HLA-DQA2 and HLA-DQB2 Genes Are Specifically Expressed in Human Langerhans Cells and Encode a New HLA Class II Molecule

Cedric Lenormand, Huguette Bausinger, Florence Gross, Francois Signorino-Gelo, Susanne Koch, Maryse Peressin, Dominique Fricker, Jean-Pierre Cazenave, Thomas Bieber, Daniel Hanau, Henri de la Salle and Sylvie Tourne
J Immunol April 15, 2012, 188 (8) 3903-3911; DOI: https://doi.org/10.4049/jimmunol.1103048

Abstract

The precise role of human epidermal Langerhans cells (LCs) in immune response is highly controversial. While studying the gene expression profile of these cells, we were intrigued to identify the HLA-DQB2 gene as potentially expressed in LCs. Despite a strong evolutionary conservation of their sequences, the concomitant expression of the poorly polymorphic HLA-DQA2/HLA-DQB2 genes, paralogous to the HLA-DQA1/HLA-DQB1 genes, has never been detected in any cell type. We confirmed by RT-PCR that the HLA-DQA2 and -DQB2 genes are both expressed in LCs, but not in monocyte-derived dendritic cells, or in blood CD1c+ or plasmacytoid dendritic cells. The presence of the HLA-DQβ2 chain in LCs could be demonstrated by Western blotting, whereas immunofluorescence revealed its localization in early endosomes. As in the case of other HLA class II molecules, the HLA-DQα2 and -DQβ2 chains formed heterodimers that had to associate with the invariant chain to reach endosomal compartments. HLA-DQα2/β2 heterodimers were expressed at the cell surface, where they could mediate staphylococcal superantigen stimulation of T cells. Interestingly, HLA-DQα2 and HLA-DQβ1 chains formed mixed heterodimers which efficiently left the endoplasmic reticulum. These observations strongly suggest that the poorly polymorphic HLA-DQA2 and -DQB2 genes should be considered to be of immunological importance. The HLA-DQα2/β2 molecules could influence the complexity of the repertoire of Ags presented by LCs.

Human leukocyte Ag class II (HLA-II) molecules, expressed by APCs, present peptidic Ags to CD4+ T cells and in this manner are essential to adaptive immune responses against pathogens and tumor development (1). HLA-II molecules are also implicated in immunologic tolerance and participate in deleterious autoimmune reactions (2). The HLA-II genes are distributed over three highly polymorphic contiguous genetic loci—HLA-DR, HLA-DP, and HLA-DQ (subsequently designated HLA-DQA1/HLA-DQB1)—which encode conventional histocompatibility molecules. This region also includes the poorly polymorphic HLA-DM and HLA-DO genes, whose products are molecular chaperones participating in HLA-II–restricted Ag presentation. The HLA-II region also includes six pseudogenes (3) and the two genes paralogous to HLA-DQA1/DQB1, HLA-DQA2, and HLA-DQB2. Although HLA-DQA2 transcripts are present in B lymphoblastoid cell lines, HLA-DQB2 mRNA has never been detected (4). The HLA-DQα2 chain detected in lymphoblastoid cells has been shown to associate with the invariant chain (Ii) and to reach the plasma membrane, but no association with HLA-DQβ–chains could be confirmed (5), questioning the meaning of these observations. Remarkably, the two genes are poorly polymorphic (6–8) and fairly well conserved among Old and New World primates (9, 10), indicating an important biologic function of their products. A comparative analysis of the now available genomic sequences confirms this conservation, except in the gorilla, whose HLA-DQB2 ortholog contains a 23-nucleotide insertion in exon 3 (Supplemental Fig. 1).

Epidermal Langerhans cells (LCs) are the resident dendritic cells (DCs) of the epidermis. Because their exact role in the immunologic response is currently under debate (11), we established the transcriptomes of skin LCs (sLCs) and LCs differentiated in vitro from CD34+ cord blood cells (CD34+LCs; available on request) and compared them to the transcriptomes of human monocyte-derived DCs (monoDCs; available at GeoDatasets). We were intrigued to find HLA-DQB2 transcripts in LCs, but not in monoDCs. In this study, we demonstrate that human LCs express both HLA-DQA2 and HLA-DQB2, and we show that their protein products might represent new HLA-II Ag-presenting molecules specifically expressed in LCs.

Materials and Methods

Abs

The following anti-tag mAbs were used: rat anti-hemagglutinin (HA; 3F10) conjugated or not with HRP (Roche Applied Science, Meylan, France) and mouse anti-V5 (SV5-Pk1) conjugated or not with biotin, Alexa488 (A488) or HRP (AbD Serotec, Düsseldorf, Germany).

We purchased the mouse anti-human Ii mAb Pin.1 and goat anti-EEA1 IgGs from Santa Cruz Biotechnology (Santa Cruz, CA), the mouse anti–HLA-DM mAb Tal18.1 (5C1) and mouse anti–HLA-DQ mAb B-K27 from Abcam (Paris, France), sheep anti-TGN46 IgGs from Abnova (Tapei, Taiwan), rabbit anti-calnexin IgGs from Stressgen (Victoria, Canada), and mouse anti-CD107a/LAMP-1 (H4A3) and anti-CD107b/LAMP-2 (H4B4) mAbs from BD Biosciences (Le Pont de Claix, France). Rabbit anti-human Ii IgGs have been described previously (12). The mouse anti–HLA-DRα mAb DA6.147 (hybridoma supernatant) was produced in our laboratory.

We purchased polyclonal goat anti-mouse [PE-conjugated F(ab′)2 fragments] and HRP-conjugated rabbit anti-mouse IgGs were purchased from DakoCytomation (Trappes, France), HRP-conjugated goat anti-mouse IgM (μ-chain specific) from Pierce (Fischer Scientific, Illkirch-Graffenstaden, France), FluoProbe (FP)547-conjugated donkey anti-rat IgG [F(ab′)2] from Interchim (Montluçon, France), and the Cy5-conjugated polyclonal donkey anti-sheep, rabbit, or goat IgG [F(ab′)2] and Cy5-conjugated polyclonal sheep anti-mouse IgG [F(ab′)2] from Jackson ImmunoResearch (West Grove, PA). All secondary Abs were depleted of IgGs reacting with Igs from other species (minimal cross-reactivity grade).

The directly conjugated mouse mAbs used were: A488-conjugated anti-CD207 (DCGM4; Dendritics, Lyon, France), FITC-conjugated anti-CD1a (BL6; Beckman-Coulter, Villepinte, France), A647-conjugated anti–HLA-DR (L243; Biolegend, San Diego, CA), allophycocyanin-conjugated anti-CD3 (Miltenyi Biotec, Paris, France), and Cy3-conjugated anti-CD63 (H5C6) (13).

Control Igs were: mouse IgG1 (679.1 Mc7) and IgG2a (U7.27; Beckman Coulter), normal mouse serum (Rockland, Gilbertsville, PA), polyclonal rabbit IgGs (Southernbiotech, Birmingham, AL), and normal preimmune rabbit serum (Proteogenix, Oberhausbergen, France).

To obtain anti–HLA-DQα2 and HLA-DQβ2 mAbs, peptides from HLA-DQα2 (MAVGKHTLEFMM) and HLA-DQβ2 (CRDFPKDFLVQFKG) were synthesized, coupled to keyhole limpet hemocyanin (Proteogenix, Oberhausbergen, France) and used in our laboratory to generate mouse anti-peptide mAbs. Hybridomas were screened by ELISA using peptide coated plates, by Western blotting using protein extracts of stably transfected M10 cells expressing V5-tagged HLA-DQα2 or HA-tagged HLA-DQβ2 chains, and by immunofluorescence staining of these cells. Two anti–HLA-DQβ2 mAbs were obtained, V31 (IgM) and I48 (IgG1). As shown in the Supplemental Fig. 2, I48 stains the endoplasmic reticulum (ER) of transfected cells expressing HLA-DQβ2 chains, whereas V31 can be used for Western blot analyses.

