There is accumulating evidence that pathogenic T cells in T1D recognize epitopes formed by post-translational modifications of β-cell antigens, including hybrid insulin peptides (HIPs). The ligands for several CD4 T-cell clones derived from the NOD mouse are HIPs composed of a fragment of proinsulin joined to peptides from endogenous β-cell granule proteins. The diabetogenic T-cell clone BDC-6.9 reacts to a fragment of C-peptide fused to a cleavage product of pro-islet amyloid polypeptide (6.9HIP). In this study, we used a monoclonal antibody (MAb) to the 6.9HIP to determine when and where HIP antigens are present in NOD islets during disease progression and with which immune cells they associate. Immunogold labeling of the 6.9HIP MAb and organelle-specific markers for electron microscopy were employed to map the subcellular compartment(s) in which the HIP is localized within β-cells. While the insulin B9-23 peptide was present in nearly all islets, the 6.9HIP MAb stained infiltrated islets only in NOD mice at advanced stages of T1D development. Islets co-stained with the 6.9HIP MAb and antibodies to mark insulin, macrophages, and dendritic cells indicate that 6.9HIP co-localizes within insulin-positive β-cells as well as intra-islet antigen-presenting cells (APCs). In electron micrographs, the 6.9HIP co-localized with granule structures containing insulin alone or both insulin and LAMP1 within β-cells. Exposing NOD islets to the endoplasmic reticulum (ER) stress inducer tunicamycin significantly increased levels of 6.9HIP in subcellular fractions containing crinosomes and dense-core granules (DCGs). This work demonstrates that the 6.9HIP can be visualized in the infiltrated islets and suggests that intra-islet APCs may acquire and present HIP antigens within islets.
A broad variety of self-antigens have been implicated in autoimmune diabetes, but only a subset of these is likely to be involved in disease initiation, with subsequent epitope spreading to other antigens. In the NOD mouse, substantial evidence supports proinsulin as the predominant instigating antigen (1–3). Efforts to define which proinsulin-derived epitopes are most pertinent to disease etiology have included a wide variety of approaches, including genetics, biochemistry, confocal and electron microscopy (EM), and targeted mass spectrometry. These studies have revealed different forms of insulin-derived epitopes, including native epitopes of insulin B-chain and C-peptide, as well as a class of modified antigens, hybrid insulin peptides (HIPs) resulting from peptide fusion, all of which contribute to the activation of pathogenic CD4 T cells.
Insulin secretory granules are considered a repository of antigenic targets involved in autoimmune diabetes (4). However, the subcellular localization of immunogenic insulin-derived epitopes in β-cell granules remains largely unexplored. The primary secretory granule in β-cells is the insulin dense-core granule (DCG), containing a crystal core of insulin. The DCG has diverse biological functions, such as proinsulin processing and insulin secretion. Maintaining insulin homeostasis in β-cells is achieved by fusing excessive DCGs with lysosomes for catabolic reduction. This crinophagic pathway generates a minor set of β-cell granules, termed crinophagic bodies or crinosomes, which possess lysosomal activities to catabolize secretory proteins in DCGs (5–7).
The subcellular distribution of insulin-derived antigens that elicit the most robust T-cell responses within β-cells has been probed using differential centrifugation to fractionate DCGs and crinosomes with subsequent mass spectrometry to identify the peptides therein (8). The DCG fraction obtained at a high speed (25,000 × g) primarily contained intact insulin, long B-chain peptides, and some C-peptides, whereas the granule fraction obtained at a lower speed (5,000 × g), representing crinosomes, was enriched for shorter insulin peptides, including shorter B-chain peptides associated with the predominant epitope B9-23 and diverse C-peptide sequences. Further investigation of crinosome-derived peptides, initially by manual investigation of unassigned spectra, revealed a few HIP sequences (8), suggesting the presence of HIPs in crinosomes. These results were later confirmed in a subsequent study (9) by searching the crinosome peptidome against an in silico HIP database (10). This approach also identified HIPs in the secretome (peptide contents exocytosed from β-cells upon glucose stimulation) (9), raising the possibility that HIPs are also present in DCGs.
