A Dominant CTL-inducing Determinant of GAD65 is processed in the Vesicular Compartment

Journal of Autoimmune and Systemic Diseases

Editor in Chief: Kota V. Ramana

Research Article

A Dominant CTL-inducing Determinant of GAD65 is processed in the Vesicular Compartment of Antigen Presenting Cells and Cross-presented on Class I MHC Molecules

Sarah Rasché1, and Anthony Quinn1*
1Department of Biological Sciences, University of Toledo, USA
*Corresponding author: Anthony Quinn, Department of Biological Sciences, University of Toledo, 2801 W. Bancroft St, Toledo, Ohio, USA, Tel: 419-530-7836; FAX: 419-530-7737; E-mail:
Received: May 18, 2017; Accepted: July 13, 2017; Published: July 20, 2017
Copyright: ©2017 Rasché S, et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Citation: Rasché S, Quinn A (2017) A Dominant CTL-inducing Determinant of GAD65 is processed in the Vesicular Compartment of Antigen Presenting Cells and Cross presented on Class I MHC Molecules. J autoimmune systemic dis 1(1): 100001.


MHC class I-restricted CTL can contribute to the pathogenesis of type 1 diabetes by directly killing pancreatic beta cells, which endogenously express the diabetes-associated autoantigen GAD65. Interestingly, the priming and expansion of antigen-specific CTL requires costimulatory signals found only on professional antigen presenting cells, which do not express GAD65, but can present captured exogenous sources of the antigen. To better understand the emergence of GAD-reactive CTL and islet autoimmunity in diabetes-prone NOD mice we investigated whether a dominant Kd-restricted GAD 65 determinant, 546-554, could be generated from both cytoplasmic and vesicular sources of GAD65. Functional cross-presentation to p546-specific T cells occurred when macrophages were cultured with soluble or insoluble polypeptides containing the GAD65 546-554 determinant. Both TAP and proteasome-independent mechanisms contribute to the cross-presentation of p546 from soluble antigen, which was dependent upon endosomal acidification. The processing of determinant 546-554 was reduced in the presence of Brefeldin A, suggesting that nascent MHC I molecules were required for functional epitope generation. While B cell lymphomas were poor at capturing insoluble GAD65 polypeptides, retroviral expression of vesicle-targeted GAD65 converted M12C3 cells into APC capable of fully activating 546-554-specific CTL, therefore, supporting the notion that processing could occur from vesicular stores of GAD65. Collectively, these data suggest that a dominant GAD65 CTL inducing-determinant can be processed and produced via a TAP-independent vacuolar pathway, and that the assembly of the Kd:546-554 can occur entirely within the endosomal system.

Keywords: gad65, autoimmune, diabetes, t cells
View PDF Download PDF


GAD65: glutamic acid decarboxylase; CTL: cytotoxic T lymphocyte; T1D: type 1 diabetes; TAP: transported associated with antigen processing; APC: antigen presenting cell


Type I diabetes (T1D), an inflammatory autoimmune disease, is characterized by the presence of autoreactive T cells that recognize and attack insulin-producing beta cells in the pancreatic islets of langerhans [1-3]. The destruction of beta cells leads to lost glucose homeostasis and the onset of hyperglycemia in human T1D and in the NOD mouse model [3-5]. Both CD4+ and CD8+ T cells are present in the inflamed islets of NOD mice prior to the onset of T1D[6,7]; however, as parenchymal cells, including beta cells, are unable to express MHC class II [6,8], their role in direct antigen presentation is limited to MHC class-I restricted CD8+ T cells. The priming and activation of MHC class I-restricted autoimmunity is necessary for the initiation and progression of the T1D in NOD mice [9,10] and mice that fail to express Kd [9] or MHC I associated-β2M [11,12] are protected from both insulitis and T1D. While these findings and others have established the significance of CD8+ T cells in NOD T1D [13], the evolution of these MHC class I-restricted responses remains to be fully elucidated. Beta cell–specific autoimmunity in T1D may be initiated in the draining pancreatic lymph nodes where antigen released from islet cells during organ development [14,15] can reach an environment containing the professional antigen presenting cells (APC) necessary for CD8+ CTL priming [16,17]. However, the requisite expression of Kd on beta cells in T1D [15] demonstrates that the targets of immune-mediated destruction, beta cells, also have a vital role in presenting autoantigens to islet-destructive CTL [15,18,19]. Therefore, in the larger scheme of pathogenesis in T1D, MHC class-I restricted CTL cytotoxicity requires the processing and presentation of key beta cell antigens as proteins captured by professional APC as well as endogenous proteins processed in beta cells, with the caveat that the proteolytic machinery of each must yield and display the same epitope, peptide determinant: MHC I complex, to be recognized by the cognate CTLs.

Immune responses to glutamic acid decarboxylase (GAD65), an enzyme expressed in the pancreatic beta cells, are indicative of heightened risk for T1D in humans and NOD mouse [20-23]. The Kd-restricted GAD65 determinant 546-554 spontaneously primes CTL capable of inducing insulitis [24,25], localized loss of insulin in vivo [26], and beta cell death in vitro [26]. Given the narrow range of tissues that normally express GAD65 [27-30], the priming of GAD65-specific autoimmunity must include mechanisms where APC responsible for activating naïve T cells are able to present an autoantigen that they fail to express autologously. Cross-presentation is an alternative method of antigen processing that allows professional APC to capture extracellular antigen, proteolytically digest it to peptide fragments that are then loaded onto MHC class I molecules to elicit CTL responses [31]. Previously, we reported that GAD65 546-554-specific CTL are responsive to APCs isolated from pancreatic lymph nodes of mice treated with the beta cell toxin streptozotocin (STZ) [24], confirming that the Kd:546-554 determinant could be processed and presented by cells that acquire GAD65 released from damaged or dying beta cells. Mechanistic studies of cross-presentation have yielded two pathways by which captured protein antigens can be proteolytically degraded and loaded onto MHC class I molecules; one is TAP (transporter associated with antigen processing), proteasome, and ER-dependent [32,33], while the other is independent of TAP and the proteasome, but is reliant upon enzymes typically localized in endosomes and lysosomes [34,35].

In our ongoing studies focusing on the antigen processing of GAD65 [24,25,36,37], a natural and clinically relevant autoantigen, T cell determinant mapping in NOD mice revealed that dominant GAD65 MHC class I epitopes (Kd) were often proximal to dominant GAD65 MHC class II restricted determinants [24], including the Kd:546-554 determinant and the I-Ag7 restricted GAD530-543 Th determinant that arise spontaneously [25,36]. Structural proximity of MHC I and MHC II dominant determinants was first reported in studies of T cell responses to beta galactosidase [38] and more recently for self-antigens associated with autoimmune disease, including insulin [39] and myelin oligodendrocyte glycoprotein (MOG) [40-42], and model antigens like hen egg-white lysozyme (HEL) [43]. These findings would be consistent with a positional relationship during processing and common enzymatic machinery for antigenic peptide generation. This clustering of epitopes could contribute to the enhanced antigenic dominance of the region, provided that immunogenic forms of both MHC class I and class II epitopes can be generated from exogenous sources of the antigen. Here, we find that processing of the GAD 546-554 determinant could occur independent of TAP and the proteasome, and were diminished in the presence of drugs that disrupt endosomal acidification. Additionally, we observed a reduction of cross-presentation in the presence of cysteine cathepsin inhibitors and Brefeldin A, suggesting that this determinant could be processed via a “vesicular” pathway and occur within the endosomal compartment.