Cell lines

Homozygous HLA-DR, DQA1, DQB1, and DPA1 cell lines (IMGT/HLA Database identification numbers: 11628, 10472, 11004, 11625, 11398, 11069, 10557, 11555, 10751, 10536, 10902, 11096, 11572, 11587, 10546), each having a different HLA-DQA1 and -DQB1 haplotype, were obtained from the International Histocompatibility Workshop; their complete HLA haplotypes are available on the IMGT/HLA Database Web site (http://www.ebi.ac.uk/imgt/hla/) (14, 15). The HLA-II+ melanoma cell line M10 (HLA-DQA1*03:01:01,*03:03; HLA-DQB1*03:02:01) has been described previously (16); HeLa and T2 cells (ATCC no. CCL-2 and CRL-1992) were obtained from LGC Standards (Molsheim, France), and HeLa-Ii (17) and T2-DR3 cells (18) were provided by P. Cresswell (Yale University School of Medicine).

The 781.82 cell line has a large deletion of the HLA-I and HLA-II regions on one chromosome and a deletion of the HLA-DR, -DQA1, and -DQB1 genes on the other one (19, 20). All cell lines were grown in RPMI 1640 supplemented with 10% FCS, penicillin, and streptomycin, except for HeLa cells that were grown in DMEM (all media from Invitrogen, Fisher Scientific, Illkirch Graffenstaden, France). The absence of HLA-DQA2 and -DQB2 expression in M10, T2, and 721.82 cell lines was checked by RT-PCR (no amplification product after 30 cycles).

Oligonucleotides

Oligonucleotides (sequences available on request) were purchased from MWG (Ebersberg, Germany).

DNA extraction and HLA gene exon analysis

The genomic DNAs of cell lines were extracted using Nucleospin kit (Macherey-Nagel, Hoerdt, France). To investigate the polymorphism of the HLA-DQA2/DQB2 genes, exons 2 and 3 of both genes were amplified according to the following protocol: one cycle (2 min at 94°C), 25 cycles (15 s at 94°C, 30 s at 56°C, 1 min at 72°C) and one cycle (7 min at 72°C). The sequences were checked by DNA sequencing using a BigDye Terminator kit (Roche Applied Science), followed by analysis on an ABI3130 sequencer (Applied Biosystems, Courtaboeuf, France).

RNA extraction and RT-PCR

RNAs were extracted from at least 106 cells using an RNeasy Plus mini kit (Qiagen, Courtaboeuf, France) and reverse transcribed with oligo-dT primers and SuperScript III reverse transcriptase (Invitrogen). PCR was performed with 100 ng of reverse-transcribed RNA and Pwo DNA polymerase (Roche Applied Sciences). Amplification was performed according to the following protocol: one cycle (2 min at 94°C), 20 cycles for actin, and 30 cycles for the other genes (15 s at 94°C, 30 s at 56°C, 1 min at 72°C) and one cycle (7 min 72°C). Amplicons were purified using a High PCR Product Purification kit (Roche Applied Science), A-tailed with Taq polymerase (Invitrogen), and cloned into a pDrive vector using a Qiagen PCR Cloning kit (Qiagen) or checked by direct sequencing.

Cell preparations

Highly enriched sLC preparations were obtained as described previously (21). Fresh normal human skin samples from patients undergoing abdominal plastic surgery were incubated with trypsin to separate the epidermal sheets, which were then incubated at 37°C in the presence of DNAse I (Sigma, Saint Quentin Fallavier, France) to produce single-cell suspensions. The cells were enriched by density gradient centrifugation on Ficoll (lymphocyte separation medium; Eurobio, Courtaboeuf, France) and subjected to MACS with anti-CD1a beads (Miltenyi Biotec). When tested, the maturation of LCs was induced by 24-h treatment of the cells with 25 μg/ml low m.w. polyinosinic-polycytidylic acid [poly(I:C)] (Invivogen-Cayla, Toulouse, France).

LCs generated in vitro (CD34+LCs) were obtained as described previously (22, 23). CD34+ precursors purified from umbilical cord blood samples using MACS (Miltenyi Biotec) were cultured in the presence of GM-CSF, TNF-α, stem cell factor, and FLT3 ligand for 4 d, after which TGF-β1 was added. CD207+ LCs were isolated by MACS (Miltenyi Biotec). The sLC and CD34+LC preparations were checked by flow cytometry for CD1a or CD207 expression.

MonoDCs were differentiated from elutriated monocytes in the presence of GM-CSF and IL-4, as described previously (12). RNAs from plasmacytoid and CD1c+ blood DCs were purchased from 3H Biomedicals (Uppsala, Sweden).

B cells were purified from PBMCs using CD19 Dynabeads (Invitrogen). Macrophages were differentiated from elutriated monocytes by culture for 5 d in the presence of 250 ng/ml M-CSF (Peprotech).

Cell line transfection and transduction

For expression studies in human cells, cDNAs were cloned in the plasmid pEGFP-N3 or pFuse (InvivoGen). The 3′ ends of cDNAs expressing V5 and HA tags were obtained by partial gene synthesis (Epoch Life Science, Sugar Land, TX) and all cloned cDNAs were verified by sequencing. For control experiments, the HLA-DQα1*03:03 and -DQβ1*03:02:01 chains were fused to V5 and HA tags, respectively. HeLa and M10 cells were transfected using Fugene 6 reagent (Roche Applied Science) while T2 cells were electroporated. Alternatively, HLA-DQA2 and DQB2 cDNAs were cloned into a pTRIPZ-derived lentivirus, followed by an internal ribosome entry site of the encephalomyocarditis virus and then a blasticidin or puromycin resistance gene, respectively. These lentiviral vectors were used to transduce 721.82 and HeLa cells.

Immunofluorescence techniques

For flow cytometry analyses, cell surface and intracellular staining were performed by standard techniques (13), after blocking of the Ig receptors of LCs and monoDCs with a human FcR blocking reagent (Miltenyi Biotec). Data were acquired on a Gallios flow cytometer and further analyzed with Kaluza software (Beckman-Coulter).

For confocal microscopy, adherent cells (M10, HeLa) were cultured overnight on a 12-mm–diameter glass coverslip, whereas LCs and monoDCs were allowed to adhere to poly-l-lysine precoated coverslips for 30 min in serum-free culture medium. The cells were then fixed in 3.2% paraformaldehyde, quenched with PBS containing 100 mM glycine, permeabilized with 0.05% saponin, and stained as described previously (13). The preparations were mounted in Mowiol (Citifluor, Biovalley, Marne-la-Vallée, France) and images were obtained with the HCX PL APO λ blue 63.0 × 1.40 objective of a Leica SP5 AOBS confocal microscope (Leica Microsystems, Heidelberg, Germany).

Immunoprecipitations and Western blot analyses

Washed cells were lysed at 20 × 106/ml in Triton X-100 lysis buffer (20 mM Tris pH 8, 150 mM NaCl, 5 mM EDTA, and 1% Triton X-100) or digitonin lysis buffer (50 mM sodium acetate pH 5, 150 mM NaCl, 1% digitonin) containing a protease inhibitor mixture (Complete Mini; Roche Applied Science) for 20 min on ice. Immunoprecipitations were performed as described (13). After removal of the nuclei, the lysates were precleared by addition of isotype matched control Abs and adsorption on 50 μl protein G or A-Sepharose (GE Healthcare, Vélizy, France). The supernatants were incubated with the appropriate Abs, and immune complexes were recovered by adsorption on protein G or A-Sepharose. Immunoadsorbed proteins, or soluble protein extracts from Triton X-100 lysates, were denatured and treated or not for 60 min at 37°C with 1000 U endoglycosidase H (EndoH) or 500 U PNGase F, according to the manufacturer’s recommendations (New England Biolabs, Beverly, CA). The samples were then denatured in Laemmli buffer, separated on 12.5% SDS-PAGE gels, and transferred to 0.2-μm nitrocellulose membranes (BioRad, Marnes la Coquette, France). The membranes were blocked, stained, and washed as described previously (24). Chemiluminescent revelation was performed with the SuperSignal West Pico substrate (Pierce, Fischer Scientific) and Hyperfilm ECL films (GE Healthcare).

