression in HeLa-CIITA cells.
Ii interference induces CD1a accumulation at the cell surface of iDCs
To further investigate the functional relevance of CD1a/Ii association on iDC, we performed RNA interference experiments to knockdown the Ii expression on iDCs. iDCs obtained at day 5 of monocyte differentiation were transiently transfected with siRNA specific for Ii and analyzed by cytofluorimetry. Expression of CD83 was not detected on iDCs transfected with Ii- or control- siRNAs (data not shown), indicating that transfection did not induce DC maturation. Under these conditions, the total expression of Ii was reduced by 95% (Fig. 5,A). This is clearly visualized in Fig. 5,B, which shows that Ii expression is not observed after siRNA Ii transfection, neither at the cell surface nor in cellular compartments. As expected, under these conditions the surface and total HLA-DR expression was decreased. (Fig. 5, B and D). In control iDCs, CD1a was detected at the plasma membrane as well as intracellularly, where it colocalized with Ii (Fig. 5,C). Silencing of Ii resulted in an increase of surface CD1a expression (∼20%) (Fig. 5 D). Therefore, these results indicate that in the absence of Ii, CD1a molecules accumulate at the cell surface and suggest that the Ii participates in CD1a endocytosis.
CD1a and its molecular partners are constitutively present in iDCs lipid rafts
Clustering of proteins at the cell surface involves different actors and among them the tetraspanins have been implicated due to their capacity to link numerous proteins and raft microdomains. For this reason, we next asked whether there was a particular localization of these molecules in specialized membrane microdomains (36, 37).
Triton X-100 lysates of iDCs were fractionated by sucrose-gradient density ultracentrifugation to separate soluble proteins from detergent resistant membranes. We then examined the distribution of CD1, HLA-DR, Ii, and CD9 molecules in each fraction (Fig. 6 A). Flotillin, a typical constituent of lipid rafts was mainly recovered in fractions 3 and 4 confirming the separation of lipid rich microdomains. The CD1 molecules were mainly detected in soluble fractions (fractions 8–10). Nevertheless, protein quantification reveals that a significant proportion (12.8%) of CD1 molecules was found in lipid rafts (fractions 3–4).
CD1a and Ii can enter cell surface lipid rafts of iDCs
To determine the localization of these four molecules in cell surface lipid rafts, copatching experiments at the surface of living cells were performed using cholera toxin β (CTX), which binds to GM1, a constituent of cell surface lipid rafts. The iDCs were labeled with specific mAbs followed by a Texas Red-conjugated secondary Ab and FITC-conjugated CTX, a treatment that led to aggregation and patching of molecules. Analysis by confocal microscopy revealed a clear and almost total colocalization of Ii molecules with GM1 and a partial one of CD1a (Fig. 6 B). These results have been observed in different iDCs preparations and were visualized on each plan of cells. Z-stack analysis on single cells have confirmed Ii/CTX and CD1a/CTX colocalization (data not shown). In agreement with previous studies (15, 38), only a minor fraction of CD9 molecules colocalized with GM1.
Thus, Ii highly colocalized with surface CD1a clusters and both CD1a and Ii clusters were found in surface lipid rafts. These results strongly suggest that CD1a associates with Ii in iDC surface lipid raft microdomains.
Lipid rafts: dependent CD1a/Ii association at iDCs surface
We then examined whether the dynamic recruitment of Ii and CD9 with CD1a at the surface of iDCs (Fig. 3) is dependent on lipid raft integrity. Cells were treated with MβCD to deplete cholesterol and disrupt lipid rafts, and colocalization experiments were performed as described above. Induced CD1a relocalization after MβCD treatment no longer led to the enrichment of Ii (Fig. 3,A) into CD1a patches. The few areas of HLA-DR/CD1a colocalization observed before treatment with MβCD were almost completely lost by cholesterol depletion (Fig. 3,B). In contrast, the CD1a/CD9 colocalization was still observed after raft disruption (Fig. 3 C). These results indicate that the recruitment of Ii with CD1a is highly dependent on lipid raft integrity.