Materials and Methods

Mice and immortalized cell lines

NOD mice were purchased from Taconic Farms (Germantown, NY) and housed in the University of Toledo Animal Care Facility. All experiments described in this report were reviewed and approved by the University of Toledo’s Institutional Animal Care and Use Committee. 4G4 B cell hybridomas were provided by Dr. E. Sercarz (Torrey Pines Institute for Molecular Studies, San Diego, CA), RAW264 cells were a kind gift from Dr. Mark Wooten (University of Toledo, Toledo, OH), and M12C3.GAD65.g7.WT.PPL were a kind gift from Dr. Dario Vignali (St. Jude’s Children’s Hospital, Memphis TN). The macrophage lines P388.D1, J774, and M12C3 were purchased from the American Tissue Culture Collection (Manassas, VA). The tap-deficient RMA-S cell line, transfected with a Kd-expressing plasmid (RMA-S-Kd), were provided by Dr. Leonard Harrison (Walter and Eliza Hall Institute, Melbourne, Australia), and cultured as previously described [24]. GAD65 p546-specific T cell lines and clones were isolated from immunized NOD mice as previously described [24]. The T cell clone 546.14.46 was isolated by limiting dilution and restimulated with peptide-pulsed irradiated NOD spleen cells and expanded by culture in IL-2 containing medium. All cells were cultured in complete medium (RPMI-1640 supplemented with 10% FBS, 1% penicillin/streptomycin, 0.001% 2-mercaptoethanol, 1% sodium pyruvate, and 1% nonessential amino acids, Mediatech, Herndon, VA).


GAD65 peptides 530-543 (APVIKARMMEYGTT), 546-554 (p546) (SYQPLGDKV), and GAD65 polypeptides 530-554 (p530) (APVIKARMMEYGTTMVSYQPLGDKV) and 524-554 (p524) (SRLSKVAPVIKARMMEYGTTMVSYQPLGDKV) were synthesized as previously described [24]. Biotin conjugated GAD65 peptide 530-554 (b-p530) was synthesized by Ohio Peptide (Powell, OH). In all experiments, p546 was used at a final concentration of 10 μg/ml, while polypeptides p530 and p524 were used at a final concentration of 30 μg/ml, these concentrations were shown to yield optimal responses in titration experiments, respectively, and represent concentrations of 0.86-1.07 × 10-5M. For microbead cross-presentation assays, 1μm latex microspheres (Sigma, St. Louis, MO) were diluted to 0.02% v/v and coated with 50 μg/ml GAD65 polypeptides p530 or p524 overnight at 25˚C. The peptide-coated microspheres were washed extensively prior to being fed to the indicated APCs.

Confocal microscopy

To detect GAD65 localization within endosomes of M12C3.g7.GAD.WT.PPL (PPL) cells, the cells were fixed with cold 4% paraformaldehyde for 10 min at 25˚C, washed, and blocked with 10% FBS/0.5% BSA for 1 hr. The cells were then stained with mouse anti-GAD65 monoclonal antibody (GAD6) followed by biotin-conjugated rat anti-CD107 (LAMP-1) (BioLegend, San Francisco, CA). After application of both primary antibodies, the cells were stained with FITC conjugated goat-anti-mouse IgG, and Texas Red-conjugated StreptAvidin. Finally, the cells were mounted on glass slides with Vectamount (Vector, Burlingame, CA) and analyzed by confocal microscopy. Alternatively, the cells were suspended to 1x106 cells in 10% FBS CM supplemented with 50 nM LysotrackerRed (Invitrogen, Carlsbad, CA) and incubated at 37°C for 2 hrs. The cells were then washed, fixed, blocked, and stained with anti-GAD65, biotin-conjugated goat-anti-mouse IgG, and StreptAvidin-Alexa 488, and mounted on glass slides with Vectamount before analysis by confocal microscopy. To detect localization of microspheres within endosomes, RAW264D or J774.1 cells were suspended to 1 × 106 cells/ml in 10% FBS CM supplemented with 0.02% v/v 1μm latex beads (Molecular Probes, Eugene, OR). The cells were incubated at 37°C for 30 minutes, washed and incubated for an additional hr at 37°C before fixation. The cells were then stained with FITC-conjugated anti-Kd and biotin-conjugated anti-LAMP-1 (both from BioLegend, San Francisco, CA) for 1 hr followed by staining with Texas Red-conjugated StreptAvidin. The resulting cells were mounted on glass slides with Vectamount and analyzed by confocal microscopy. The slides were viewed and images were captured with a Leica TSC-SP5 confocal microscope. No filter was utilized and 488 and 561 lasers were used to obtain all images. Magnification for individual photographs is listed in the figure legends. Protein co-localization was analyzed using Leica LAS software.

p530-554 binding studies

To detect the binding of GAD65 polypeptides to the Kd allele, TAP-deficient RMAS-Kd cells (1 × 105 cells/ml) were incubated for 24 hrs at 25˚C to stabilize MHC I surface expression [24]. The next day, GAD65 polypeptides were added to the cell cultures and incubated at 25˚C for 1 hr. The cells were then placed at 37°C for 3 hrs before staining with biotin-conjugated-anti-Kd (Biolegend, San Francisco, CA) and StreptAvidin-Alexa 488, and analysis by flow cytometry. To detect binding of biotin-conjugated p530-554 (b-p530), P388 cells (1 × 105) were incubated with b-p530 (1 or 30 μg/ml) for 1 hr at 37°C. The cells were then washed and set aside at 4°C or fixed with 2% paraformaldehyde. The fixed cells were permeabilized with 0.1% saponin/0.5% BSA for 15 min at room temperature. All cells were then stained with StrepAvidin-Alexa488 for 15 minutes 25˚C, washed twice, and analyzed by flow cytometry.