Lymphocyte stimulation assays

Freshly isolated PBMCs from healthy donors were first depleted of B cells, monocytes, and HLA-II positive-activated T cells by negative magnetic sorting using a mixture of anti-CD19, anti-CD14, anti-pan–HLA-II and anti–HLA-DR mAbs and anti-IgG Dynabeads (Invitrogen). Purified HLA-II negative T cells were labeled with CFSE according to the manufacturer’s instructions (Invitrogen), and 1 × 105 cells were incubated at a 10:1 ratio with mitomycin C-treated transfected cells expressing HLA-DQα2/β2 molecules, in the presence of superantigen, in 200 μL RPMI supplemented with 10% heat-inactivated human serum (PAA, Les Mureaux, France). After 3 d at 37°C, the T lymphocytes were labeled with an allophycocyanin-conjugated anti-CD3 mAb and analyzed by flow cytometry. Highly purified endotoxin-reduced (<5 EU/mg) staphylococcal enterotoxin B (SEB) was purchased from Toxin Technology (Sarasota, FL).

Ethics

Biological samples were obtained with written informed consent and the protocols were approved by the local ethics committee (Comité de Protection des Personnes Est IV).

Results

The HLA-DQA2 and HLA-DQB2 genes are transcribed in LCs

To check that the HLA-DQB2 gene is expressed in LCs, RT-PCR experiments were performed using RNA from sLCs, CD34+LCs, monoDCs, B lymphocytes, or macrophages derived from monocytes using M-CSF. Although no RT-PCR product could be obtained from RNA of monoDCs, B lymphocytes, or macrophages, two fragments were amplified from LC samples, the longer being the most abundant (Fig. 1A). Partial sequencing of the two fragments showed that they corresponded to HLA-DQB2 transcripts. Complete HLA-DQB2 cDNAs were then generated from three independent samples and characterized. Three forms were identified, one including the six exons present in the gene and potentially coding for a full-length transmembrane β-chain. The two others lacked exon 4, encoding the transmembrane domain, or exons 4 and 5, and thus could be translated into soluble proteins. Analysis of the full-length HLA-DQB2 cDNAs derived from different donors revealed nucleotide polymorphisms, some of them resulting in amino acid substitutions (Fig. 1B). To better assess the HLA-DQB2 polymorphism, exons 2 and 3 from 15 cell lines homozygous for different HLA-DR, DQA1 and DQB1 alleles were amplified by PCR and directly sequenced. We observed no additional polymorphisms.

FIGURE 1.
FIGURE 1.

HLA-DQA2 and HLA-DQB2 are transcribed in LCs. (A) The expression of the HLA-DQA2 and -DQB2 genes in skin LCs (sLC), CD34+ LCs (CD34LC), monoDCs (mDCs), B lymphocytes (B) and macrophages (M) was analyzed by RT-PCR (30 cycles). As a control, the expression of actin (Act) was checked (20 cycles). (B) The polymorphism of amino acids in the HLA-DQβ2 chain was deduced from the analysis of independent cDNA clones. Putative initiator methionines are indicated. Dashes denote the absence of a coding sequence, owing to the absence of an in-frame upstream AUG codon, and dots indicate amino acid entities. The amino acids are numbered relative to the first residue, after cleavage of the signal peptide (SP). (C) The expression of the HLA-DQA2 and -DQB2 genes in CD1c+ DCs and plasmacytoid blood DCs was studied by RT-PCR (+; 30 cycles). As controls, the expression of actin (Act, 20 cycles), BDCA2, CD1C (30 cycles) and HLA-DRA (25 cycles) was also analyzed. −, Control without reverse transcriptase; TM, transmembrane part.

The possible coexpression of HLA-DQA2 was then investigated. No RT-PCR product was generated from monoDCs, but two fragments were amplified from LCs (Fig. 1A). Partial sequencing of the fragments showed that they corresponded to HLA-DQA2 transcripts. Complete HLA-DQA2 cDNAs were then generated by RT-PCR, cloned, and sequenced. The cloned cDNAs differed by the presence or absence of exon 4, which encodes the transmembrane and cytoplasmic domains, and thus potentially encoded a transmembrane or a soluble protein, respectively. Analysis of three independent samples identified two synonymous allelic polymorphisms. The polymorphism of HLA-DQA2 was also evaluated by analyzing exons 2 and 3 of the gene in the 15 IHW cell lines, by direct sequencing of PCR fragments. A total of four nucleotide polymorphisms were found in exon 2 and one in exon 3. One in exon 2 resulted in a T41M amino-acid substitution, whereas the others were synonymous.

To assess the specificity of this expression to LCs, among DC subsets, additional RT-PCR experiments were performed using RNA from CD1c+ blood DCs, plasmacytoid DCs, B lymphocytes, and macrophages. HLA-DQA2 and -DQB2 transcripts were detected only in LCs (Fig. 1C), suggesting that expression of the HLA-DQA2/DQB2 genes is strongly restricted to LCs.

HLA-DQα2 and HLA-DQβ2 chains form heterodimers which egress from the ER

To determine their cellular and biochemical properties, the HLA-DQα2 and -DQβ2 chains were expressed in the M10 melanoma cell line, respectively fused to a V5 or an HA C-terminal peptide tag. For comparison purposes, the HLA-DQα1 and -DQβ1 chains endogenous to the M10 cell line were also expressed, fused to the V5 or HA tag, respectively. Immunofluorescence revealed that when separately expressed, the tagged α and β-chains displayed a predominant ER localization (Fig. 2Aa–d). In contrast, when coexpressed, the tagged HLA-DQα1/-DQβ1 and HLA-DQα2/-DQβ2 pairs of chains colocalized in vesicular compartments, pointing to an endosomal localization (Fig. 2Ae–j). These observations strongly suggest that in M10 cells, for both forms of HLA-DQ, transgenic α− and β-chains remain unpaired in the absence of their molecular partner because of their overexpression and are consequently mostly retained in the ER. Conversely, coexpression of α1/β1 or α2/β2 chains would allow the assembly of significant amounts of heterodimers, which subsequently leave the ER.

FIGURE 2.
FIGURE 2.

Coexpressed HLA-DQα2 and -DQβ2 chains form heterodimers which egress from the ER. (A) Stably transfected M10 cells expressing DQα1-V5 (a), DQβ1-HA (b), DQα2-V5 (c), DQβ2-HA (d), DQα1-V5 and DQβ1-HA (e–g) or DQα2-V5 and DQβ2-HA (h–j) were indirectly immunolabeled with anti-tag mAbs and analyzed by confocal microscopy. Anti-V5 (A488) (a, c, e, h); anti-HA (red-FP 547; b, d, f, i), merged (g, j). Scale bars, 25 μM. (B) Stably transfected M10 cells coexpressing HLA–DQα2-V5 and HLA–DQβ2-HA (α2β2) were lysed in 1% TX100, and proteins were immunoprecipitated (IP) with anti-HA or anti-V5 mAbs. Immunoadsorbed proteins were treated (+) or not (−) with endoglycosidases PNGase F (F) or EndoH (H) and analyzed by Western blotting using anti-tag mAbs (WB). IP/WB indicates the combination of Abs used for immunoprecipitation and then Western blotting. R and S denote EndoH-resistant and -sensitive species, respectively. Lysates of untransfected cells (WT) and equivalent amounts of protein from lysates of transfected cells expressing only HLA-DQα2 or -DQβ2 chains (1:1 pool; α2 + β2) were processed under the same conditions. The position of the 30-kDa molecular mass standard is indicated on the left side of the Western blots. Asterisks identify a recurrent nonspecific 29-kDa band in PNGase F-treated samples. (C) HLA-DQα2 (upper panel) or -DQβ2 (lower panel) molecules were immunoprecipitated from solubilized extracts of stably transfected M10 cells expressing only one chain (M10α2 or M10β2) and processed as described in (B). (D) Transfected M10 cells expressing HLA–DQα2-V5 or HLA–DQβ2-HA alone or together were lysed in 1% TX100, and proteins were immunoprecipitated with the mAb B-K27. Immunoadsorbed proteins were treated (+) or not (−) with PNGase F (F) or EndoH (H) and analyzed by Western blotting (WB) with anti-tag mAbs. R, S, and asterisks are as in (B).