To further investigate the CD1a/Ii association in iDC surface lipid rafts, we have pulled-down CD1a from sucrose gradient fractions of surface biotinylated iDCs and revealed surface protein with streptavidin Western blotting. We observed that CD1a was equally distributed throughout the gradient, including the lipid raft enriched fractions, confirming that CD1a is present in cell surface lipid rafts of iDCs in steady state conditions (Fig. 6 C). Moreover, a p33 protein corresponding to the m.w. of Ii p33, was exclusively detected in CD1a precipitates from raft fractions (fraction 3–4).
Altogether these results demonstrate that the CD1a/Ii association is constitutive in iDCs and can take place during the dynamic recruitment of CD1a in lipid rafts.
iDCs membrane lipid rafts are essential for CD1a-restricted T cell activation
To investigate the functional impact of CD1a-lipid raft localization on Ag presentation and T cell activation, we have disrupted iDCs lipid rafts by cholesterol depletion and analyzed the resulting response of the CD1a-restricted T cell clone K34B9.1 to the sulfatide Ag.
Using fixed iDCs to avoid serum-dependent lipid raft reconstitution, we have observed that MβCD treatment inhibits TNF-α production by the T cell clone (Fig. 6,D) and this is particularly evident at low concentrations of sulfatide. To exclude toxic and inhibitory effects on T cells, we have performed the same analysis with iDCs MβCD-treated but not fixed and showed that under these conditions which allow raft reconstitution, the T cell response was restored (Fig. 6 D, inset). These results demonstrate that CD1a concentration in lipid rafts is crucial to properly activate CD1a-restricted T cells.
Our study demonstrates that in human iDCs CD1a associates with the Ii, HLA-DR, and the tetraspanin CD9 and partitions into lipid rafts. The CD1a/Ii association takes place during recruitment of CD1a into surface lipid rafts and this type of membrane localization promotes efficient Ag presentation and T cell activation. We also provide evidence that CD1a and Ii control the pool of surface Ii and that Ii participates in CD1a endocytosis.
Our biochemical analysis revealed that several molecules, such as MHC II, Ii p35, Ii p33, and CD9 coprecipitate with CD1a. The amounts of Ii and CD9 proteins detected in CD1a precipitates are strikingly more abundant than those of coprecipitated HLA-DR. Cocapping experiments are in agreement with this finding, because confocal microscopy showed a faint presence of HLA-DR within CD1a clusters, and a prominent colocalization of CD1a with Ii and CD9. The scarce number of CD1a-HLA-DR clusters argues against their physiological relevance.
The association of the CD9 tetraspanin with CD1a, which resembles the CD9/HLA-DR clusters present on human monocytes (15), appears to be more important. Another CD1 isoform, CD1d, has been found associated with the CD82 tetraspanin and MHC II molecules. CD1d may associate with MHC II independently of CD82 (28). The reason why CD1d does not coprecipitate with tetraspanins remains unclear. Tetraspanins associate with numerous surface molecules and form networks which organize the local distribution of interacting molecules. These networks are also formed in membrane domains outside lipid rafts, thus demonstrating the important general role of tetraspanins in membrane organization (38, 39). The presence of large amounts of CD9 molecules that laterally associate with CD1a, may have the double effect of organizing CD1a in the membrane and enhancing CD1a-mediated Ag presentation, as has been reported for HLA-DR in DCs (29, 33).
A second mechanism used by proteins to concentrate at the cell surface is to enter lipid rafts microdomains (36). Our results reveal that CD1a molecules are constitutively present in raft compartments of total iDC lysates and in cell surface lipid rafts. A significant portion of Ii and MHC II molecules that are associated with CD1a distribute in lipid rafts. Previous studies have described the association of MHC II with lipid rafts in monocyte-derived human iDCs and have shown that only 10–20% of the total MHC II pool localizes in rafts, the remaining MHC II molecules being constitutively distributed in nonraft compartments of iDCs (17, 18). Similarly to MHC II, we find that only 10–15% of CD1a molecules are detected in iDCs lipid rafts. To evaluate the impact of raft integrity on the CD1a proximity to Ii and CD9, we performed cholesterol depletion experiments. We found that CD1a/Ii colocalization is, to a great extent, dependent of the integrity of lipid rafts, whereas CD1a/CD9 colocalization is not. This observation can result from physical differences between lipid rafts and tetraspanin-enriched microdomains. Despite the potential of tetraspanins to associate with gangliosides and cholesterol, which typically are resident in lipid rafts, the tetraspanin-enriched microdomains are resistant to cholesterol depletion (15, 40).
