Figures
Abstract
Synthetic peptides are widely used in immunological research as epitopes to stimulate their cognate T cells. These preparations are never completely pure, but trace contaminants are commonly revealed by mass spectrometry quality controls. In an effort to characterize novel major histocompatibility complex (MHC) Class I-restricted β-cell epitopes in non-obese diabetic (NOD) mice, we identified islet-infiltrating CD8+ T cells recognizing a contaminating peptide. The amount of this contaminant was so small to be undetectable by direct mass spectrometry. Only after concentration by liquid chromatography, we observed a mass peak corresponding to an immunodominant islet-specific glucose-6-phosphatase catalytic subunit-related protein (IGRP)206-214 epitope described in the literature. Generation of CD8+ T-cell clones recognizing IGRP206-214 using a novel method confirmed the identity of the contaminant, further underlining the immunodominance of IGRP206-214. If left undetected, minute impurities in synthetic peptide preparations may thus give spurious results.
Citation: Brezar V, Culina S, Østerbye T, Guillonneau F, Chiappetta G, Verdier Y, et al. (2011) T Cells Recognizing a Peptide Contaminant Undetectable by Mass Spectrometry. PLoS ONE 6(12): e28866. https://doi.org/10.1371/journal.pone.0028866
Editor: Matej Oresic, Governmental Technical Research Centre of Finland, Finland
Received: August 31, 2011; Accepted: November 16, 2011; Published: December 14, 2011
Copyright: © 2011 Brezar et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This research was supported by the Juvenile Diabetes Research Foundation (JDRF grant 1-2008-106), the European Foundation for the Study of Diabetes (EFSD/JDRF/Novo Nordisk European Programme in Type 1 Diabetes Research 2007) and the Domaine d'Intérêt Majeur: Cardiovasculaire – Obésité – Diabète (Ile-de-France CODDIM). R.M. is an INSERM Avenir Investigator. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Synthetic peptides produced by fluorenylmethyloxycarbonyl (FMOC) chemistry are widely used in a variety of research fields. In immunology, the main interest of peptides is that they can be synthesized to reproduce the amino acid sequence of antigenic epitopes, which are the protein fragments recognized by T cells upon processing and presentation by antigen-presenting cells (APCs). When tailored at an optimal amino acid length, they do not require uptake and processing by APCs, as they can directly bind major histocompatibility complex (MHC) Class I (8–11 amino acid length) and Class II (∼12–20 amino acid length) molecules on the APC surface for immediate presentation to CD8+ and CD4+ T cells, respectively.
Another major attractiveness is that production of peptides in suitable quantities and purity grades is much easier than that of recombinant antigens, which frequently carry over small protein or endotoxin contaminants from the bacterial, yeast or baculovirus systems where they are generated [1]. Thus, peptides are employed in different immunological studies both in vitro, to stimulate epitope-specific T cells, and in vivo, as immunizing agents in mice and as vaccination antigens in human trials [2].
Similarly, peptides are largely used in type 1 diabetes research to pinpoint the molecular targets of autoreactive T cells [3] which ignite recognition and destruction of insulin-producing pancreatic β cells [4]. Once identified, these epitopes are invaluable reagents to track the corresponding T cells and to neutralize their pathogenic potential through suitable tolerogenic delivery [5].
Despite these major advantages, even synthetic peptides cannot be produced to complete purity. This can frequently lead to misleading results, which are seldom reported in the literature [6], [7], [8]. To minimize these drawbacks, the workflow of peptide synthesis includes quality control steps, where purity is assessed through high performance liquid chromatography (HPLC) and mass spectrometry (MS) analyses before delivery to the final users. These analyses are frequently useful a posteriori to troubleshoot unexpected results, as contaminant peaks identified by MS can be scrutinized more carefully.
In this report, we describe an instance of a peptide contaminant which went undetected in several quality control MS analyses. Nonetheless, this minute contaminant was capable of eliciting peptide-specific T-cell responses in vitro. The troubleshooting procedure demonstrating the involvement of such contaminant in the observed results is described, including a novel technique to select antigen-specific CD8+ T-cell clones from infiltrated islets of non-obese diabetic (NOD) mice.