To confirm that HLA-DQα2 and -DQβ2 chains associate together, immunoprecipitations were performed with antitag mAbs and the immunoprecipitated proteins were treated with endoglycosidases and analyzed by Western blotting. These experiments showed that HLA-DQα2 coprecipitated with HLA-DQβ2 and reciprocally (Fig. 2B, lanes 6 and 14). This interaction was not a methodological artifact because no coprecipitation was detected when lysates from cells expressing only HLA-DQα2 or HLA-DQβ2 (M10α2 and M10β2 cells) were pooled and processed under the same conditions (Fig. 2B, lanes 5 and 13). Endoglycosidase digestion revealed that some of the HLA-DQα2 and -DQβ2 chains engaged in heterodimeric complexes had EndoH-resistant glycosylations (Fig. 2B, lanes 8 and 16), attesting to their egress from the ER. EndoH-resistant HLA-DQβ2 chains were not detected in DQα2-negative M10β2 cells, demonstrating that DQβ2 chains are retained in the ER in the absence of HLA-DQα2 proteins (Fig. 2C, lane 8). Conversely, EndoH-resistant HLA-DQα2 chains were readily detected in DQβ2-negative M10α2 cells (Fig. 2C, lane 4), indicating that some HLA-DQα2 molecules leave the ER even in the absence of the HLA-DQβ2 protein.

In view of the strong homology between the HLA-DQA1/DQB1 and HLA-DQA2/DQB2 genes, we checked whether the monomorphic anti–HLA-DQ mAb B-K27 recognizes HLA-DQα2/β2 heterodimers. HLA-DQα2 and -DQβ2 subunits were transiently expressed, alone or together, in the HLA-II negative cell line HeLa and then analyzed by immunofluorescence. B-K27 stained cells coexpressing HLA-DQα2 and -DQβ2 molecules, but not cells expressing only one chain (Supplemental Fig. 3, left panel, first two rows, right panel, first row), showing that it recognizes HLA-DQα2/β2 heterodimers.

The HLA-DQα2/-DQβ2 heterodimers recognized by the mAb B-K27 were then biochemically analyzed using transfected M10 cells expressing peptide-tagged chains. Most of the HLA-DQα2 and -DQβ2 chains immunoprecipitated by B-K27 were EndoH-resistant (Fig. 2D, lanes 5–7 and 10–12), suggesting that this mAb predominantly recognizes HLA-DQα2 and -DQβ2 molecules that have already left the ER. Whereas no HLA-DQβ2 molecules were immunoprecipitated from soluble extracts of M10β2 cells (Fig. 2D, lane 9), EndoH-resistant HLA-DQα2 subunits were recovered from lysates of M10α2 cells (Fig. 2D, lanes 2–4), pointing to the possible formation of mixed heterodimers between HLA-DQα2 and endogenously expressed HLA-DQβ1 chains.

We investigated whether HLA-DQα2 and -DQβ2 molecules associate with HLA-DQβ1 and -DQα1 chains, respectively. HeLa cells were transfected to transiently express V5- and HA-tagged chains, alone or in combinations, and then stained with the mAb B-K27 and analyzed by immunofluorescence microscopy. Cells expressing α1/β1, α2/β2, and α2/β1 pairs, but not the α1/β2 combination, were stained by the mAb (Supplemental Fig. 3, right panels), suggesting that HLA-DQα2 and -DQβ1, but not HLA-DQα1 and -DQβ2, efficiently associate to form mixed heterodimers.

The ability of DQα- and DQβ-chains to form mixed heterodimers and subsequently leave the ER was then explored in stably transfected M10 cells expressing tagged α2/β1 or α1/β2 combinations (M10α2β1 and M10α1β2 cells). Biochemical experiments showed that some of the HLA-DQα2 and -DQβ1 molecules formed mixed heterodimers, which were mostly EndoH resistant and thus had left the ER (Fig. 3, lanes 12–14 and 26–28). In contrast to what had been observed in immunofluorescence experiments with HeLa cells (Supplemental Fig. 3), tagged HLA-DQα1 and -DQβ2 chains also coimmunoprecipitated in M10 cells (Fig. 3, lanes 9–11 and 23–25). However, the majority of the coprecipitated molecules were EndoH sensitive (Fig. 3, lanes 11 and 25), suggesting a strongly impaired ability of these mixed HLA-DQα1/β2 heterodimers to egress from the ER. The discrepancies between the HeLa and M10 models could result from 1) the absence of the Ii in the former, whose presence could facilitate the association between HLA-DQα1 and HLA-DQβ2, 2) the nonrecognition of HLA-DQα1/β2 heterodimers by the mAb B-K27, or 3) differences in sensitivity between the two assays.

FIGURE 3.
FIGURE 3.

Assembly and biochemical maturation of HLA-DQα1/β2 and HLA-DQα2/β1 mixed heterodimers. Stably transfected M10 cells expressing HLA–DQα2-V5 and HLA–DQβ1-HA (α2β1) or HLA–DQα1-V5 and HLA–DQβ2-HA (α1β2) were lysed, and proteins were immunoprecipitated with anti-tag mAbs. Immunoadsorbed proteins were treated (+) or not (−) with PNGase F (F) or EndoH (H) and analyzed by Western blotting using anti-HA or -V5 mAbs.

In conclusion, our results showed that HLA-DQα2 and -DQβ2 subunits associate to form heterodimers that egress from the ER. Mixed HLA-DQα2/β1 and HLA-DQα1/β2 heterodimers were also observed, although they differed remarkably in their ability to leave the ER.

HLA-DQα2/β2 molecules associate with the Ii to reach endosomal compartments

To determine whether the HLA-II chaperone Ii assists in the transport of HLA-DQα2/β2 complexes, the two subunits were expressed in HeLa cells transfected or not with the p33 isoform of Ii, and the cellular distribution of HLA-DQα2/β2 molecules was studied by confocal microscopy using the mAb B-K27. In the absence of Ii, HLA-DQα2/β2 molecules displayed an ER distribution (Fig. 4A, upper panels), whereas when coexpressed with Ii they reached vesicular, CD63+ late endosomal compartments (Fig. 4A, lower panels). The presence of HLA-DQα2/β2 molecules in the endosomal pathway was confirmed in the M10α2β2 cell model by confocal immunofluorescence microscopy. Colocalization of tagged HLA-DQα2 and -DQβ2 subunits with different markers of the endo-lysosomal pathway (i.e., EEA1 for sorting endosomes, CD63 for late endosomes, HLA-DR for MIIC, LAMP1 and LAMP2 for lysosomes) was observed, although these events appeared to be rather infrequent, with the result that in transfected M10 cells, the precise identity of the HLA-DQα2/β2+ vesicular compartments remains to be ascertained (Supplemental Fig. 4).

FIGURE 4.
FIGURE 4.