Primary macrophage and dendritic cell isolation

To generate bone marrow-derived phagocytes (dendritic cells and macrophages), bone marrow was expelled from the femurs and fibulas of NOD mice, homogenized to single cell suspension, and plated at 1.5 × 107 cells per 12 cm cell culture dish. The cells were cultured for 72 hrs in the presence of 4nM GM-CSF or M-CSF (Peprotech, Rock Hill, NJ) in 10 ml of 20% FBS-CM to enrich the growth of dendritic cells or macrophages, respectively. On Day 4, the culture media was replaced with fresh growth factor supplemented media. On Day 6, culture media was replaced with 20% FBS-CM without growth factors. The resulting cells were trypsinized and the expression of CD11b, CD11c, and CD8α verified by flow cytometry. The DCs were defined as CD11b+/CD11c+/CD8 α+, and macrophages were CD11b+/CD11clo/CD8 α(-). Additionally, peritoneal macrophages (p-macs) were isolated by flushing NOD mouse peritoneal cavities with 5 ml PBS. The cells from the peritoneal lavage were washed, and evaluated for CD11b, CD11c, and CD8α expression by flow cytometry, as above. P-macs were defined CD11b+/CD11clo/CD8 α-. The cells described were then used as APCs in T cell stimulation assays. The APCs (3 × 105 cells/well) were pulsed with p546, p530, or p530-coated 1 μm latex beads for 4 hrs, washed, and plated in 96-well plates with 546.14.46 CTL clones (5 × 104 cells/well). Twenty-four hrs later, supernatants were screened for IFN-γ production by ELISA as previously described [24].

RMAS-Kd cross-presentation

To determine whether TAP was required for p546 cross-presentation, TAP-deficient RMAS-Kd cell were incubated with p546, p524, or p524-coated 1 μm latex beads for 24 hrs at 37°C. The next day, the RMAS-Kd cells (2 × 105 cells/well) were co-cultured with the CTL clone 546.14.46 (5 × 104 cells/well) at 37°C. Twenty-four hrs later, supernatants were screened for IFN-γ production by ELISA.

Inhibition of Antigen Processing

To inhibit cross-presentation, P388, J774.1, or RAW264 cells were suspended to 1 × 106/ml in 10% FBS CM supplemented with MG132 (1.6 μM) or chloroquine (50 μM) for 30 minutes at 37°C. Peptide p546 or polypeptide p530 were then added to a final concentration of 10 or 30 μg/ml, respectively, for 3 hrs at 37°C. The cells were then washed twice and plated at 1 × 105 cells/well with GAD65 546-554-specific CTL clone 546.14.46 (5 × 104 cells/well) and incubated for 24 hrs at 37°C. The next day, an IFN-γ ELISA was used to measure CTL activation. Additionally, RMAS-Kd cells (2 × 106) were pretreated with Brefeldin A (5 μg/ml) for 30 minutes at 25°C. p546 (10 μg/ml) or p530 (30 μg/ml) were then added for 2 hrs at 25°C, washed, and incubated for an additional 3 hrs in the presence of Brefeldin A. After washing, the treated cells were plated with CTL clone 546.14.46 (5 × 104 cells/well) overnight at 37°C. The next day, an IFN-γ ELISA was used to measure CTL activation. To inhibit SIINFEKYL processing, ovalbumin (OVA)-expressing EG7.OVA cells (ATCC, Manassas, VA) were acid stripped (citric acid buffer; pH=4.0) to remove surface MHC I: peptide complexes followed by extensive washing with PBS. The cells were then suspended in 10% FBS CM supplemented with MG132 (1.6μM ) or DMSO for 4 hrs at 37°C, then washed and fixed with 2% paraformaldehyde for 10 minutes at room temperature. The cells were then plated at 3 × 105 cells/well with the CD8+ T cell hybridoma B3Z.D7 (2 × 105 cells/well), which is specific for the OVA epitope, SIINFEKYL. B3Z.D7 cells express IL-2 and β-galactosidase under the IL-2 promoter when activated [44]. Twenty-four hrs later IL-2 in the culture medium was measured by ELISA.


The results from assays were compared using Student’s paired t tests and GraphPad Prism software (La Jolla, CA). p values < 0.05 were considered significant. The figures represent results found in 3 or more experiments.


Functional cross-presentation of the Kd:546-554 epitope occurs when captured antigen is targeted to phagosomes

Collectively, our previous works indicated that the 524-543 region of the GAD65 molecule contains at least 3 distinct T cell epitopes that are recognized by spontaneously arising T cells in young prediabetic NOD mice [25,36]. Interestingly, two of the epitopes can be engaged by Th cells [36], while the third stimulates GAD546-554-specific CTL [25], presumably from cytoplasmic stores of antigen.

Figure 1

Figure 1: The GAD65 polypeptide p524-554 activates both CTLs and Th cells. (A) Organization of T cell determinants in the GAD65 polypeptides 524-554 and 530-554. (B) Immunization with GAD65 polypeptide induces both Th and CTL responses. The draining lymph nodes from three NOD mice immunized subcutaneously with polypeptide 530-554 in CFA were tested for proliferative recall responses to in vitro challenge with GAD65 peptides (2 or 10 µg/ml) in tritiated thymidine (3H) uptake assay. The results are expressed as the mean of stimulation indices relative to unstimulated lymph node cells and are representative of four experiments. (C) APCs generate a Kd-binding epitope from GAD65 polypeptide sources. Triplicate cultures of J774 cells were pulsed with GAD65 polypeptides 524-554 or 530-554, washed, and plated with the CTL clone 546.14.46. The Kd-binding peptides GAD65 546-554 and influenza peptide NP147 served as positive and negative controls, respectively. IFN-? in 24 hr culture supernatants was quantified by ELISA and the results expressed as the mean of triplicate wells. (D) APCs generate MHC II-binding epitopes from GAD65 polypeptide sources. Triplicate cultures of NOD SPC were pulsed with GAD65 polypeptides 524-554 or 530-554 and plated with clonal Th cell hybridomas specific for GAD65 530-543 (530.38.3) or GAD65 524-538 (524.169.1). IL-2 in culture supernatants was quantified by ELISA 24 hr later and the results expressed as the mean of triplicate wells. *p< 0.05 when compared to negative control peptides or media control

To investigate the potential for endosomal evolution of the Kd:546-554 epitope, we synthesized two GAD65 polypeptides p530-554 (p530) and p524-554 (p524) - each contains the 546-554 CTL-inducing determinant and an adjacent I-Ag7-restricted Th-inducing determinant, 530-543 and/or 524-538, respectively (Figure 1a) [36]. Importantly, both polypeptides are much longer than the optimal Kd-binding sequence [45,46] and therefore would be expected to require proteolytic processing prior to binding to the MHC I molecule and acquisition of the ability to stimulate CTL. When emulsified in CFA and used as an immunogen in NOD mice, the p530 polypeptide was able to engage and expand both Kd-restricted and I-Ag7-restricted responses in antigen recall assays (Figure 1b), suggesting that both classes of determinants were processed from insoluble GAD65 polypeptide captured by APC - albeit the I-Ag7 determinant more efficiently. In additional experiments designed to show that exogenous sources of antigen could be used to generate the Kd:546-554 epitope, NOD spleen cells fed GAD65 polypeptides p524 or p530 were each able to activate a 546-554-specific CTL (Figure 1c). As predicted, polypeptide p524 could be used as a substrate to produce the epitopes needed for activation of GAD65 524-538 and GAD65 530-543-reactive Th cells (Figure 1d), while polypeptide p530 was only able to stimulate 530-543-specific Th cells (Figure 1d).