Results
Identification of a MHC Class I-restricted candidate epitope in NOD mice
We set up to investigate the kinetics of CD8+ T-cell IFN-γ responses to a selected panel of previously described β-cell epitopes [9]. After isolating the islet infiltrates and culturing them for 7 d in the presence of IL-2 to enrich the population of CD8+ T cells, we observed that IFN-γ responses were very weak vis à vis most peptides, except for 3 of them: IGRP206-214, PIB15-23 and PIA7-21 (Fig. 1A–B). As reported [10], responses to IGRP206-214 and its mimotope NRP-V7 were detectable in NOD mice of different ages, while PIB15-23-reactive T cells were only detected until 12 weeks (Fig. 1B).
(A) Representative staining showing IFN-γ responses of spontaneous NOD islet-infiltrating CD8+ T cells recalled in vitro with different peptides. Percent cells present in each quadrant are given. (B) IFN-γ-producing CD8+ T-cell responses in islets from NOD mice of different ages (n = 4/group) following in vitro peptide stimulation. (C) PIA7-21-specific responses in islet infiltrates originate from CD8+, but not from CD4+ T cells. The percent of IFN-γ+ CD4+ and CD8+ T cells after in vitro peptide recall (n = 4/group) is shown.
While IGRP206-214 and PIB15-23 are well described Class I Kd-restricted epitopes [11], [12], PIA7-21 has been previously described as a Class II-restricted epitope [13]. To confirm that the observed PIA7-21-induced responses were Class I-restricted, we measured IFN-γ release upon peptide stimulation in both CD4+ and CD8+ T cells and found that only CD8+ T cells responded to PIA7-21 (Fig. 1C).
As the PIA7-21 peptide was unexpectedly long to bind to MHC Class I registries, we hypothesized that there could be a shorter variant either present in the peptide preparation or processed during peptide pulsing. However, testing of potential 9–11-mer core epitopes within the PIA7-21 sequence, either in spontaneous islet infiltrates or in splenocytes following peptide immunization, did not yield any similar IFN-γ response (data not shown). Indeed, we found two epitopes, PIA9-17 and PIA13-21, which showed significant T-cell responses after peptide immunization, but no spontaneous T-cell responses were observed in islet infiltrates (Fig. 1C).
New peptide syntheses fail to reproduce the same reactivity
The PIA7-21 peptide was resynthesized by another manufacturer at two different purity grades (75.7% and 96.9%). Neither of these new peptide preparations was capable of eliciting T-cell responses similar to our previous synthesis (Fig. 2). We thus hypothesized that the observed reactivity could be due to a contaminant present in the first peptide preparation.
IFN-γ production by islet-infiltrating CD8+ T cells following in vitro stimulation with 3 different preparations of PIA7-21 peptide. The percent of IFN-γ+CD8+ cells is given.
The original PIA7-21 preparation contains a PIB15-23 contaminant
A first MS analysis on the original PIA7-21 preparation highlighted a 1009,51 m/z mass peak compatible with the MH+ adduct of the PIB15-23 peptide (Fig. 3A). However, this contaminant could not explain the observed T-cell responses since, as shown in Fig. 1B, the PIB15-23 peptide does not induce T-cell responses as strong as contaminated PIA7-21 and is not recognized in islet infiltrates of older NOD mice. To definitely rule out the involvement of PIB15-23, we used the contaminated PIA7-21 preparation to stimulate PIB15-23-specific splenocytes from G9C8 transgenic mice expressing a PIB15-23-specific T-cell receptor [14]. No response was observed (data not shown), suggesting that the amount of PIB15-23 contaminant was insufficient to elicit T-cell responses. Thus, we hypothesized that there could be another peptide contaminating the original PIA7-21 preparation.
(A) MS analysis: a contaminant with a mass peak of 1009,516 compatible with the MH+ adduct of PIB15-23 is detected in the contaminated PIA7-21 preparation. (B) One hundred microliters of a 50 µM solution of contaminated PIA7-21 peptide were subjected to nanochromatography and separated in 96 fractions. Each of these fractions was tested for its ability to induce IFN-γ responses in islet-infiltrating CD8+ T cells. Results obtained with one single positive fraction compared with one of the remaining 95 negative fractions are shown. (C) The T-cell-positive fraction identified was reanalyzed by MS. The obtained mass list shows two species at 1009,516 and 1096,619 m/z, compatible with the MH+ adduct of PIB15-23 and IGRP206-214, respectively.