HIPs are the product of a unique post-translational modification whereby insulin peptide fragments are fused to sequences from endogenous proteolytically processed β-cell secretory granule proteins. One such HIP, the 6.9HIP, is composed of a fragment of C-peptide fused to IAPP2 (DLQTLAL/NAARD) and is the antigenic ligand of diabetogenic CD4+ T-cell clone BDC-6.9 from the NOD mouse (11). T cells reactive to the human counterpart of this HIP and others have been identified in PBMCs from T1D patients (12) and distinct HIP-reactive T cells have been cultured from the islets of deceased T1D organ donors (13–16). Importantly, a nearly identical sequence to the 6.9HIP (LQTLAL/NAARD) was identified in the MHC-II (I-Ag7) peptidomes of both pancreatic islets and draining lymph nodes in NOD mice (9), indicating that 6.9HIP is a bona fide neoepitope that is presented in vivo by MHC-II.
Mapping the intracellular site of HIP formation is challenging, yet essential for putative intervention strategies. Crinosomes and DCGs are both favorable environments for HIP formation due to high concentrations of the peptide components of HIPs and enzymes that function at low pH. Proposed mechanisms for HIP formation via enzymes cathepsin D and cathepsin L have been reported from in vitro studies (17, 18). A monoclonal antibody (MAb) specific for the insulin B-chain sequences B9-23, but not native insulin, has been employed to localize the chief epitope B9-23 to crinosomes (8). The B9-23 MAb co-localizes with the lysosomal marker LAMP1 in vesicles distinct from DCGs, providing a punctate staining pattern with immunofluorescent microscopy. Diabetogenic effector T cells reactive to the B12-20 epitope are responsive to the contents of these crinosomes, suggesting that other critical antigens could be present in the same organelles (8, 19). In the current study, we developed a MAb to the C-peptide/IAPP2 (6.9) HIP, previously identified by mass spectrometry, to determine whether such rare post-translationally modified peptides can be detected through microscopic analysis of antibody staining. The primary goals were to establish in which cellular compartment the 6.9HIPs are localized and with what type of antigen-presenting cells (APCs) they may be associated.
NOD/ShiLtJ (NOD) mice were originally obtained from The Jackson Laboratory. All mice were bred, maintained, and used in experiments in our pathogen-free animal facility in accordance with the guidelines of the Division of Comparative Medicine at Washington University School of Medicine (Association for Assessment and Accreditation of Laboratory Animal Care accreditation no. A3381-01). NOD.IAPP−/− mice were previously bred in the Haskins mouse colony by backcrossing C57BL6.IAPP−/− mice (20) onto the NOD background (21).
BALB/c mice were immunized three times, 2 weeks apart with 100 µ g/each of the NOD 6.9HIP (SLDQLALNAARDPN) conjugated to KLH and combined 1:1 with CFA, IFA, and IFA, respectively. Sera were monitored for MAb titer with enzyme-linked immunosorbent assays (ELISAs) and radio binding assays [as in (22)] using a recombinant probe expressing the NOD 6.9HIP. A MAb targeting the NOD 6.9HIP sequence was developed as previously described (23). Specificity of the 6.9HIP MAb was determined using Western blots and in-house ELISAs, using anti-mouse HRP 1:10,000 (Pierce) and quantitation with o-phenylenediamine dihydrochloride substrate at 490 nm as per vendor recommendations (Sigma).
Mouse HIPs were cloned in-frame with fusion partner NUS in vector pET43a (Novagen) for expression in bacteria. Induced proteins were purified on Ni-Agarose (Qiagen) according to the manufacturer’s instructions. 6.9HIP and control peptides (>95% purity, Genscript or Peptide 2.0) were resuspended at a concentration of 10 mg/mL according to solubility recommendations.
Ninety-six-well ELISA plates were coated with 6.9HIP peptide (LQTLAL/NAARD, 2 μ M) in carbonate buffer overnight at 4°C. Plates were washed and subsequently blocked with DMEM/5% fetal bovine serum (FBS). The 6.9HIP MAb was serially diluted and directly added to the coated plate. Horseradish peroxide (HRP)-conjugated goat anti-mouse IgG (1:10,000) antibody was then added for 2 h at 4°C, the responses were measured using the OptEIA TMB Substrate (BD), and the data [optical density (OD) at 450 nm] were collected using an iMark Microplate Reader (Bio-Rad Laboratories).