To rule out the possibility that the GAD65 polypeptides were binding directly to surface Kd molecules, and not requiring processing, we utilized the TAP-deficient RMAS-Kd cell line, which poorly expresses surface MHC I due to a mutation in the TAP2 subunit. TAP is needed to shuttle short cytoplasmic peptides into the ER for loading and stabilization of nascent MHC class I molecules and the efficient transport of the complex to the cell surface [47]. For TAP-deficient RMAS-Kd cells, exogenously supplied Kd-binding peptides can directly stabilize empty surface MHC I molecules and increase the surface expression on the cells [24]. We observed a significant increase in Kd expression on RMAS-Kd cells pulsed with p546-554, while no change in Kd expression occurred in the presence of the polypeptides p530 (Figure 2a) or p524 (data not shown). To provide additional evidence that the 530-554 sequence does not bind directly to cell surface Kd molecules we synthesized polypeptide GAD65 b-p530, which has biotin attached to the carboxyl terminus of the GAD65 530-554 sequence. P530-554 did not bind to Kd-expressing P388.D1 cells (Figure 2b), while it bound well to I-Ag7-expressing P388 cells (Figure 2b).

Finally, to show that the 546-554 sequence within the GAD65 polypeptides required processing prior to loading onto Kd molecules, we compared live and glutaraldehyde-fixed NOD spleen cells for the ability to cross-present the GAD65 polypeptide p530-554. While both live and fixed splenic APC readily stimulated GAD65-specific CTL when provided the 9-mer peptide 546-554 (Figure 2c), only the live splenocytes were able to display the Kd:546-554 epitope after capture of the p530-554 polypeptide in culture (Figure 2c).

Figure 2

Figure 2: GAD65 polypeptides do not bind directly to Kd molecules. (A) TAP-deficient RMAS-Kd cells were incubated for 24 hrs at 25°C to stabilize MHC I surface expression. The next day, GAD65 peptide p546-554 or polypeptide p530-554 was added to triplicate cell cultures, incubated at room temperature for 1 hr, and then at 37°C for an additional 3 hrs. The resultant cells were washed and stained with biotin-conjugated-anti-Kdand StrepAvidin-Alexa488, and analyzed by flow cytometry. Filled histograms denote cells without peptide, while open histograms denote peptide-pulsed cells. The mean fluorescence intensity (MFI) listed refers to the peptide-pulsed cells. (B) GAD65 polypeptides bind I-Ag7+. M12C3 cells (Kd+/I-Ag7+) or P388 cells (Kd+ /I-Ag7-) were incubated with biotin-conjugated p530-554 (biot-p530) for 1 hr at 37°C. The cells were washed and stained with StrepAvidin-Alexa488 to detect surface-bound peptide by flow cytometry. The MFI from triplicate cultures of the biotin-p530-pulsed cells is noted in the figure. (C) GAD65 polypeptide requires processing to generate a CTL-inducing epitope. Live or glutaraldehyde-fixed NOD spleen cells were pulsed for 2 or 8 hrs with GAD65 peptide 546-554 or polypeptide 530-554, washed, and added to triplicate cultures of CTL clone GAD546.14.86 in 96-well plates. Forty-eight hrs later the supernatants were tested for IFN-? by ELISA, and the results expressed as the mean of triplicate wells. *p<0.05 when compared to negative control peptides

The GAD546-554 CTL determinant can be processed from GAD65 molecules targeted to endocytic vesicles

The GAD65 determinant 546-554 is displayed in the context of Kd on the surface of NIT-1 insulinoma cells [26] and M12C3.g7.GAD.48.16 lymphoma cells [24] both of which express GAD65 in the cytoplasm. On the other hand, antigen presenting cells from streptozotocin-treated NOD mice were able to use captured GAD65 to activate a similar group of Kd-restricted 546-554 specific CTL [24], showing that the Kd:546-554 epitope could be created from vesicular stores of GAD65 as the substrate. Here, we utilized retrovirally-transduced M12C3 B cells as APC to determine if the same Kd:546-554 epitope could be created and displayed when GAD65 is targeted to intracellular vesicles. The M12C3.GAD65.g7.WT.PPL cell line expresses GAD65 with a preprolactin (PPL) signal sequence attached, causing the beta cell protein to preferentially localize in vesicles including lysosomes (Figure 3a) and late endosomes (Figure 3b). CTL specific for the GAD65 Kd:546-554 epitope were readily activated when cultured in the presence of M12C3.GAD65.g7.WT.PPL cells, as shown by IFN-γ production (Figure 3c), providing further evidence that the 546-554 determinant could be processed from vesicular sources of GAD65. We also observed localization of Kd molecules within late endosomes, suggesting that both processing and loading of peptides could occur within these vesicles (data not shown). Th cells specific for GAD65 I-Ag7-restricted determinants 530-43 and 268-83 were also able to detect their cognate epitopes on the surface of M12C3.GAD65.g7.WT.PPL cells (Figure 3d). Influenza- and HEL-specific Th cells were unresponsive to the GAD65-expressing cells (Figures 3c and 3d).

Figure 3

Figure 3: GAD65 in endosomal compartments can be processed and presented to Kd-restricted CTL. (A) GAD65 localizes to lysosomes in M12C3.g7.GAD.WT.PPL cells. The cells were labeled with LysotrackerRed (50 nm) and stained with antibodies to GAD65. The resultant cells were analyzed by confocal microscopy. The images, representative of four experiments, depict single and overlaid fluorescent channels from a single plane, as well as the overlaid channels of a Z-stack. The yellow structures depict localization of GAD65 (green) to lysosomes (red). Scale bar = 10 µm; 2040x total magnification. (B) GAD65 localizes to late endosomes in M12C3.g7.GAD.WT.PPL cells. Fixed cells were stained with rat anti-CD107 (LAMP-1) and anti-GAD65 and analyzed by confocal microscopy. The cells were then fixed and analyzed by confocal microscopy. Scale bar = 10 µm; 2040x total magnification. The images are representative of four experiments. (C) Endosomal-targeted GAD65 can generate CTL-inducing epitopes. CTL clones 546.14 46 and APCs (M12C3, M12C3.g7.GAD.WT.PPL, M12C3 + p546) were co-cultured for 24 hrs at 37°C. Supernatants were collected and IFN? was quantified by ELISA. An influenza-specific CTL clone (Flu.1) served as the negative control for GAD p546 specificity. The results represent mean values of triplicate samples. (D) Endosomal-targeted native GAD65 can generate Th-inducing epitopes. B cell APCs (4G4, M12C3, M12C3.GAD.PPL) were co-cultured with Th cell hybridomas specific for GAD65-530-43(530.1), GAD65 268-283 (268.11) or hen egg white lysozyme 11-25 (HEL.8) for 24 hrs at 37°C. Supernatants were collected and IL-2 was quantified by ELISA. The results represent the mean of triplicate wells. *p<0.05 when compared to media control

To definitively demonstrate that captured sources of GAD65 shuttled to the vesicular compartment could give rise to the Kd:546-554 epitope, we used GAD65 polypeptides (p530-54 or p524-54) adsorbed to 1 µm latex beads as the source of antigen. This approach ensures that the only supply of specific antigen is indeed derived from molecules captured into endocytic vesicles. Therefore, when added to cultures of J774.1 APC the latex beads co-localized to the same endocytic vesicles as those visualized with captured red dextran (Figure 4a), reaching near maximal uptake within 1hr (Figure 4b). Further analysis of subcellular localization revealed that the microbeads were captured and entered Kd-containing endocytic vesicles (Figure 4c), which co-localized with lysosomal-associated membrane protein (LAMP-1) positive vesicles (Figure 4c).