Fractionation of the original PIA7-21 preparation reveals a mass peak corresponding to IGRP206-214 that elicits IGRP206-214-specific T-cell responses
To concentrate other possible peptide contaminant(s), we used HPLC fractionation. Ninety-six fractions were collected, concentrated 10-fold and subsequently tested for CD8+ T-cell IFN-γ responses from islet infiltrates. Only one of these 96 fractions was able to elicit a T-cell response (Fig. 3B). This fraction was reanalyzed by MS (Fig. 3C). The previously identified 1009.51 m/z peak corresponding to PIB15-23 was again detected, along with a 1096.62 m/z peak not previously visualized in the unfractionated PIA7-21 sample (compare with Fig. 3A). This peak is compatible with the MH+ adduct of the IGRP206-214 peptide. One further MS analysis attempted to conclusively identify IGRP206-214 but could only sequence part of the peptide (VYLKTNVFL) (Fig. 4). This is probably due to the extremely low amount of this contaminant.
(A) MS analysis of the contaminated PIA7-21 preparation. (B) MS/MS analysis of the 1096.34 m/z compatible with the MH+ adduct of IGRP206-214.
To verify whether this minute IGRP206-214 peptide contaminant was responsible for the observed T-cell responses, we performed a parallel IFN-γ/TMr staining. We stimulated cultured islet infiltrates with the contaminated PIA7-21 preparation and double-stained them for intracellular IFN-γ and NRP-V7-loaded Kd TMrs (Fig. 5). The same CD8+ T cells which secreted IFN-γ in response to the contaminated PIA7-21 preparation also bound the NRP-V7 TMr (Fig. 5A), similar to IGRP206-214-stimulated cells (Fig. 5B). Staining with a control TUM-loaded Kd TMr or stimulation with peptide diluent alone did not yield any IFN-γ/TMr double-positive population (Fig. 5C–D). This experiment proves that the IGRP206-214 contaminant present in our PIA7-21 peptide sample is capable of eliciting IGRP206-214-specific CD8+ T-cell responses. We can also exclude the presence of other contaminants capable of stimulating CD8+ T cells, as ∼90% of contaminant-stimulated IFN-γ+ cells were also NRP-V7 TMr+ (Fig. 5A).
Parallel IFN-γ and Kd TMr staining after stimulation with contaminated PIA7-21 or IGRP206-214 peptide. (A) Stimulation with contaminated PIA7-21 peptide and NRP-V7 Kd TMr staining. (B) Stimulation with IGRP206-214 peptide and NRP-V7 Kd TMr staining. (C) Stimulation with contaminated PIA7-21 peptide and TUM Kd TMr negative control staining. (D) Stimulation with peptide diluent alone and NRP-V7 Kd TMr staining.
CD8+ T-cell clones raised against the contaminated PIA7-21 peptide preparation are IGRP206-214- specific
To finally prove that IGRP206-214 was the contaminant inducing T-cell responses, we expanded and cloned CD8+ T cells responding to our contaminated peptide. A diagram of the cloning procedure is shown in Fig. 6A. Total islet infiltrates were initially stimulated and expanded using islets from NOD-RIPB7 mice as a source of antigen. These islets further express the co-stimulatory molecule B7.1 [15]. The expanded infiltrates were subsequently challenged with the contaminated PIA7-21 preparation, stained with the cytotoxic granule marker CD107a/b [16] and sorted based on positive staining (Fig. 6B). Upon this peptide recall, 19.5% of islet-stimulated CD8+ T cells responded by CD107a/b upregulation. This fraction was single-cell-sorted. After one further round of stimulation with RIPB7 islets, 3 growing wells were identified out of the 60 seeded (5.0%). These 3 clones were tested for specificity by TMr staining (Fig. 6C). All of them stained uniformly positive for the NRP-V7 Kd TMr, but not for the PIB15-23 nor for the control TUM Kd TMr.
(A) Schematic of the T-cell expansion and cloning technique. (B) Sorting of epitope-reactive CD8+ T cells based on CD107a/b upregulation after in vitro peptide recall. Left plot shows negative control (no stimulation); right plot shows CD107a/b staining after stimulation with contaminated PIA7-21 peptide. (C) Representative TMr staining of one of the 3 T-cell clones generated. Staining for NRP-V7, PIB15-23 and negative control TUM Kd TMr is shown from left to right.