Ii controls the transport of HLA-DQα2/β2 heterodimers to CD63+ endosomal compartments. (A) HeLa cells expressing or not Ii were transduced to stably express untagged HLA-DQα2 and -DQβ2 chains. Fixed and permeabilized cells were indirectly stained with the anti-DQ mAb B-K27 (Alexa488) and the anti-CD63 mAb H5C6 (Cy3) and analyzed by confocal microscopy (arrows indicate colocalization events). Scale bars, 25 μM. (B) TX100-solubilized membrane extracts of M10α2β2 cells were immunoprecipitated (IP) with anti-Ii rabbit serum (Ii) or normal rabbit serum (NRS). Immunoadsorbed HLA-DQα2–V5 and -DQβ2–HA chains were treated (+) or not (−) with PNGase F (F) or EndoH (H), separated by SDS-PAGE and revealed by Western blotting with anti-V5 (upper panel) or anti-HA (lower panel) mAbs. Symbols are as in Fig. 2B. (C) Immunoprecipitations and Western blot analyses were performed on M10α2 and M10β2 cell extracts as in B, but without endoglycosidase treatment. Western blotting with the anti-Ii mAb Pin.1 (lower panel) was performed to control the immunoprecipitation efficiencies. The position of the IgG L chain is indicated by an asterisk.

The association of HLA-DQα2/β2 dimers with Ii was biochemically confirmed (Fig. 4B, lanes 3–5 and 8–10). Not only did the two chains coimmunoprecipitate with Ii, but in addition a significant enrichment in mature glycosylated HLA-DQα2 and -DQβ2 molecules was observed (Fig. 4B, lanes 5 and 10), demonstrating that their interaction with Ii occurs not only in the ER but also in post-ER compartments. As expected, no association between HLA-DQβ2 and Ii was detected in M10β2 cells (Fig. 4C, lane 11), whereas HLA-DQα2 molecules were coimmunoprecipitated with Ii in M10α2 cells (Fig. 4C, lane 6). This finding is consistent with the view that in M10 cells HLA-DQα2 molecules can leave the ER, probably in association with endogenous HLA-DQβ1 chains.

To conclude, HLA-DQα2/β2 heterodimers associate with Ii and this association is critical for their transport to the endosomal network.

HLA-DQα2 and -DQβ2 subunits interact with HLA-DM

We next checked that HLA-DQα2 and -DQβ2 molecules interact with the endosomal chaperone HLA-DM. M10α2β2 cells were lysed in a buffer containing 1% digitonin at pH 5, which conserves the association between HLA-DM and HLA-DR (25). Western blot analyses revealed that HLA-DMα-chains coimmunoprecipitated with HLA-DQα2 and -DQβ2 chains (Fig. 5, lanes 2–4 and 6–8). A significant proportion of the HLA-DMα molecules associated with HLA-DQα2 subunits were EndoH resistant (Fig. 5, lane 4), in agreement with a biologic interaction in acidic endosomal compartments. In contrast, a vast majority of the HLA-DMα molecules associated with HLA-DQβ2 subunits were EndoH sensitive (Fig. 5, lane 8), which may indicate a significant interaction with ER-localized unpaired HLA-DQβ2 chains, as already described for unpaired DRβ molecules (26). EndoH-resistant HLA-DMα molecules were nevertheless also found in association with HLA-DQβ2 chains. These results strongly suggest that HLA-DQα2/β2 molecules interact with HLA-DM in acidic post-Golgi compartments.

FIGURE 5.
FIGURE 5.

Interaction of HLA-DQα2/β2 molecules with HLA-DM. Soluble membrane extracts of M10α2β2 cells were prepared in 1% digitonin at pH 5. HLA-DQα2 and -DQβ2 chains were immunoprecipitated (IP) with anti-V5 and anti-HA mAbs, respectively. Coimmunoprecipitated proteins were treated (+) or not (-) with PNGase F (F) or EndoH (H) and analyzed by Western blotting (WB) using the anti-HLA-DMα mAb Tal18.1. Symbols are as in Fig. 2B.

HLA-DQα2/β2 heterodimers are expressed on the plasma membrane and present superantigens to T lymphocytes

We then investigated whether HLA-DQα2/β2 molecules are expressed on the cell surface, which would be an important clue supporting an Ag-presenting function of these proteins. To overcome the lack of a mAb specific for HLA-DQα2/β2 complexes, we expressed these molecules in HLA-II–negative T2 cells and in HLA-DR– and HLA-DQα1/β1–negative but HLA-DP–positive 721.82 cells. Flow cytometric analysis of cells stained with the mAb B-K27 demonstrated that HLA-DQα2/β2 complexes were expressed on the plasma membrane of cells coexpressing HLA-DQα2 and -DQβ2 chains (Fig. 6A).

FIGURE 6.
FIGURE 6.

HLA-DQα2/β2 complexes are expressed on the plasma membrane and mediate superantigen activation of T cells. (A) HLA-DR− and HLA-DQα1−/DQβ1− 721.82 cells and HLA-II− T2 cells were transduced or transfected to stably express untagged (721.82 cells) or V5-tagged (T2 cells) HLA-DQα2 chains (α2), untagged HLA-DQβ2 chains (β2), or both chains (α2β2). Cell surface expression was quantified by flow cytometric analysis of nonpermeabilized cells stained with the monomorphic anti-DQ mAb B-K27. Light gray, isotype control; dark gray, B-K27 staining; black, B-K27 staining of parental cells (right panels). (B) CFSE-labeled HLA-II negative T cells (1 × 105) were cocultured with mitomycin C-treated T2 cells (1 × 104), transfected or not to express HLA-DQα2-V5/DQβ2 (T2DQ) or HLA-DR3 (T2DR), in the presence or absence of SEB (100 ng/ml). After 3 d, the T cells were stained with an allophycocyanin-conjugated anti-CD3 mAb and analyzed by flow cytometry. Three independent experiments were performed with lymphocytes from different donors, and the results of one representative experiment are shown.

To determine whether the expression of HLA-DQα2/β2 on the cell surface might lead to T cell stimulation, we evaluated the ability of the complexes to mediate bacterial superantigen activation of T cells. Purified HLA-II–negative T cells from healthy donors were labeled with CFSE and cocultured, in the presence or absence of the staphylococcal superantigen SEB, with mitomycin C-treated HLA-DQα2/β2– or HLA-DR3–expressing T2 cells or untransfected T2 cells. T cell proliferation was observed in the presence of superantigen and APCs expressing HLA-DQα2/β2 or HLA-DR3 (Fig. 6B, lower panels), but not in the absence of superantigens or HLA-II molecules (Fig. 6B, upper panels). These results demonstrate that, like HLA-DR molecules, HLA-DQα2/β2 molecules mediate superantigen activation of T cells.

Proteins encoded by the HLA-DQB2 gene are present in LCs

Two mAbs (V31 and I48) specific for the HLA-DQβ2 chain (Supplemental Fig. 2) were used to check for the presence of HLA-DQβ2 chains in sLCs by Western blotting and immunofluorescence, respectively.

Proteins displaying molecular weights corresponding to those of the transmembrane and soluble forms of HLA-DQβ2 were detected in sLC but not monoDC extracts (Fig. 7A). Remarkably, some of the HLA-DQβ2 molecules were EndoH resistant, indicating that they had left the ER of sLCs (Fig. 7B).

FIGURE 7.
FIGURE 7.