To demonstrate that primary APC could present antigen coated onto beads to CTL, bone marrow-derived dendritic cells, bone marrow-derived macrophages, and peritoneal macrophages were fed GAD polypeptides p530 or p524 coated beads and shown to efficiently activate GAD65 546-554-specific CTL (Figures 4d-4f). In contrast, as the 4G4 B cells proved poor at capturing latex beads (data not shown), they were unable to activate the CTL when cultured with beads (Figure 4g), but were fully capable of activating the CTL cells when the APC were pulsed with soluble GAD65 peptide p546 or polypeptides p530 (Figure 4g). These findings confirmed that the source of CTL stimulating peptide was not dissociated or unbound polypeptide available prior to internalization of beads. Additionally, these results are supportive of our hypothesis that endosomal processing contributes to the generation of the Kd:546-554 epitope.

Figure 4

Figure 4: The Kd:546-554 epitope can be crosspresented from extracellular sources of antigen. A) Exogenously provided latex microbeads are taken up and directed into the endocytic vesicles of J774 cells. J774 cells were pulsed with 1µm latex beads (0.02% w/v) and TexasRed-conjugated 10 kDa dextran for 15 min at 37°C, washed, and incubated for an additional 1 hr at 37°C. The cells were then fixed and analyzed by fluorescent microscopy. Representative images from four experiments are shown – the red channel (left) or the red and DIC overlay channels (right) depict Dextran (red) co-localization with the microbeads (white). Scale bar = 5 µm; total magnification is 3600x. (B) J774 cells were pulsed with 1µm latex beads (0.02% w/v) and Texas Red-conjugated 10 kDa dextran for 15 min at 37°C, washed, and incubated ("chased") for an additional 15, 60, or 240 min at 37°C. The cells were then fixed and beads were located by DIC and determined to be localized to vesicles (red) or not. The results depict the percentage of beads co-localized with red dextrans from at least 100 beads per chase time point in one of four experiments. (C) RAW264 or J774 cells were pulsed with 1µm latex beads (0.02% v/v) at 37°C for 15 minutes, washed, and chased for 1 hr at 37°C. The cells were then fixed and stained with anti-Kd and anti-LAMP-1. The resultant cells were analyzed by confocal microscopy. Shown are representative images of single fluorescent channels as well as merged fluorescent channels and fluorescent channel/DIC overlays. Yellow depicts co-localization of Kd (green) with LAMP-1 (red). The microbeads localized with Kd+/LAMP-1+ vesicles (yellow) are visualized in the DIC overlays (right). Scale bars = 2 µm; total magnifications are 2520x for RAW264 and 3600x for J774. (D) Primary NOD phagocytes can crosspresent antigen from GAD65 polypeptide coated microbeads. Triplicate cultures of NOD bone marrow derived (D) dendritic cells, (E) macrophages, (F) peritoneal macrophages, or (G) 4G4 B cell lymphoma cells were pulsed with GAD65 polypeptides or GAD65 polypeptide-coated microbeads for 4 hrs, washed, and plated with 546.14.46 CTL clones. IFN? in supernatants was quantified by ELISA 24 hrs later. The results are the mean response of triplicate wells. *p<0.05 when compared to media control

P546 processing is dependent upson endosomal acidification

To determine whether the cross-presentation of the Kd:546-554 epitope from captured antigen involves proteolytic mechanisms associated with endocytic vesicles, three different macrophage lines were pulsed with GAD65 polypeptide in the presence of chloroquine (CQ), an inhibitor of endosomal acidification. Chloroquine significantly inhibited the cross-presentation of GAD65 polypeptide p524-554 by macrophage lines J77.4, P388.D1, and RAW264D cells as shown by the drug’s ability to significantly reduce activation of 546-specific CTL (Figure 5a). On the other hand, we saw little inhibition of p524-554 cross-presentation in the presence of the proteasome inhibitor MG132 (Figure 5a). Neither MG132 nor chloroquine inhibited presentation of the 9-mer peptide p546-554 (Figure 5b) demonstrating that the drugs did not adversely affect the interaction between the epitope and TCR, or the stability of MHC I expression. Similar results were observed when peritoneal macrophages were used as the APC (data not shown). Under similar assay conditions, MG132 inhibited processing of the SIINFEKYL epitope from cytoplasmic OVA in EG7.OVA cells, and reduced the subsequent activation of cognate specific CTL B3Z.D7 by 40% (Figure 5c).

Figure 5

Figure 5: Endosomal Acidification is required for GAD65 p546 crosspresentation. (a) APCs (RAW264D, J774, or P388) were pretreated with indicated concentrations of the antigen processing inhibitors chloroquine (CQ) or MG132 for 30 min at 37°C. GAD65 polypeptide 530-554 (A) or peptide 546-554 (b), were then added at a final concentration of 30 or 10 µg/ml, respectively, for 3 hrs at 37°C. The APCs were then washed and plated with triplicate cultures of CTL clone 546.14.46 for 24 hrs at 37°C. The supernatants were screened for IFN? production by ELISA. The results are expressed as the mean of triplicates cultures. (c) Proteasomal processing is required for the generation of CTL-inducing epitopes produced from cytoplasmic antigen. Cytoplasm-targeted ovalbumin (OVA)-expressing EG7.OVA cells were subjected to citric acid stripping to remove surface peptide:MHC-I complexes. The cells were then treated with MG-132 or DMSO (media) for 4 hrs at 37°C. The cells were then washed, fixed, and plated with the OVA-specific CD8+ T cell hybridoma, B3Z.D7. Twenty-four hrs later, IL-2 secretion was quantified by ELISA. The results are expressed as the mean of triplicates cultures. *p<0.05 when compared to media controls

TAP is not required for p546 cross-presentation

The role of TAP in cross-presentation appears to differ depending on the model antigen and cell type; however, for peptides processed within the endosomal network, the chaperone activity of TAP is not required. We found this mechanism was not required for the cross-presentation of the GAD65 polypeptide p530, as we observed robust GAD65 546-554-specific CTL activation when the TAP-deficient RMAS-Kd cell line was fed p530-coated latex beads (Figure 6a) - which enter the cells via phagosomes, but are unable to cross the plasma membrane for direct entry into the cytosol. Additionally, when the RMAS-Kd cells are pulsed with the GAD65 polypeptide in the presence of Brefeldin A, the cross-presentation is diminished, suggesting that nascent MHC I molecules are required for functional Kd:546 epitope display (Figure 6b).