Discussion
When we set out to investigate islet-infiltrating CD8+ T-cell responses to a panel of different β-cell peptides, we identified a candidate Class I-restricted epitope derived from the insulin A chain. New preparations of the same peptide did not however elicit CD8+ T-cell responses, and MS analyses on concentrated fractions revealed that the original preparation was contaminated with an IGRP206-214 epitope that was responsible for the observed responses. When validating epitope panels for T-cell recognition, it is therefore important that strong responses measured against a given peptide are reproduced on a separate preparation to rule out these pitfalls.
This IGRP206-214 contaminant was introduced during peptide synthesis and purification and not during laboratory manipulation. Indeed, peptide sampling from different portions of the vial including its very bottom gave similar T-cell responses, which would not be the case if cross-contamination from an IGRP206-214 preparation occurred during weighing of the lyophilized powder. Further supporting this conclusion, the original PIA7-21 preparation, although >90% pure, also contained traces of other peptides originated from insulin (i.e. PIB15-23) that were detected by MS. Impurities in synthetic peptides are mostly accounted for by products of deletion, additional residue incorporation, truncation and incomplete de-protection in solid phase synthesis [17]. This is not the case for our findings, since peptide contaminants corresponded to different amino acid sequences. Thus, the most likely possibility is that the peptide preparation was contaminated by improper handling of HPLC equipment during purification. Insufficient washing steps may have left traces of other peptides previously purified on the same column. Indeed, the manufacturer's log records revealed that PIB15-23 and IGRP206-214 preparations were purified on the same day, by the same operator using the same equipment.
The main interest of this case report is that the IGRP206-214 peptide contaminant remained undetected by conventional MS analyses, and a compatible mass peak could finally be revealed only by concentrating the sample 10-fold. This emphasizes the exquisite sensitivity of T cells in sensing even minute amounts of their cognate antigens. We used different technical approaches to identify this contaminant. To conclusively prove that IGRP206-214 was eliciting CD8+ T-cell responses, we produced clones selected based on upregulation of the cytotoxic granule marker CD107a/b [16] following peptide exposure. This approach is attractive for cloning epitope-specific CD8+ T cells for which the corresponding TMrs are not available. Indeed, very few CD8+ T-cell clones recognizing β-cell epitopes have been isolated from spontaneous islet infiltrates in the NOD mouse [18], [19], [20].
Our results further emphasize the impressive immunodominance of the IGRP206-214 epitope within the NOD islet infiltrates. They also exemplify how a fraction of IGRP206-214-specific CD8+ T cells infiltrating the islets is of very high avidity, as it was capable of responding to minute amounts of peptide.
In conclusion, high purity peptides may hide traces of contaminating peptides potentially leading to experimental pitfalls. Functional T-cell assays are extremely sensitive to these impurities, which may be missed by MS analysis. While this sensitivity makes T-cell assays particularly vulnerable, it can also provide means to conclusively identify these contaminants.
Materials and Methods
Peptides
Proinsulin (PI)B24-C1 (FFYTPMSRRE), PIC15-30 (SFGDLQTLALEVARQK), islet-specific glucose-6-phosphatase catalytic subunit-related protein (IGRP)21-29 (TYYGFLNFM), glutamic acid decarboxylase (GAD)206-220 (TYEIAPVFVLLEYVT), GAD247-266 (NMYAMLIARYKMFPEVKEKG), and GAD524-543 (SRLSKVAPVIKARMMEYGTT) were produced by Schafer-N. PIB24-C36 (FFYTPMSRREVED), PIC32-A12 (KRGIVDQCCTSICS) and non-contaminated PIA7-21 peptides (CTSICSLYQLENYCN) were produced by GL Biochem. All these peptides were >80% pure. IGRP206-214 (VYLKTNVFL), PIB15-23 (LYLVCGERG), dystrophia myotonica kinase (DMK)138-146 (FQDENYLYL) and contaminated PIA7-21 (CTSICSLYQLENYCN) were produced by “manufacturer X” at >90% purity.
Mice
NOD mice were bred and housed in specific pathogen-free conditions. NOD mice were crossed with C57BL/6J-RIP-B7-1 mice [15] to obtain a first generation (F1) of offspring that was used to prepare islet feeders [18]. G9C8 transgenic mice expressing the T-cell receptor from a G9C8 CD8+ clone recognizing the PIB15-23 peptide [14] were also used. All experiments were conducted according to ethic guidelines. This study was approved by the local ethics committee for animal experimentation (Comité régional d'éthique pour l'expérimentation animale, approval number: P2.AL.116.09).