The association of CD1a with Ii and the dynamic recruitment of these complexes in lipid rafts has important functional consequences on CD1a Ag presentation. Indeed, CD1a association with lipid rafts facilitates activation of CD1a-restricted T cells, especially at low Ag doses. This result is in agreement with the finding that lipid raft recruitment of MHC II molecules promotes the response of peptide-specific T cells (12). That study, together with our findings, make the important point that the membrane localization of the presenting molecule is important independent of the type of presented Ag and responding T cell. It is tempting to speculate that clustering of Ag-loaded CD1a molecules has a major impact in the presence of low Ag amounts because it can facilitate TCR focusing and signaling cascade in T cells despite the low numbers of CD1a-Ag complexes.
Another important question raised in this study is the role of Ii in CD1a trafficking to the plasma membrane and through the endocytic pathway. αβIi complexes may use two distinct trafficking pathways (41). The first pathway is dependent on the AP-1, which promotes direct traffic to endosomes. The second, which is the major pathway, is dependent on AP-2, which allows rapid passage of these complexes to the cell surface. Therefore, Ii may impart different traffic behavior to associated proteins, according to the adaptor protein it is associated with. Ii has been implicated in direct membrane trafficking of MHC II. For example, in CD1a-negative APCs, such as B cells and monocytes, a minor fraction of newly synthesized MHC II molecules can directly reach the cell surface as αβ/Ii-p33 complexes (42). Our coprecipitation data together with the increased Ii surface expression in the presence of CD1a suggest that a fraction of Ii molecules directly traffic to the plasma membrane when associated with CD1a. This mechanism could operate in iDCs, a cell population expressing large amounts of CD1a, in which the majority of newly formed αβIi complexes appear firstly on the cell surface before they start recycling (43). However, the association with Ii is not an absolute requirement for surface appearance, because CD1a directly reaches the cell surface after egress from the endoplasmic reticulum and trans Golgi network in both Ii-positive or Ii-negative cells (44).
Our experiments also show that extinction of Ii in primary human DCs induces an increased surface expression of CD1a, which we interpret as a result of diminished internalization. Thus, Ii is not required for surface display of newly synthesized CD1a molecules, but might be involved in CD1a internalization and recycling.
Once CD1 molecules have reached the plasma membrane, they are internalized, transported through clathrin-coated vesicles to early endosomes, and then recycled to the cell surface (44, 45). Endocytosis of other CD1 molecules is largely dependent on intracytoplasmic tail motifs allowing interaction with AP-2 (46). The inability of the CD1a cytoplasmic tail to interact with AP-2 raises the question of the mechanism involved in CD1a internalization. We propose that CD1a molecules that associate with Ii benefit from the efficient internalization signal in the Ii cytoplasmic tail and then follow a fast internalization pathway. As a large pool of CD1a molecules binds exogeneous ligands present in the extracellular environment on the cell surface (35), it is possible that only CD1a molecules associated with Ii recycle in clathrin-coated vesicles and early endosomes and sample intracellular Ags.
Further investigations are needed to further examine the physiological significance of Ii in CD1a endocytosis and presentation of intracellular Ags.
We are grateful to Alain Haziot for critical reading of the manuscript and James Vigneron for plasmid preparation.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by Institut National de la Santé et de la Recherche Médicale, Assistance Publique-Hôpitaux de Paris, TRANSNET, and Allostem Grants MRTN-CT-2001-512253 and FP6 503319, and by the Swiss National Science Foundation Grant 3100AO-109918.
Abbreviations used in this paper: MHC I, MHC class I; MHC II, MHC class II; iDC, immature dendritic cell; DC, dendritic cell; CT, cholera toxin; MβCD, Mβ-cyclodextrin; Ii, invariant chain.