Editing of an Immunodominant Epitope of Glutamate Decarboxylase by HLA-DM

Cell lines and Ag

B-LCL Priess, JAH, and BSM are homozygous for the expression of HLA-DR4 (DRA1*0101, DRB1*0401), while B-LCL Frev and Frev.pa express HLA-DR4 and DR1 (DRA1*0101, DRB1*0401, and DRB1*0101) alleles. The line Frev.pa was isolated by subcloning and expresses lower steady state levels of intracellular DM compared with Frev (15). Retroviral transduction of the parental cell line Priess with the gene encoding the 65-kDa form of human GAD resulted in a cell line constitutively expressing this endogenous Ag, termed PriessGAD (30, 31). To obtain APC with high levels of functional DM, PriessGAD cells were transfected with cDNAs encoding DMα and DMβ (L. Denzin, Sloan Kettering, NY). Stable DM⁺ transfectants were obtained following APC selection for DMα in 0.5 μg/ml puromycin (Sigma-Aldrich, St. Louis, MO) and DMβ in 0.8 μg/ml hygromycin (Sigma-Aldrich). T2 is a B×T cell hybrid with a large deletion in the MHC class II locus and does not express endogenous class II proteins (32). The cell lines T2.DR4 and T2DR4.DM expressing functional class II DR4 or DR4 and DM have been described, and DR4 expression on all APC was determined by FACS using Ab 3.5.9-13F10 (32, 33). All APC were maintained in Iscove’s medium (Life Technologies, Gaithersburg, MD) supplemented with 10% heat-inactivated calf serum, 100 U/ml penicillin, and 100 μg/ml streptomycin. For functional assays, the following T cell hybridomas were used: 33.1, specific for GAD_273–285; 17.9, specific for human serum albumin (HSA)_64–76; and 2.18, specific for Igκ_188–203 (33, 34, 35). The T cell hybridomas used in this study recognize their cognate peptide presented in the context of HLA-DR4 and were maintained in RPMI (Life Technologies) supplemented with 10% FCS, 0.1% β-ME, 100 U/ml penicillin, and 100 μg/ml streptomycin. Control studies with the class II-negative cell line, T2, confirmed the DR4 restriction of these hybridomas.

The recombinant 65-kDa form of human GAD was produced in bacteria and was purified to homogeneity using affinity chromatography as previously published (36). GAD_273–285 (IAFTSEHSHFSLK) was synthesized using an ABI peptide synthesizer (PE Applied Biosystems, Foster City, CA) and F-moc technology. Peptide purity and structure were confirmed by reverse phase HPLC and mass spectroscopy.

Subcellular fractionation and immunoblotting

To localize intracellular GAD, HLA-DR, and HLA-DM, membrane organelles were isolated from APC as previously described (31, 37) and assayed by Western blot analysis. For each sample, equal amounts of protein were either boiled in reducing sample buffer or incubated at room temperature in nonreducing sample buffer and separated by 10% SDS-PAGE. The proteins were transferred to a nitrocellulose membrane (Micron Separations, West Brook, MA) and probed with the mAbs anti-GAD65 (Sigma-Aldrich), DM.K8 recognizing DMβ (38) (A. Vogt, Roche Institute, Basil, Switzerland), DA6.147 specific for DRα, the rabbit polyclonal anti-DMα, or a rabbit polyclonal anti-DOβ (E. Mellins, Stanford, CA). Filter-bound Abs were visualized using goat anti-mouse HRP or goat anti-rabbit HRP (Jackson ImmunoResearch Laboratories, West Grove, PA), followed by epichemiluminescence (Amersham Pharmacia Biotech, Arlington Heights, IL). Exposure times for chemiluminescence were varied to ensure accurate and quantitative signal detection.

Acid elution of peptides from surface MHC class II

APC were washed twice in ice-cold 160 mM NaCl then resuspended in 160 mM NaCl/citric acid, pH 4.0, and incubated 30 min on ice. These cell preparations were neutralized with ice-cold medium and subsequently washed twice with fresh tissue culture medium. No loss in cell viability was observed using these conditions, and APC retained their ability to activate T cells upon addition of antigenic peptides.