Preparation of pancreatic islet cell infiltrates
Mice were sacrificed and pancreatic islets isolated after perfusion with 0.75 mg/ml collagenase P (Roche), as described [21], [22]. To measure antigen-specific CD8+ T-cell responses, islets were isolated and cultured for 7 d in the presence of 50 U/ml recombinant human interleukin (IL)-2 (R&D). Infiltrating cells were then collected and analyzed.
Flow cytometry
Cells were stained with PE-labeled Kd tetramers (TMrs) synthesized by the National Institute of Health Tetramer Core Facility and loaded with NRP-V7 (a mimotope of the IGRP206-214 epitope; KYNKANVFL) [11] and control TUM peptide (KYQAVTTTL). Peptide-stimulated interferon (IFN)-γ responses were measured using intracellular IFN-γ staining after incubation for 5 h in the presence of 10 µM peptide and 10 µg/ml Brefeldin A (Sigma). Phorbol myristate acetate (PMA)/ionomycin (1 µg/ml/each) were used as a polyclonal positive control stimulus. CD8-Alexa Fluor 700 (clone 53-6.7; eBioscience), CD4-PerCP (clone RM4-5; BD), IFN-γ-PE (clone XMG1.2; BD), CD45-FITC (clone 30-F11; BD) and CD107a/b-Alexa Fluor 647 (clone 1D4B; eBioscience) were used. For intracellular IFN-γ staining, Cytofix/Cytoperm kit (BD) was used.
Generation of epitope-specific CD8+ T-cell clones
Isolated islets from prediabetic NOD mice were cultured for 7 d in the presence of 50 U/ml recombinant human IL-2.
Restimulation.
On the 7th day of culture, 8–10 irradiated (2,500 rad) islets/well from NOD-RIPB7 (F1) mice were added and cultured for 7 d in the presence of 50 U/ml IL-2. Fresh IL-2 (50 U/ml final concentration) was added every 3 d. Fresh medium was added when needed (every 7 d throughout the cloning). Fourteen days after the addition of NOD-RIPB7 islets, another restimulation was performed in the same way.
Sorting of epitope-specific CD8+ T cells.
14 d after the second restimulation, recovered cells were stained with CD107a/b in the presence of 10 µM peptide and 0.7 µM monensin (Sigma) for 5 h at 37°C. One hundred CD8+CD107a/b+ cells/well were sorted in 96-well V-bottom plates containing 10 fresh irradiated islets from NOD-RIPB7 mice as feeders. IL-2 (50 U/ml) was added and replenished every 3 d at the same final concentration. Fresh medium was added approximately every 7 d.
Nanochromatography
Peptide solutions (5 µl, 10 mM) were diluted 1∶200 in 10% acetonitrile (ACN; Carlo Erba), 0.1% trifluoroacetic acid (TFA; Pierce). These preparations were either directly analysed by matrix-assisted laser desorption/ionisation (MALDI) MS as described below or separated and concentrated by reverse phase nanochromatography as follows. Ten microliters of diluted peptides were injected and separated with an Ultimate 3000 series HPLC (Dionex). Injected peptides were trapped using solvent A (0.1% TFA, 2% ACN) at 30 µl/min loading flow for 3 min in a C18 trap column (Acclaim reverse phase C18 pepmap 100 phase, 5 µm particles, 100 Å pores, 5 mm length, 300 µm internal diameter). The microvalve then oriented the elution to back flow (300 nl/min) the peptides towards the analytical column (C18 pepmap 100, 3 µm particle, 100 Å pores, 15 cm length, 75 µm i.d.) with a gradient rising from 10% solvent B (80% ACN, 20% solvent A) at microvalve switch to 60% B in 34 min. Fractions (20 s/each, i.e. 100 nl) were deposited in 96 wells of a 384-well plate using a Probot fraction collector (Dionex). In order to increase the concentration of collected peptides in each well, the injection/fractionation was repeated 20 times and collected in the same 96 plate positions to accumulate identical fractions. Fractions were subsequently dehydrated in a speed vacuum concentrator (Eppendorf). After solubilization in 5 µl 10% ACN 0.1% TFA, each well represented a 10-fold concentration factor of the initial peptide solution. Half microliter of each fraction was deposited on a MALDI target to perform MS analysis.