Figure 6

Figure 6: TAP is not required for GAD65 p546 crosspresentation. (A) TAP-deficient RMAS-Kd cells were incubated with GAD65 peptide 546-554 (10 µg/ml), GAD65 polypeptide 530-554 (30 µg/ml) or GAD65 polypeptide-coated latex microbeads at 37°C. Twenty-four hrs later, the cells were washed and plated with triplicate cultures of CTL clone 546.14.46 at 37°C. Twenty-four hrs later, supernatants were screened for IFN? production by ELISA. The results are expressed as the mean of triplicates cultures.*p<0.05 when compared to media control. (B) Nascent Kd molecules are needed to crosspresent GAD65 p530 polypeptide. RMAS-Kd cells were pre-treated with Brefeldin A (5 µg/ml) followed by addition of GAD65 peptide 546-554 or polypeptide 530-554 at 25°C. The cells were washed and chased for an additional 3 hrs with Brefeldin A. The treated cells were washed and plated with triplicate cultures of CTL clone 546.14.46. Twenty-four hrs later, supernatants were screened for IFN? production by ELISA. The results are expressed as the mean of triplicates cultures. *p<0.05 when compared to buffer controls


Type 1 Diabetes is a multigenic autoimmune disease associated with the expression of several immune-related genes, including particular HLA and MHC alleles in humans and mice, respectively [4]. Peculiar MHC alleles may increase susceptibility to autoimmune disease by shaping the responding T cell repertoire (clonotypes) via preferential binding and display of unique self-peptide determinants during thymic selection and peripheral activation of pathogenic autoreactive T cells [48]. Although insulitis may occur in the absence of MHC I on beta cells, direct interaction between beta cells and CTL contributes significantly to the progression of T1D, as mice lacking MHC on beta cells fail to develop hyperglycemia [18]. Here we show that dominant GAD65 determinants for the two MHC alleles requisite for type 1 diabetes in NOD mice, I-Ag7 and Kd , can be loaded in the endocytic compartment of antigen presenting cells, including the CTL-inducing determinant GAD65 546-554, which is associated with islet inflammation and T1D [24,26,49].

Previously we suggested that regions of GAD65 could preferentially give rise to dominant self-determinants capable of binding MHC class I or MHC class II, and contribute to the priming cellular autoimmune immune responses [24]. We found that four of the eight immunogenic peptides capable of inducing GAD65-specific Kd-restricted CTL were also proximal to sequences known to induce I-Ag7-restricted Th cell responses in prediabetic NOD mice [24] - their selection during antigen processing may be owing to the availability of distinct determinants to proteolytic enzymes [50]. Kd-restricted CTL specific for GAD65 peptides 88-78, 268-78, and 546-554 were also responsive to GAD65 determinants crosspresented and displayed on the surface of APC recovered from the pancreatic lymph nodes of streptozotocin-treated mice [24]. Furthermore, peptide 546-554 was the most immunogenic of these peptides, displayed the highest affinity for Kd, and prompted the most robust cytotoxic and cytokine responses in naïve mice [24], all of which is consistent with its immunodominant nature [51]. Similarly, GAD65 peptide 524-543 contains an immunodominant determinant recognized by splenic Th cells in naïve NOD [22]. Extracellular antigens captured into phagosomes are subject to proteolytic degradation within specialized endocytic compartments prior to loading onto MHC class II molecules. Given their immunodominance, we hypothesized that the same proteolytic mechanisms that yield the Th determinant within the GAD65 524-543 sequence might also contribute to processing of the 546-554 CTL determinant.

Several factors may contribute to determinant hierarchy, including the tertiary structure of the protein antigen [51]. The carboxyl-terminal domain of GAD65, which contains multiple immunodominant determinants [22,36], may be more readily accessible and easily processed compared to other determinants. In fact, the crystal structure of human GAD65 shows that this loop is “visible” as opposed to being hidden in the center of the molecule [52], perhaps avoiding steric hindrance of proteases. As the immunogenicity of the CTL determinant may also be enhanced by its proximal relation to Th-inducing determinants [38], it is possible the determinants capable of inducing CTL and Th determinants following capture by macrophages and dendritic cells are preferentially selected early in the immune response and thus indicative of heightened susceptibility.

While dendritic cells have dominated the discussion on cross-presentation in humans and are clearly sufficient for priming CTL in mouse models, macrophages are also capable of priming CTL through this alternative process of antigen presentation [53-55], perhaps via macropinocytosis, and are necessary for the progression of autoimmune diabetes [56,57]. The proteolytic machinery responsible for generating antigenic specific epitopes may also depend on whether the captured protein is shuttled to early endosomes or late endosomes [43]. The requirements for the endosomal cathepsins in cross-presentation vary, possibly due to differential antigen uptake [43] or variable structural, chemical, and catalytic properties of individual antigens. Evidence for Cathepsin S (Cat S), a cysteine protease, in generating MHC class I binding peptides generation is equivocal. Cat S facilitates cross-presentation of ovalbumin epitopes in mouse DC [35,58,59] and influenza epitopes in mice in vivo [35]. Conversely, Cat S-mediated epitope degradation hinders cross-presentation of lysozyme epitopes in mouse DC and macrophages [43]. Cathepsin-mediated degradation may play an even more formidable role in macrophages, which contain high levels of endocytic proteases [60]. Discovering the role of specific cathepsins in the processing of individual GAD65 determinants may prove challenging as this family of proteases appears to have overlapping specificities [61].

Bone marrow-derived dendritic cells and macrophages were able to cross-present the 546-554 determinant from both particulate and soluble sources of GAD65 polypeptides. However, 4G4 B cells were unable to efficiently crosspresent the epitope from bead-bound polypeptide antigen. On the other hand, M12C3 cells expressing GAD65 in the cytoplasm [24] or in vesicles could activate p546-554s-specific CTL, demonstrating that Kd:546 epitope could be produced by both endocytic and proteosome driven processing mechanisms. As such, beta cells expressing GAD65 are also able to generate the Kd:546 epitope [24], completing the cycle needed for peripheral priming of CTL through cross-presentation in professional antigen presenting cells and engagement of cognate CTL effector function by parenchymal cells.