Ag presentation assays

For exogenous Ags, APC were incubated with or without Ag for 18 h at 37°C. GAD Ag was tested over a range of concentrations (1–50 μg/ml) or the fixed amount of 20 μg/ml, while HSA Ag was used at 250 μg/ml. Control samples were incubated for 16 h without Ag, then pulsed with 10 μM GAD_273–285 and incubated for an additional 2 h. The remainder of the assay was conducted as described previously (31). All T cell assays were performed in triplicate with the mean and SD calculated for each data point. All figures are representative of a minimum of three independent experiments.

Editing of a class II-restricted epitope from the IDDM autoantigen GAD

T cells responsive to the islet cell Ag GAD are detectable in IDDM patients as well as transgenic mice expressing disease-associated DR and DQ alleles (28, 39). Studies of human T cells responsive to GAD have proven difficult, potentially due to a number of factors, including low T cell precursor frequencies, the insolubility of this protein, and genetic variability in APC. With respect to the latter, the efficiency of HLA-DR4-restricted presentation of exogenous GAD was shown to vary widely among a panel of B-LCL, suggesting that genetic differences might guide GAD epitope display in humans and contribute to disease susceptibility (29). Here, the relative importance of the MHC-encoded peptide editor, HLA-DM, on the efficiency of GAD epitope presentation by APC was assessed using functional assays (Fig. 1⇓). Class II presentation of the epitope GAD_273–285 derived from exogenous GAD was significantly reduced in the DM-positive APC T2.DR4.DM compared with T2.DR4 cells lacking this editor (Fig. 1⇓, A and B). The enhanced display of naturally processed GAD_273–285:class II complexes on T2.DR4 was particularly striking, as nearly 80% of the cell surface class II molecules on these cells are loaded with the Ii-derived CLIP ligand and are inaccessible to the majority of peptides generated during Ag processing (8, 9). The expression of HLA-DM in T2.DR4 cells results in the editing of unstable CLIP-class II complexes (32), suggesting that GAD_273–285 may also be susceptible to DM-mediated release from class II DR4. Differences in GAD epitope presentation were not linked to cell surface class II levels on APC^+/− DM, as flow cytometry revealed greater amounts of DR4 displayed on T2.DR4.DM (mean channel fluorescence, 149.2) compared with T2.DR4 cells (mean channel fluorescence, 77.8). To confirm that both T2.DR4 and T2.DR4.DM cells can function to activate GAD-specific T cells, these APC were incubated with the synthetic GAD_273–285 peptide (Fig. 1⇓C). Indeed, HLA-DR4-restricted presentation of this synthetic peptide was readily detected using each APC, although T2.DR4.DM cells appear to present the synthetic GAD peptide better, possibly due to higher levels of surface DR4. The requirement for DM in generating accessible class II molecules through CLIP release has been widely documented in studies of a large number of Ags (12, 32, 40). Indeed, presentation of the exogenous Ag, HSA, was impaired in T2.DR4 cells compared with that in DM-positive cells (Fig. 1⇓D). These results suggest that, like CLIP, an epitope derived from GAD can be released from DR4 in the presence of DM. In addition, editing of this GAD epitope was only observed following intracellular Ag processing, as the display of surface-loaded synthetic GAD_273–285 was not diminished with increasing DM levels.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

HLA-DR4-restricted presentation of exogenous GAD is diminished in the presence of DM. T2.DR4 and T2.DR4.DM cells were incubated for 18 h in the presence 20 μg/ml GAD (A), variable concentrations of GAD (B), or HSA Ag (D). C, APC were pulsed for 2 h with increasing concentrations of synthetic GAD_273–285. Following these incubations, APC were plated with the appropriate T cell hybridomas (33.1 specific for GAD and 17.9 specific for HSA), and DR4-restricted presentation was determined as described in Materials and Methods. The data shown are representative of at least three individual experiments.