The initial peptide dilution or the concentrated fractions were mixed 1∶1 on a MALDI target plate with 1 µl of 5 mg/ml of α-cyano-4-hydroxycinnamic acid (CHCA) matrix (Laser Biolabs) in 70% ACN 0.1% TFA and 60 nM Glu-fibrinopeptide (Sigma) as internal standard and analysed using a 4800 MALDI time-of-flight (TOF) TOF analyser (ABSciex).
MS
Spectra acquisition and processing were performed using the 4000 series explorer software (ABSciex) version 3.5.28193 build 1011 in positive reflectron mode at fixed laser fluency with low mass gate and delayed extraction. External plate calibration was performed using 4 calibration spots at each corner of the plate. Additional internal calibration was performed using Glu-fibrinopeptide (m/z 1570.677). For each acquisition, 10 increments of 50 spectra in the range of m/z 600 to 4500 were acquired at a 200 Hz laser shot frequency. Five-hundred MS spectra per sample were summed and processed to obtain monoisotopic values from isotope clusters with a raw spectra s/n ratio of 20.
MS/MS
In each MS spectrum, the 12 most abundant peaks were selected for fragmentation, starting with the least abundant. A cut-off was applied at a minimum s/n of 20. Neighboring precursors within 200 resolution were excluded. 1000 MS/MS spectra per precursor were summed by increments of 50. Processing included baseline subtraction and Stavitsky Golay smoothing with 3 points across peak and a polynomial order of 4. Generated peaklists reflected monoisotopic values from isotope clusters with a minimum s/n ratio of 22.
Peptide identification
Peak lists were submitted to an in-house mascot (Matrix science) version 2.2 search engine [23]. The Swiss-Prot database release 2010_05 (516603 sequences; 181919312 residues) was used with restriction to mammalia (60,000 sequences). No enzymes were selected as cleavage specificity. Parent and fragment mass tolerances were respectively set to 50 ppm and 0.3 Da; variable modification (oxidation) of methionines was allowed. A filter was applied to the search in order to reduce false positives and matching redundancies of the same peptide in several hits. All matches above 5% risks of random matching were eliminated (p<0.05). Peptide lists only included proteins matching the same set of peptide and, if shared subsets were present, at least one non-shared peptide was required. The minimal peptide score was 25. With these parameters, the minimal protein score was 50.
Acknowledgments
We wish to thank the National Institute of Health Tetramer Core Facility for production of Kd tetramers. Servier Medical Art was used for Fig. 6A.
Author Contributions
Conceived and designed the experiments: VB RM. Performed the experiments: VB SC TO FG GC YV. Analyzed the data: VB FG GC YV RM. Contributed reagents/materials/analysis tools: TO JV FSW SB. Wrote the paper: VB RM.
References
- 1. Mallone R, Mannering SI, Brooks-Worrell BM, Durinovic-Bello I, Cilio CM, et al. (2011) Isolation and preservation of peripheral blood mononuclear cells for analysis of islet antigen-reactive T cell responses: position statement of the T-Cell Workshop Committee of the Immunology of Diabetes Society. Clin Exp Immunol 163: 33–49.
- 2. Culina S, Boitard C, Mallone R (2011) Antigen-based immune therapeutics for type 1 diabetes: magic bullets or ordinary blanks? Clin Dev Immunol 2011: 286248.
- 3. Mallone R, Brezar V, Boitard C (2011) T cell recognition of autoantigens in human type 1 diabetes: clinical perspectives. Clin Dev Immunol 2011: 513210.
- 4. Brezar V, Carel JC, Boitard C, Mallone R (2011) Beyond the hormone: insulin as an autoimmune target in type 1 diabetes. Endocr Rev 32: 623–669.
- 5. Mallone R, Nepom GT (2005) Targeting T lymphocytes for immune monitoring and intervention in autoimmune diabetes. Am J Ther 12: 534–550.
- 6. de Graaf MT, de Beukelaar JW, Burgers PC, Luider TM, Kraan J, et al. (2008) Contamination of synthetic HuD protein spanning peptide pools with a CMV-encoded peptide. Cytometry A 73: 1079–1085.