CD4+ and CD8+T cell responses to beta cell antigens characterize autoimmunity in the NOD mouse. While the genesis of this response is still under investigation, antigens released during normal islet organ development could incite pancreatic inflammation in the neonatal NOD, which subsequently releases additional cellular debris [14]. Given that CD8+ T cells must be primed and activated by professional APC before acquiring pathogenic CTL effector function, cross-presentation of self-antigens by macrophages and dendritic cells is likely important in T cell mediated autoimmune diseases like T1D [16]. The ability to generate the same dominant MHC class I epitope from endogenous and exogenous sources of autoantigen may also be an important factor in breaking self-tolerance and increasing susceptibility to autoimmune disease.


  1. Cahill GF Jr, McDevitt HO (1981) Insulin-dependent diabetes mellitus: the initial lesion. N Engl J Med 304: 1454-1465.
  2. Cahill GF Jr (1981) Diabetes mellitus-an overview. Curr Concepts Nutr 10: 145-151.
  3. Eisenbarth GS (1986) Type I diabetes mellitus. A chronic autoimmune disease. N Engl J Med 314: 1360-1368.
  4. Wicker LS, Todd JA, Peterson LB (1995) Genetic control of autoimmune diabetes in the NOD mouse. Annu Rev Immunol 13: 179-200.
  5. Yang Y, Santamaria P (2006) Lessons on autoimmune diabetes from animal models. Clin Sci (Lond) 110: 627-639.
  6. Weringer EJ, Like AA (1988) Identification of T cell subsets and class I and class II antigen expression in islet grafts and pancreatic islets of diabetic BioBreeding/Worcester rats. Am J Pathol 132: 292-303.
  7. Jarpe AJ, Hickman MR, Anderson JT, Winter WE, Peck AB (1990) Flow cytometric enumeration of mononuclear cell populations infiltrating the islets of Langerhans in prediabetic NOD mice: development of a model of autoimmune insulitis for type I diabetes. Reg Immunol 3: 305-317.
  8. McInerney MF, Rath S, Janeway CA Jr (1991) Exclusive expression of MHC class II proteins on CD45+ cells in pancreatic islets of NOD mice. Diabetes 40: 648-651.
  9. Katz J, Benoist C, Mathis D (1993) Major histocompatibility complex class I molecules are required for the development of insulitis in non-obese diabetic mice. Eur J Immunol 23: 3358-3360.
  10. Christianson SW, Shultz LD, Leiter EH (1993) Adoptive transfer of diabetes into immunodeficient NOD-scid/scid mice. Relative contributions of CD4+ and CD8+ T-cells from diabetic versus prediabetic NOD.NON-Thy-1a donors. Diabetes 42: 44-55.
  11. Wicker LS, Leiter EH, Todd JA, Renjilian RJ, Peterson E, et al. (1994) Beta 2-microglobulin-deficient NOD mice do not develop insulitis or diabetes. Diabetes 43: 500-504.
  12. Sumida T, Furukawa M, Sakamoto A, Namekawa T, Maeda T, et al. (1994) Prevention of insulitis and diabetes in beta 2-microglobulin-deficient non-obese diabetic mice. Int Immunol 6: 1445-1449.
  13. Liblau RS, Wong FS, Mars LT, Santamaria P (2002) Autoreactive CD8 T cells in organ-specific autoimmunity: emerging targets for therapeutic intervention. Immunity 17: 1-6.
  14. Turley S, Poirot L, Hattori M, Benoist C, Mathis D (2003) Physiological beta cell death triggers priming of self-reactive T cells by dendritic cells in a type-1 diabetes model. J Exp Med 198: 1527-1537.
  15. Yamanouchi J, Verdaguer J, Han B, Amrani A, Serra P, et al. (2003) Cross-priming of diabetogenic T cells dissociated from CTL-induced shedding of beta cell autoantigens. J Immunol 171: 6900-6909.
  16. Kurts C, Heath WR, Carbone FR, Kosaka H, Miller JF (1988) Cross-presentation of self-antigens to CD8+ T cells: the balance between tolerance and autoimmunity. Novartis Found Symp 215: 172-81; discussion 81-90.
  17. Heath WR, Carbone FR (2001) Cross-presentation in viral immunity and self-tolerance. Nat Rev Immunol 1: 126-134.
  18. Hamilton-Williams EE, Palmer SE, Charlton B, Slattery RM (2003) Beta cell MHC class I is a late requirement for diabetes. Proc Natl Acad Sci U S A 100: 6688-6693.
  19. Pang S, Zhang L, Wang H, Yi Z, Li L, et al. (2009) CD8(+) T cells specific for beta cells encounter their cognate antigens in the islets of NOD mice. Eur J Immunol 39: 2716-2724.
  20. Uibo R, Lernmark A (2008) GAD65 autoimmunity-clinical studies. Adv Immunol 100: 39-78.
  21. Ludvigsson J (2009) Therapy with GAD in diabetes. Diabetes Metab Res Rev 25: 307-315.
  22. Kaufman DL, Clare-Salzler M, Tian J, Forsthuber T, Ting GS, et al. (1993) Spontaneous loss of T-cell tolerance to glutamic acid decarboxylase in murine insulin-dependent diabetes. Nature 366: 69-72.
  23. Tisch R, Wang B, Atkinson MA, Serreze DV, Friedline R (2001) A glutamic acid decarboxylase 65-specific Th2 cell clone immunoregulates autoimmune diabetes in nonobese diabetic mice. J Immunol 166: 6925-6936.
  24. Busick RY, Aguilera C, Quinn A (2007) Dominant CTL-inducing epitopes on GAD65 are adjacent to or overlap with dominant Th-inducing epitopes. Clin Immunol 122: 298-311.
  25. Quinn A, McInerney MF, Sercarz EE (2001) MHC class I-restricted determinants on the glutamic acid decarboxylase 65 molecule induce spontaneous CTL activity. J Immunol 167: 1748-1757.
  26. Rasche S, Busick RY, Quinn A (2009) GAD65-Specific Cytotoxic T Lymphocytes Mediate Beta-Cell Death and Loss of Function. Rev Diabet Stud 6: 43-53.
  27. Gottlieb DI, Chang YC, Schwob JE (1986) Monoclonal antibodies to glutamic acid decarboxylase. Proc Natl Acad Sci U S A 83: 8808-8812.
  28. Kaufman DL, McGinnis JF, Krieger NR, Tobin AJ (1986) Brain glutamate decarboxylase cloned in lambda gt-11: fusion protein produces gamma-aminobutyric acid. Science 232: 1138-1140.
  29. Kaufman DL, Houser CR, Tobin AJ (1991) Two forms of the gamma-aminobutyric acid synthetic enzyme glutamate decarboxylase have distinct intraneuronal distributions and cofactor interactions. J Neurochem 56: 720-723.
  30. Christgau S, Schierbeck H, Aanstoot HJ, Aagaard L, Begley K, et al. (1991) Pancreatic beta cells express two autoantigenic forms of glutamic acid decarboxylase, a 65-kDa hydrophilic form and a 64-kDa amphiphilic form which can be both membrane-bound and soluble. J Biol Chem 266: 21257-21264.