Differential DM expression among APC is well established, with these variations influencing CLIP editing and accessibility of class II molecules (16). To determine whether naturally occurring differences in DM expression can modulate GAD epitope display, B-LCL with variations in their endogenous DM levels were tested for functional epitope presentation (Fig. 2⇓A). DR4-restricted presentation of the 273–285 epitope from exogenous GAD was significantly higher in Priess cells compared with Frev, BSM, and JAH B-LCL. When the expression of DM was examined in these B-LCL, Priess cells exhibited the lowest relative level of DM expression (Fig. 2⇓, B and D). B-LCL with low or high levels of DM can be isolated by subcloning (16); thus, the cell line Frev.pa with reduced DM expression was isolated and tested for GAD peptide presentation. Again, the reduced levels of DM in Frev.pa cells correlated with increased presentation of the GAD epitope. At neutral or mildly acidic pH (>5.0), the editing function of DM can be regulated via another MHC-encoded heterodimer, HLA-DO found within the endocytic pathway of B cells (41). Steady state expression of DOβ determined for each B-LCL failed to correlate with functional studies of GAD epitope presentation (Fig. 2⇓, C and D). DO expression was nearly identical in Frev and Frev.pa cells, although differential DM levels and class II-restricted presentation of GAD were observed. These results suggest that GAD epitope editing by HLA-DM takes place in mature endosomal or prelysosomal compartments (pH <5.0) where HLA-DO cannot regulate DM function. For each B-LCL analyzed, efficient and comparable presentation of the synthetic GAD_273–285 peptide was detected (Fig. 2⇓A). These results strongly suggest that variations in DM expression levels within APC modulate the presentation of epitopes derived from the autoantigen GAD.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

HLA-DR4-restricted presentation of exogenous GAD differed among human B-LCL and was correlated with DM expression. A, APC were incubated for 18 h in the presence of GAD Ag, then cocultured with the T cell hybridoma, 33.1. HLA-DR4-restricted presentation of GAD_273–285 was quantitated as described in Materials and Methods. Relative surface expression of HLA-DR4 on each APC was determined by flow cytometry: BSM, 110.39 mean channel fluorescence (mcf); JAH, 93.21 mcf; Priess, 147.59 mcf; Frev, 115.40 mcf; Frev.pa, 332.15 mcf. B–D, Quantitation of steady state levels of HLA-DM and DO. B-LCL lysates were analyzed on 10% PAGE, transferred to membranes for Western blotting, and probed with the DMβ-specific mAb DMK.8 (B and D) or an Ab to DOβ (C and D). Steady state levels of DM and DO expression were expressed as relative intensity units (RIU) for each cell line, as determined by densitometric analysis. These results are representative of three distinct experiments.

Dissociation of GAD peptide:class II complexes at acidic pH

Reduced presentation of the GAD epitope following processing within APC expressing abundant DM suggested that this peptide forms an unstable complex with class II DR4. The low affinity of this GAD epitope for DR4 has been established (30). Previous studies have demonstrated that unstable class II ligands such as CLIP can be displaced by HLA-DM or upon exposure of such peptide:class II complexes to low pH (8). B cells that had been preloaded with synthetic GAD peptide 273–285 and then exposed to low pH, compared with APC incubated at neutral pH, exhibited a diminished capacity to activate T cells specific for this immunodominant epitope (Fig. 3⇓A). These results suggested that GAD_273–285, like CLIP, forms an unstable complex with MHC class II molecules. PriessGAD cells constitutively express high levels of intracellular GAD and present the immunodominant epitope GAD_273–285 in the context of HLA-DR4. A decrease in GAD-specific T cell activation was also observed following acidic pH incubation and aldehyde fixation of these APC expressing endogenous Ag (Fig. 3⇓, A and B). Aldehyde fixation of these APC after low pH peptide stripping was required, as live PriessGAD cells within a matter of hours, generate newly formed GAD_273–285:DR4 complexes that transit to the cell surface (Fig. 3⇓B). Acidic elution of peptides from cell surface class II molecules was also examined using an epitope known to form high affinity, stable complexes with HLA-DR4. As predicted, an epitope derived from the endogenous Ag Igκ that binds to HLA-DR4 was found to be resistant to low pH dissociation (Fig. 3⇓B). Prior studies using in vitro binding and dissociation assays have established the formation of high affinity, stable complexes of this κ peptide and DR that are not susceptible to DM (42). These latter studies suggest that acid treatment of cell surface class II dimers did not radically alter MHC structure or trigger indiscriminant release of ligands. Both the synthetic and naturally processed GAD_273–285 were susceptible to elution from DR4 under low pH conditions. These results support the formation of unstable or loosely bound GAD peptide:class II complexes in APC that may serve as a target for DM-mediated editing.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Instability of GAD epitope complexed with DR4 at very low pH. A, Priess or PriessGAD cells treated with the GAD_273–285 synthetic peptide were incubated in ice-cold neutral or acidic pH saline buffer for 30 min, followed by washing to remove released class II ligands. The cells were then paraformaldehyde fixed and cocultured with the GAD-specific T cell hybridoma 33.1 as described in Materials and Methods. B and C, In a separate experiment, PriessGAD cells were incubated in ice-cold acidic pH saline buffer for 30 min to promote unstable peptide dissociation. Samples were divided and either paraformaldehyde-fixed or untreated before coculture with the GAD-specific T cell 33.1 (B) or the Igκ-specific T cell 2.18 (C). Representative results are shown from three distinct experiments. Data are expressed as the mean of triplicate samples analyzed with the SE indicated.