- 7. Currier JR, Galley LM, Wenschuh H, Morafo V, Ratto-Kim S, et al. (2008) Peptide impurities in commercial synthetic peptides and their implications for vaccine trial assessment. Clin Vaccine Immunol 15: 267–276.
- 8. Mannering SI, Purcell AW, Honeyman MC, McCluskey J, Harrison LC (2003) Human T-cells recognise N-terminally Fmoc-modified peptide. Vaccine 21: 3638–3646.
- 9. Di Lorenzo TP, Peakman M, Roep BO (2007) Translational mini-review series on type 1 diabetes: Systematic analysis of T cell epitopes in autoimmune diabetes. Clin Exp Immunol 148: 1–16.
- 10. Amrani A, Verdaguer J, Serra P, Tafuro S, Tan R, et al. (2000) Progression of autoimmune diabetes driven by avidity maturation of a T-cell population. Nature 406: 739–742.
- 11. Lieberman SM, Evans AM, Han B, Takaki T, Vinnitskaya Y, et al. (2003) Identification of the beta cell antigen targeted by a prevalent population of pathogenic CD8+ T cells in autoimmune diabetes. Proc Natl Acad Sci USA 100: 8384–8388.
- 12. Wong FS, Karttunen J, Dumont C, Wen L, Visintin I, et al. (1999) Identification of an MHC class I-restricted autoantigen in type 1 diabetes by screening an organ-specific cDNA library. Nat Med 5: 1026–1031.
- 13. Daniel D, Wegmann DR (1996) Protection of nonobese diabetic mice from diabetes by intranasal or subcutaneous administration of insulin peptide B-(9-23). Proc Natl Acad Sci USA 93: 956–960.
- 14. Wong FS, Siew LK, Scott G, Thomas IJ, Chapman S, et al. (2009) Activation of insulin-reactive CD8 T cells for development of autoimmune diabetes. Diabetes 58: 1156–1164.
- 15. Wong S, Guerder S, Visintin I, Reich EP, Swenson KE, et al. (1995) Expression of the co-stimulator molecule B7-1 in pancreatic beta-cells accelerates diabetes in the NOD mouse. Diabetes 44: 326–329.
- 16. Betts MR, Brenchley JM, Price DA, De Rosa SC, Douek DC, et al. (2003) Sensitive and viable identification of antigen-specific CD8+ T cells by a flow cytometric assay for degranulation. J Immunol Methods 281: 65–78.
- 17. De Spiegeleer B, Vergote V, Pezeshki A, Peremans K, Burvenich C (2008) Impurity profiling quality control testing of synthetic peptides using liquid chromatography-photodiode array-fluorescence and liquid chromatography-electrospray ionization-mass spectrometry: the obestatin case. Anal Biochem 376: 229–234.
- 18. Wong FS, Visintin I, Wen L, Flavell RA, Janeway CA Jr (1996) CD8 T cell clones from young nonobese diabetic (NOD) islets can transfer rapid onset of diabetes in NOD mice in the absence of CD4 cells. J Exp Med 183: 67–76.
- 19. DiLorenzo TP, Graser RT, Ono T, Christianson GJ, Chapman HD, et al. (1998) Major histocompatibility complex class I-restricted T cells are required for all but the end stages of diabetes development in nonobese diabetic mice and use a prevalent T cell receptor alpha chain gene rearrangement. Proc Natl Acad Sci USA 95: 12538–12543.
- 20. Santamaria P, Utsugi T, Park BJ, Averill N, Kawazu S, et al. (1995) Beta-cell-cytotoxic CD8+ T cells from nonobese diabetic mice use highly homologous T cell receptor alpha-chain CDR3 sequences. J Immunol 154: 2494–2503.
- 21. Jarchum I, Takaki T, DiLorenzo TP (2008) Efficient culture of CD8(+) T cells from the islets of NOD mice and their use for the study of autoreactive specificities. J Immunol Methods 339: 66–73.
- 22. Diana J, Brezar V, Beaudoin L, Dalod M, Mellor A, et al. (2011) Viral infection prevents diabetes by inducing regulatory T cells through NKT cell-plasmacytoid dendritic cell interplay. J Exp Med 208: 729–745.
- 23. Perkins DN, Pappin DJ, Creasy DM, Cottrell JS (1999) Probability-based protein identification by searching sequence databases using mass spectrometry data. Electrophoresis 20: 3551–3567.