  31. Carbone FR, Kurts C, Bennett SR, Miller JF, Heath WR (1998) Cross-presentation: a general mechanism for CTL immunity and tolerance. Immunol Today 19: 368-373.
  32. Ackerman AL, Giodini A, Cresswell P (2006) A role for the endoplasmic reticulum protein retrotranslocation machinery during cross-presentation by dendritic cells. Immunity 25: 607-617.
  33. Ackerman AL, Cresswell P (2004) Cellular mechanisms governing cross-presentation of exogenous antigens. Nat Immunol 5: 678-684.
  34. Harding CV, Song R (1994) Phagocytic processing of exogenous particulate antigens by macrophages for presentation by class I MHC molecules. J Immunol 153: 4925-4933.
  35. Shen L, Sigal LJ, Boes M, Rock KL (2004) Important role of cathepsin S in generating peptides for TAP-independent MHC class I cross-presentation in vivo. Immunity 21: 155-165.
  36. Quinn A, McInerney B, Reich EP, Kim O, Jensen KP, et al. (2001) Regulatory and effector CD4 T cells in nonobese diabetic mice recognize overlapping determinants on glutamic acid decarboxylase and use distinct V beta genes. J Immunol 166: 2982-2991.
  37. Dai YD, Jensen KP, Marrero I, Li N, Quinn A, et al. (2008) N-terminal flanking residues of a diabetes-associated GAD65 determinant are necessary for activation of antigen-specific T cells in diabetes-resistant mice. Eur J Immunol 38: 968-976.
  38. Krzych U, Fowler AV, Sercarz EE (1985) Repertoires of T cells directed against a large protein antigen, beta-galactosidase. II. Only certain T helper or T suppressor cells are relevant in particular regulatory interactions. J Exp Med 162: 311-323.
  39. Martinez NR, Augstein P, Moustakas AK, Papadopoulos GK, Gregori S, et al. (2003) Disabling an integral CTL epitope allows suppression of autoimmune diabetes by intranasal proinsulin peptide. J Clin Invest 111: 1365-1371.
  40. Sun D, Whitaker JN, Huang Z, Liu D, Coleclough C, et al. (2001) Myelin antigen-specific CD8+ T cells are encephalitogenic and produce severe disease in C57BL/6 mice. J Immunol 166: 7579-7587.
  41. Ford ML, Evavold BD (2005) Specificity, magnitude, and kinetics of MOG-specific CD8+ T cell responses during experimental autoimmune encephalomyelitis. Eur J Immunol 35: 76-85.
  42. Mayo S, Quinn A (2007) Altered susceptibility to EAE in congenic NOD mice: Altered processing of the encephalitogenic MOG35-55 peptide by NOR/LtJ mice. Clin Immunol 122: 91-100.
  43. Belizaire R, Unanue ER (2009) Targeting proteins to distinct subcellular compartments reveals unique requirements for MHC class I and II presentation. Proc Natl Acad Sci U S A 106: 17463-17468.
  44. Sanderson S, Shastri N (1994) LacZ inducible, antigen/MHC-specific T cell hybrids. Int Immunol 6: 369-376.
  45. Rammensee H, Bachmann J, Emmerich NP, Bachor OA, Stevanović S (1999) SYFPEITHI: database for MHC ligands and peptide motifs. Immunogenetics 50: 213-219.
  46. Quesnel A, Hsu SC, Delmas A, Steward MW, Trudelle Y, et al. (1996) Efficient binding to the MHC class I K(d) molecule of synthetic peptides in which the anchoring position 2 does not fit the consensus motif. FEBS Lett 387: 42-46.
  47. Day PM, Esquivel F, Lukszo J, Bennink JR, Yewdell JW (1995) Effect of TAP on the generation and intracellular trafficking of peptide-receptive major histocompatibility complex class I molecules. Immunity 2: 137-147.
  48. Corper AL, Stratmann T, Apostolopoulos V, Scott CA, Garcia KC, et al. (2000) A structural framework for deciphering the link between I-Ag7 and autoimmune diabetes. Science 288: 505-511.
  49. Sévère S, Gauvrit A, Vu AT, Bach JM (2007) CD8+ T lymphocytes specific for glutamic acid decarboxylase 90-98 epitope mediate diabetes in NOD SCID mouse. Mol Immunol 44: 2950-2960.
  50. Schneider SC, Ohmen J, Fosdick L, Gladstone B, Guo J, et al. (2000) Cutting edge: introduction of an endopeptidase cleavage motif into a determinant flanking region of hen egg lysozyme results in enhanced T cell determinant display. J Immunol 165: 20-23.
  51. Sercarz EE, Lehmann PV, Ametani A, Benichou G, Miller A, et al. (1993) Dominance and crypticity of T cell antigenic determinants. Annu Rev Immunol 11: 729-766.
  52. Capitani G, De Biase D, Gut H, Ahmed S, Grütter MG (2005) Structural model of human GAD65: prediction and interpretation of biochemical and immunogenic features. Proteins 59: 7-14.
  53. Norbury CC, Hewlett LJ, Prescott AR, Shastri N, Watts C (1995) Class I MHC presentation of exogenous soluble antigen via macropinocytosis in bone marrow macrophages. Immunity 3: 783-791.
  54. Norbury CC, Chambers BJ, Prescott AR, Ljunggren HG, Watts C (1997) Constitutive macropinocytosis allows TAP-dependent major histocompatibility complex class I presentation of exogenous soluble antigen by bone marrow-derived dendritic cells. Eur J Immunol 27: 280-288.
  55. Pozzi LA, Maciaszek JW, Rock KL (2005) Both dendritic cells and macrophages can stimulate naive CD8 T cells in vivo to proliferate, develop effector function, and differentiate into memory cells. J Immunol 175: 2071-2081.
  56. Hutchings P, Rosen H, O'Reilly L, Simpson E, Gordon S, et al. (1990) Transfer of diabetes in mice prevented by blockade of adhesion-promoting receptor on macrophages. Nature 348: 639-642.
  57. Burkart V, Kolb H (1996) Macrophages in islet destruction in autoimmune diabetes mellitus. Immunobiology 195: 601-613.
  58. Kurotaki T, Tamura Y, Ueda G, Oura J, Kutomi G, et al. (2007) Efficient cross-presentation by heat shock protein 90-peptide complex-loaded dendritic cells via an endosomal pathway. J Immunol 179: 1803-1813.
  59. Chen L, Jondal M (2009) TLR9 activation increases TAP-independent vesicular MHC class I processing in vivo. Scand J Immunol 70: 431-438.
  60. Delamarre L, Pack M, Chang H, Mellman I, Trombetta ES (2005) Differential lysosomal proteolysis in antigen-presenting cells determines antigen fate. Science 307: 1630-1634.
  61. Colbert JD, Matthews SP, Miller G, Watts C (2009) Diverse regulatory roles for lysosomal proteases in the immune response. Eur J Immunol 39: 2955-2965.