Cytoplasmic-derived peptides bound to class II proteins intersect and can serve as substrates for DM

Class II MHC proteins have the capacity to present peptide ligands derived from both exogenous and endogenous Ags (43, 44). While DM functions to enhance class II accessibility for most Ags, select endogenous proteins are presented even in the absence of DM (33). Epitopes derived from these endogenous Ags may intersect peptide-receptive class II αβ complexes within specialized sites in the endo/lysosomal pathway (45). Little is known about the mechanisms regulating cytoplasmic peptide presentation by class II molecules, although studies have demonstrated that cytoplasmic processing is a prerequisite for the display of these antigenic epitopes (31). To determine whether complexes of cytoplasmic peptides and class II αβ intersect DM and undergo selective editing, the presentation of epitopes derived from cytoplasmic GAD was examined in cells expressing high or low levels of DM. PriessGAD cells expressing cytoplasmic GAD were transfected with HLA-DM and subclones isolated with increased DM levels. Western blot analysis of the cells PriessGAD and PriessGAD.DM confirmed the enhanced expression of DMα and DMβ in the latter cell line (Fig. 4⇓A). Notably, the low expression of DM in the parental PriessGAD cell line was functionally demonstrated by an abundance of class II αβ dimers exhibiting instability under nonreducing SDS-PAGE (Fig. 4⇓B). Enhanced expression of DM resulted in an increase in SDS-stable dimers in the PriessGAD.DM cells (Fig. 4⇓C). The ability of HLA-DM to enhance the formation of stable peptide:class II complexes in cells has long been shown to correlate with the appearance of SDS-stable class II dimers on PAGE (10, 12). Remarkably, the enhanced expression of DM also correlated with a significant reduction in HLA-DR4-restricted presentation of cytoplasmic-derived GAD_273–285 (Fig. 5⇓). This reduction occurred even though intracellular GAD protein levels remained equivalent in the two cell lines (Fig. 4⇓D). The addition of exogenous GAD peptide to these cells resulted in similar levels of T cell activation, in line with flow cytometric data indicating similar levels of DR4 expression by these cells (data not shown). Together these data suggest that DM encounters and edits DR4 molecules complexed with GAD_273–285 regardless of the route of Ag delivery. Furthermore, these results suggest that cytoplasmic-derived epitopes may access class II molecules in the absence of DM, yet the resulting peptide:αβ complexes later can undergo DM-meditated editing before display on the cell surface.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Overexpression of HLA-DM in APC enhances the formation of SDS-stable class II:peptide complexes. Western blot analysis was used to evaluate relative HLA-DM, DR, and GAD levels in PriessGAD and PriessGAD.DM cells. Membrane fractions from PriessGAD and PriessGAD.DM were solubilized, and equal amounts of protein from each were fractionated by SDS-PAGE followed by Western immunoblotting. A, C, and D, Samples prepared in reducing SDS sample buffer and boiled for 5 min were probed for DMα and DMβ, DR, or GAD levels. B, Samples prepared in nonreducing SDS sample buffer and incubated for 15 min at room temperature were probed for SDS-stable class II dimers and dissociated DRα. The Ab used here, DA6.147, preferentially recognizes free DRα compared with dimerized DRαβ, thus accounting for the ready detection of dissociated DRα in cells lacking DM. In all cases identical amounts of cell proteins were loaded in each lane, separated on 10% PAGE, and transferred to membranes, and Ab were probed. Each blot is representative of at least three independent experiments.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

DM-mediated inhibition of cytoplasmic GAD presentation by HLA-DR4. PriessGAD or PriessGAD.DM cells were cultured in fresh medium with or without the GAD_273–285 peptide for 2 h before washing and were cocultured with the GAD-specific T cell hybridoma, 33.1. Ag-specific activation of T cell hybridoma lines was measured as described in Materials and Methods. Representative results are shown from three separate experiments, with the data expressed as the mean of triplicate samples and the SE indicated.