ELISA plates were coated with 6.9HIP peptide (2 μ M) in carbonate buffer overnight at 4°C. Plates were washed and subsequently blocked with DMEM/5% FBS. Soluble competitive inhibitors (synthetic peptides) were pre-incubated with the 6.9HIP MAb (50 ng/mL) for 30 min and the mixture was incubated with the plate-bound antigens for 2 h at 37°C. In the absence of soluble competitive inhibitors, the concentration of the MAb resulted in approximately 50% binding to the plate-bound antigen. HRP-conjugated goat anti-mouse IgG (1:10,000) antibody was then added for 2 h at 4°C, the responses were quantified using the OptEIA TMB Substrate (BD), and the data (OD at 450 nm) were collected using an iMark Microplate Reader (Bio-Rad Laboratories).
Mouse peritoneal cavities were opened to clamp the common bile duct leading to the duodenum. Type XI collagenase (0.4 mg/mL in serum-free DMEM; Sigma-Aldrich) digestion buffer was injected through the bile duct to inflate the pancreas, which was removed and incubated in digestion buffer at 37°C (12 min) and then shaken vigorously for 90 s. Pancreata were washed with serum-free (SF) DMEM (3×) through a stainless-steel strainer. The flow-through, containing islets, was then filtered through a 70-µ m cell strainer. Retained islets were then flushed into a Petri dish for hand-picking. The NOD.IAPP−/− islets were prepared similarly by infusion of pancreata with CIzyme RI (VitaCyte) through the bile duct followed by gravity filtration washes (as above). Islets were enriched with gradient centrifugation using Lympholyte 1.1 (Cedarlane Laboratories) where they concentrate at the interface and were hand-dissected. Pure hand-picked islets were dispersed using Cell Dissociation Solution Non-Enzymatic (Sigma-Aldrich) for 10 min at 37°C.
Intact islets were fixed in 4% formaldehyde, permeabilized, and blocked in 0.2% saponin/3% bovine serum albumin (BSA). Primary 6.9HIP or AIP antibody (2 µ g/mL) was used to stain islets overnight in 0.2% saponin/3% BSA. Anti-mouse-AF594 secondary antibody (4 µ g/mL) was permitted to bind islets for 1 h in saponin/BSA. Islets were washed with saponin/BSA and mounted on slides with prolong diamond antifade mounting media (Invitrogen). Islets were blocked with saponin/BSA and then stained with antibodies to cell surface markers (CD11c-488 or CD4-BV480, SIRP α-AF488, and F4/80-AF594) at 2 µ g/mL for 1 h followed by washing with saponin/BSA for imaging. For staining with conjugated 6.9HIP MAb, islets were fixed in 4% formaldehyde, permeabilized and blocked with saponin/BSA, and stained with 6.9HIP-AF647 at 0.033 µ g/mL overnight in saponin/BSA. Islets were washed and mounted as above for imaging using an inverted Leica SP8 confocal scanning microscope with Leica’s LAS X software. The microscope was equipped with a 63× 1.40 numerical aperture (NA) oil-immersion objective and a white light laser. Optical sections were taken every 0.75 µ m. Images shown were taken with Imaris 9.0 software.
For immunolocalization at the ultrastructural level, islets were fixed in 4% paraformaldehyde/0.05% glutaraldehyde (Polysciences) in 100 mM PIPES/0.5 mM MgCl 2, pH 7.2, for 1 h at 4°C. Samples were then embedded in 10% gelatin and infiltrated overnight with 2.3 M sucrose/20% polyvinyl pyrrolidone in PIPES/MgCl 2 at 4°C. Samples were trimmed, frozen in liquid nitrogen, and sectioned with a Leica Ultracut UCT7 cryo-ultramicrotome (Leica Microsystems). Ultrathin sections of 50 nm were blocked with 5%