LCs express the HLA-DQβ2 chain. (A) Soluble membrane extracts of two preparations of skin LCs or monoDCs were separated by SDS-PAGE and analyzed by Western blotting using the anti–HLA-DQβ2 mAb V31, or an anti-calnexin (CNX) or anti–HLA-DRα mAb. Stably transfected M10 cells expressing HA-tagged soluble (sβHA) or transmembrane (βHA) HLA-DQβ2 or untagged HLA-DQβ2 (β) molecules and untransfected (−) M10 cells were used as positive and negative controls. The asterisk indicates a nonspecific band migrating slightly below the specific HLA-DQβ2 band. (B) Equal amounts of LC and M10β2 membrane extracts (2.5 μg) were treated (+) or not (−) with PNGase F (F) or EndoH (H) and processed as in (A). R and S indicate EndoH-resistant and -sensitive species, respectively. Asterisks identify a recurrent nonspecific 29-kDa band in PNGase F-treated samples. (C) Fixed and permeabilized sLCs were stained with I48 (revealed with Cy3-conjugated donkey anti-mouse IgG, red) and then counterstained with either anti-EEa1 (revealed with A488-conjugated donkey anti-goat IgG, green) or A488-conjugated anti-CD207 (green; preceded by a blocking step with 1% mouse serum). MonoDCs were used as a control and processed in the same way as LCs. (D) Soluble membrane extracts of freshly isolated immature skin LCs (iLC), or mature LCs (mLC) obtained by incubating the cells for 24 h with poly(I:C), were separated by SDS-PAGE and analyzed by Western blotting using the anti–HLA-DQβ2 mAb V31, or anti-calnexin (CNX; 3 μg protein extract per lane). Stably transfected M10 cells expressing HA-tagged soluble (sβH) or transmembrane (βH) HLA-DQβ2 and untransfected (−) M10 cells were used as positive and negative controls. (E) LCs from the same experiment as in (D) were stained with the mAb I48 (revealed with Cy3-conjugated secondary Abs) and then with A488-conjugated anti–HLA-DR mAb, before analysis by confocal microscopy.

Immunofluorescence analyses of cells stained with the mAb I48 confirmed that HLA-DQβ2 proteins were expressed in CD207+ sLCs, but not in monoDCs, and revealed a predominant vesicular distribution (Fig. 7C). Colocalization experiments with an anti-EEA1 Ab identified these vesicular structures as sorting early endosomes (Fig. 7C).

Finally, we checked whether the expression of HLA-DQβ2 molecules was modulated during sLC maturation. Freshly isolated sLCs, or sLCs cultured for 24 h in the presence of the TLR3 agonist poly(I:C) to induce their maturation, were processed for Western blotting and confocal microscopy. Western blot analyses using the mAb V31 showed that the quantity of HLA-DQβ2 chains remained stable after maturation (Fig. 7D), whereas confocal microscopy revealed that the mAb I48 stained only immature sLCs (Fig. 7E). As expected, control staining with the mAb L243 showed that HLA-DR molecules were present in the intracellular compartments of freshly isolated LCs, but were redistributed on the plasma membrane of cells cultured in the presence of poly(I:C), confirming that the freshly isolated sLCs were immature and became mature after culture.

Discussion

This work constitutes the first demonstration, to our knowledge, that the HLA-DQA2 and -DQB2 genes are constitutively coexpressed in physiologic cells, namely in epidermal LCs. Moreover, this expression is shown to be highly specific to LCs, because other members of the DC family such as immature monoDCs, a model of interstitial myeloid DCs, and CD1c+ blood myeloid DCs or plasmacytoid DCs do not express these two genes. Induction of the maturation of monoDCs with LPS failed to induce transcription of the two genes (data not shown), nor did we detect any expression of HLA-DQA2 or -DQB2 either in HLA-DR–purified B lymphocytes or in macrophages differentiated from purified monocytes using M-CSF. Of course we cannot exclude that, in particular cytokine or immunopathologic environments, these genes could be expressed in cells other than LCs. Nevertheless, in vivo footprinting experiments performed in B lymphoblastoid cells revealed a lack of DNA-protein binding in the regulatory regions of the HLA-DQA2 and -DQB2 promoters (27); therefore, it is tempting to speculate that among professional APCs, only LCs express the specific transcription factors mandatory for the concomitant expression of HLA-DQA2 and -DQB2.

Use of anti-HLA-DQβ2 mAbs allowed us to confirm the presence of the β subunit in freshly isolated LCs. Interestingly, a significant fraction of the total pool of these β-chains was EndoH-resistant and therefore had egressed from the ER, strongly supporting the biologic relevance of this expression. Moreover, because HLA-DQβ2 molecules mostly remained EndoH-sensitive in the absence of their α2 partner in M10β2 and M10α1β2 transfected cells (Figs. 2C, 3B, 7B), these results indirectly support the view that HLA-DQα2 chains are also present in LCs and associate with HLA-DQβ2 molecules to form biologic heterodimers. Consistent with these biochemical observations, confocal immunofluorescence analyses with the mAb I48 indicated that HLA-DQβ2 subunits reach the early endosomal compartment of LCs. In contrast, in transfected M10 cells the I48 staining revealed a typical ER localization. This difference between transfected cells and LCs might result from the presence of particular ligands, specific to epidermal LCs, which could confer a better accessibility of the epitope recognized by I48. Although Western blot analyses using the mAb V31 attested that HLA-DQβ2 molecules were present at similar levels in immature and mature LCs, these molecules could be detected by immunofluorescence with the mAb I48 only in immature cells. Such paradoxic data suggest that the 13-amino acid N-terminal part of the HLA-DQβ2 chain, which is targeted by the two mAbs, is engaged in different molecular interactions in immature and mature LCs. Finally, we were unable to test whether HLA-DQα2/β2 complexes were expressed on the plasma membrane of LCs because in flow cytometry experiments, the mAb I48 failed to stain plasma membrane-expressed molecules, as assayed using transfected 721.81 cells (Supplemental Fig. 2C). This unusual recognition behavior of the mAb I48 will require additional investigation, and the development of an Ab specific for the heterodimer will be necessary to further study the cellular distribution of HLA-DQα2/β2 molecules in LCs.

In transfected M10 cells, tagged HLA-DQα2 and -DQβ2 transmembrane proteins formed heterodimers that associated with the Ii and left the ER to reach vesicular compartments, some of which were CD63+ or HLA-DR+ late endosomes or lysosomes. In agreement with the well-established chaperone function of Ii in the transport of classical HLA-II molecules to endosomes (28), we found that the p33 isoform of Ii also mediated the transport of HLA-DQα2/β2 heterodimers to the CD63+ compartments of HLA-II− HeLa cells. Because the interaction of classical HLA-II molecules with the endosomal chaperone HLA-DM is crucial to the building of their antigenic peptide repertoire, we investigated whether HLA-DQα2/β2 complexes interact with this nonclassical HLA-II molecule. EndoH-resistant HLA-DMα proteins did indeed coimmunoprecipitate with HLA-DQα2 and -DQβ2 subunits, as described for HLA-DR molecules (25). This interaction with HLA-DM strongly suggests that HLA-DQα2/β2 heterodimers have peptide ligands, like classical HLA-II molecules.

Cell surface expression of HLA-DQα2/β2 dimers could be demonstrated in two different HLA-DQα1– and HLA-DQβ1–negative lymphoblastoid cell lines, T2 (HLA-II−) and 721.82 (HLA-DP+). Because T2 cells, but not 721.82 cells, are HLA-DM and -DO negative, these findings indicate that the presence of the two chaperone molecules did not significantly affect the cell surface expression of HLA-DQα2/β2 complexes, at least qualitatively. This surface expression, compatible with an Ag-presenting function, prompted us to test whether HLA-DQα2/β2 molecules interact with the TCR of T lymphocytes. Using a superantigen presentation assay, we showed that HLA-DQα2/β2 transfected T2 cells induced the proliferation of T cells from healthy donors in the presence of the staphylococcal superantigen SEB, almost as efficiently as T2 cells expressing HLA-DR (T2-DR3 cells). These results further support the idea of the participation of HLA-DQα2/β2 molecules in Ag presentation.

In view of the high degree of homology between the HLA-DQA1/-DQA2 and HLA-DQB1/-DQB2 genes, we wondered whether their protein products formed HLA-DQα1/β2 and -DQα2/β1 mixed heterodimers. In agreement with this hypothesis, HLA-DQα2 chains were able to egress from the ER of M10α2 cells in the absence of their HLA-DQβ2 partner. To further explore this question, we coexpressed peptide-tagged HLA-DQβ1 and -DQα2 or HLA-DQβ2 and -DQα1 chains in M10 cells and demonstrated that HLA-DQα2/β1 mixed heterodimers could assemble and egress from the ER. This observation differs from a previous report (5). The discrepancy might arise from differences in the methods or the particular HLA-DQβ1 variants used in our work and the latter study. Moreover, the assembly of HLA-DQα1/β2 heterodimers was also observed, but these molecules were mostly EndoH sensitive, which would question their biologic relevance. Isotype-matched heterodimers resulting from trans-complementation have already been described for HLA-DR, -DQ, and -DP molecules. In addition, the formation of HLA-DRα/DQβ complexes has been reported (29). If HLA-DQ mixed heterodimers exist in LCs, they may represent an unexpected way of diversifying the repertoire of peptide Ags presented by HLA-DQ in these cells. In addition, they could also be a source of an additional level of complexity in the panel of HLA-DQ Ags to be considered in the context of allograft matching. Biochemical and functional experiments using different HLA-DQ allelic variants will be required to draw a general conclusion on this point.

Our results suggest that HLA-DQα2/β2 heterodimers represent a new type of HLA-II molecule specific to LCs. These proteins share the characteristics of conventional Ag-presenting HLA-II molecules except in regard to the polymorphism of their amino acid sequences. This remarkably poor polymorphism, which was confirmed by sequencing exons 2 and 3 of the HLA-DQA2 and -DQB2 genes in a panel of HLA-DR, DQA1, DQB1 and DPA1 homozygous IHW cell lines, is reminiscent of the CD1 molecule system; it suggests that HLA-DQα2/β2 antigenic ligands share conserved biochemical or biophysical properties, which would make their characterization crucial. It is worth noting that LCs are also present in oral and genital mucosae, and cells expressing Langerin have been described in the human lung (30), liver (31) and cornea (32). Determining whether the HLA-DQA2 and -DQB2 genes are likewise expressed in such Langerin-positive DCs residing in tissues outside the skin might provide a clue as to the function of HLA-DQα2/β2 molecules and why their expression appears to be confined to LCs.

Disclosures

The authors have no financial conflicts of interest.

Acknowledgments

We thank S. Schnautz for technical assistance.

Footnotes

  • This work was supported by grants from Association de Recherche et de Développement en Médecine et en Santé Publique (to C.L.), the International Research Group on Hematopoietic Stem Cell Transplantation, La Ligue Contre le Cancer, Etablissement Français du Sang-Alsace and Grant SFB704 from the Deutsche Forschungsgemeinschaft.

  • The online version of the article contains supplemental material.

  • Abbreviations used in this article:

    A488
    Alexa488
    CD34+LCs
    Langerhans cells differentiated in vitro from CD34+ cord blood cells
    DC
    dendritic cell
    EndoH
    endoglycosidase H
    ER
    endoplasmic reticulum
    HA
    hemagglutinin
    HLA-II
    HLA class II
    Ii
    invariant chain
    LC
    Langerhans cell
    monoDC
    monocyte-derived dendritic cell
    poly(I:C)
    polyinosinic-polycytidylic acid
    SEB
    staphylococcal enterotoxin B
    sLC
    skin Langerhans cell.

  • Received October 25, 2011.
  • Accepted February 8, 2012.

References

  1. ↵
    1. Wang R. F.
    2003. Identification of MHC class II-restricted tumor antigens recognized by CD4+ T cells. Methods 29: 227–235.
    OpenUrlCrossRefPubMed
  2. ↵
    1. Fernando M. M.,
    2. C. R. Stevens,
    3. E. C. Walsh,
    4. P. L. De Jager,
    5. P. Goyette,
    6. R. M. Plenge,
    7. T. J. Vyse,
    8. J. D. Rioux
    . 2008. Defining the role of the MHC in autoimmunity: a review and pooled analysis. PLoS Genet. 4: e1000024.
    OpenUrlCrossRefPubMed
  3. ↵
    1. Shiina T.,
    2. H. Inoko,
    3. J. K. Kulski
    . 2004. An update of the HLA genomic region, locus information and disease associations: 2004. Tissue Antigens 64: 631–649.
    OpenUrlCrossRefPubMed
  4. ↵
    1. Yu L. P.,
    2. M. J. Sheehy
    . 1991. The cryptic HLA-DQA2 (“DX alpha”) gene is expressed in human B cell lines. J. Immunol. 147: 4393–4397.
    OpenUrlAbstract
  5. ↵
    1. Rudy G. B.,
    2. A. M. Lew
    . 1997. The nonpolymorphic MHC class II isotype, HLA-DQA2, is expressed on the surface of B lymphoblastoid cells. J. Immunol. 158: 2116–2125.
    OpenUrlAbstract
  6. ↵
    1. Berdoz J.,
    2. J. M. Tiercy,
    3. P. Rollini,
    4. B. Mach,
    5. J. Gorski
    . 1989. Remarkable sequence conservation of the HLA-DQB2 locus (DX beta) within the highly polymorphic DQ subregion of the human MHC. Immunogenetics 29: 241–248.
    OpenUrlCrossRefPubMed
    1. Gyllensten U. B.,
    2. D. Lashkari,
    3. H. A. Erlich
    . 1990. Allelic diversification at the class II DQB locus of the mammalian major histocompatibility complex. Proc. Natl. Acad. Sci. USA 87: 1835–1839.
    OpenUrlAbstract/FREE Full Text
  8. ↵
    1. Rowe J. R.,
    2. M. H. Neme de Gimenez,
    3. C. A. Emler,
    4. M. J. Sheehy
    . 1990. HLA-DQA2 (DX alpha) polymorphism and insulin dependent diabetes. Hum. Immunol. 29: 256–262.
    OpenUrlCrossRefPubMed
  9. ↵
    1. Gaur L. K.,
    2. E. R. Heise,
    3. P. S. Thurtle,
    4. G. T. Nepom
    . 1992. Conservation of the HLA-DQB2 locus in nonhuman primates. J. Immunol. 149: 2530.
    OpenUrlPubMed
  10. ↵
    1. Diaz D.,
    2. M. Naegeli,
    3. R. Rodriguez,
    4. J. J. Nino-Vasquez,
    5. A. Moreno,
    6. M. E. Patarroyo,
    7. G. Pluschke,
    8. C. A. Daubenberger
    . 2000. Sequence and diversity of MHC DQA and DQB genes of the owl monkey Aotus nancymaae. Immunogenetics 51: 528–537.
    OpenUrlCrossRefPubMed
  11. ↵
    1. Merad M.,
    2. F. Ginhoux,
    3. M. Collin
    . 2008. Origin, homeostasis and function of Langerhans cells and other langerin-expressing dendritic cells. Nat. Rev. Immunol. 8: 935–947.
    OpenUrlCrossRefPubMed
  12. ↵
    1. Saudrais C.,
    2. D. Spehner,
    3. H. de la Salle,
    4. A. Bohbot,
    5. J. P. Cazenave,
    6. B. Goud,
    7. D. Hanau,
    8. J. Salamero
    . 1998. Intracellular pathway for the generation of functional MHC class II peptide complexes in immature human dendritic cells. J. Immunol. 160: 2597–2607.
    OpenUrlAbstract/FREE Full Text
  13. ↵
    1. Angenieux C.,
    2. J. Salamero,
    3. D. Fricker,
    4. J. P. Cazenave,
    5. B. Goud,
    6. D. Hanau,
    7. H. de La Salle
    . 2000. Characterization of CD1e, a third type of CD1 molecule expressed in dendritic cells. J. Biol. Chem. 275: 37757–37764.
    OpenUrlAbstract/FREE Full Text
  14. ↵
    1. Robinson J.,
    2. K. Mistry,
    3. H. McWilliam,
    4. R. Lopez,
    5. P. Parham,
    6. S. G. Marsh
    . 2011. The IMGT/HLA database. Nucleic Acids Res. 39(Database issue): D1171–D1176.
    OpenUrlAbstract/FREE Full Text
  15. ↵
    1. Robinson J.,
    2. A. Malik,
    3. P. Parham,
    4. J. G. Bodmer,
    5. S. G. Marsh
    . 2000. IMGT/HLA database—a sequence database for the human major histocompatibility complex. Tissue Antigens 55: 280–287.
    OpenUrlCrossRefPubMed
  16. ↵
    1. Mackensen A.,
    2. L. Ferradini,
    3. G. Carcelain,
    4. F. Triebel,
    5. F. Faure,
    6. S. Viel,
    7. T. Hercend
    . 1993. Evidence for in situ amplification of cytotoxic T-lymphocytes with antitumor activity in a human regressive melanoma. Cancer Res. 53: 3569–3573.
    OpenUrlAbstract/FREE Full Text
  17. ↵
    1. Marks M. S.,
    2. P. A. Roche,
    3. E. van Donselaar,
    4. L. Woodruff,
    5. P. J. Peters,
    6. J. S. Bonifacino
    . 1995. A lysosomal targeting signal in the cytoplasmic tail of the beta chain directs HLA-DM to MHC class II compartments. J. Cell Biol. 131: 351–369.
    OpenUrlAbstract/FREE Full Text
  18. ↵
    1. Denzin L. K.,
    2. N. F. Robbins,
    3. C. Carboy-Newcomb,
    4. P. Cresswell
    . 1994. Assembly and intracellular transport of HLA-DM and correction of the class II antigen-processing defect in T2 cells. Immunity 1: 595–606.
    OpenUrlCrossRefPubMed
  19. ↵
    1. Erlich H.,
    2. J. S. Lee,
    3. J. W. Petersen,
    4. T. Bugawan,
    5. R. DeMars
    . 1986. Molecular analysis of HLA class I and class II antigen loss mutants reveals a homozygous deletion of the DR, DQ, and part of the DP region: implications for class II gene order. Hum. Immunol. 16: 205–219.
    OpenUrlCrossRefPubMed
  20. ↵
    1. Spies T.,
    2. C. C. Morton,
    3. S. A. Nedospasov,
    4. W. Fiers,
    5. D. Pious,
    6. J. L. Strominger
    . 1986. Genes for the tumor necrosis factors alpha and beta are linked to the human major histocompatibility complex. Proc. Natl. Acad. Sci. USA 83: 8699–8702.
    OpenUrlAbstract/FREE Full Text
  21. ↵
    1. von Bubnoff D.,
    2. H. Bausinger,
    3. H. Matz,
    4. S. Koch,
    5. G. Häcker,
    6. O. Takikawa,
    7. T. Bieber,
    8. D. Hanau,
    9. H. de la Salle
    . 2004. Human epidermal langerhans cells express the immunoregulatory enzyme indoleamine 2,3-dioxygenase. J. Invest. Dermatol. 123: 298–304.
    OpenUrlCrossRefPubMed
  22. ↵
    1. Allam J. P.,
    2. E. Geiger,
    3. S. Kraft,
    4. T. Bieber
    . 2003. Limited reliability of E-cadherin as a specific marker for in vitro-generated Langerhans cells. Arch. Dermatol. Res. 295: 263–268.
    OpenUrlCrossRefPubMed
  23. ↵
    1. Kohl K.,
    2. S. Schnautz,
    3. M. Pesch,
    4. E. Klein,
    5. M. Aumailley,
    6. T. Bieber,
    7. S. Koch
    . 2007. Subpopulations of human dendritic cells display a distinct phenotype and bind differentially to proteins of the extracellular matrix. Eur. J. Cell Biol. 86: 719–730.
    OpenUrlCrossRefPubMed
  24. ↵
    1. Maître B.,
    2. C. Angénieux,
    3. J. Salamero,
    4. D. Hanau,
    5. D. Fricker,
    6. F. Signorino,
    7. F. Proamer,
    8. J. P. Cazenave,
    9. B. Goud,
    10. S. Tourne,
    11. H. de la Salle
    . 2008. Control of the intracellular pathway of CD1e. Traffic 9: 431–445.
    OpenUrlCrossRefPubMed
  25. ↵
    1. Sanderson F.,
    2. C. Thomas,
    3. J. Neefjes,
    4. J. Trowsdale
    . 1996. Association between HLA-DM and HLA-DR in vivo. Immunity 4: 87–96.
    OpenUrlCrossRefPubMed
  26. ↵
    1. Neumann J.,
    2. A. König,
    3. N. Koch
    . 2007. Detection of aberrant association of DM with MHC class II subunits in the absence of invariant chain. Int. Immunol. 19: 31–39.
    OpenUrlAbstract/FREE Full Text
  27. ↵
    1. Indovina P.,
    2. F. Megiorni,
    3. G. Fontemaggi,
    4. P. Coni,
    5. B. Mora,
    6. M. C. Mazzilli
    . 2001. Absence of in vivo DNA-protein interactions in the DQA2 and DQB2 promoter regions. Hum. Immunol. 62: 504–508.
    OpenUrlCrossRefPubMed
  28. ↵
    1. Lamb C. A.,
    2. J. W. Yewdell,
    3. J. R. Bennink,
    4. P. Cresswell
    . 1991. Invariant chain targets HLA class II molecules to acidic endosomes containing internalized influenza virus. Proc. Natl. Acad. Sci. USA 88: 5998–6002.
    OpenUrlAbstract/FREE Full Text
  29. ↵
    1. Lotteau V.,
    2. L. Teyton,
    3. D. Burroughs,
    4. D. Charron
    . 1987. A novel HLA class II molecule (DR alpha-DQ beta) created by mismatched isotype pairing. Nature 329: 339–341.
    OpenUrlCrossRefPubMed
  30. ↵
    1. Van Pottelberge G. R.,
    2. K. R. Bracke,
    3. I. K. Demedts,
    4. K. De Rijck,
    5. S. M. Reinartz,
    6. C. M. van Drunen,
    7. G. M. Verleden,
    8. F. E. Vermassen,
    9. G. F. Joos,
    10. G. G. Brusselle
    . 2010. Selective accumulation of langerhans-type dendritic cells in small airways of patients with COPD. Respir. Res. 11: 35.
    OpenUrlCrossRefPubMed
  31. ↵
    1. Harada K.,
    2. S. Shimoda,
    3. H. Ikeda,
    4. M. Chiba,
    5. M. Hsu,
    6. Y. Sato,
    7. M. Kobayashi,
    8. X. S. Ren,
    9. H. Ohta,
    10. S. Kasashima,
    11. et al
    . 2011. Significance of periductal Langerhans cells and biliary epithelial cell-derived macrophage inflammatory protein-3α in the pathogenesis of primary biliary cirrhosis. Liver Int. 31: 245–253.
    OpenUrlCrossRefPubMed
  32. ↵
    1. Mayer W. J.,
    2. U. M. Irschick,
    3. P. Moser,
    4. M. Wurm,
    5. H. P. Huemer,
    6. N. Romani,
    7. E. U. Irschick
    . 2007. Characterization of antigen-presenting cells in fresh and cultured human corneas using novel dendritic cell markers. Invest. Ophthalmol. Vis. Sci. 48: 4459–4467.
    OpenUrlAbstract/FREE Full Text
Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools

Jump to section