From Disposal to Display: Crinophagy as a β‐Cell Source of Neoantigens in Type 1 Diabetes

ABSTRACT

Type 1 diabetes (T1D) arises when autoreactive lymphocytes target pancreatic β‐cells, yet the mechanisms that convert β‐cell self‐proteins into disease‐relevant antigens remain incompletely defined. β‐cells are uniquely positioned to generate neoantigens because they devote extraordinary protein synthesis capacity to insulin production and maintain thousands of dense‐core insulin granules. While a small fraction of granules undergoes glucose‐stimulated exocytosis, excess, aged, or immature granules can be degraded through crinophagy, a lysosomal pathway in which secretory granules fuse with lysosomes to form crinosomes. Recent immunopeptidomic studies suggest that crinosomes are not merely disposal compartments but antigen‐editing organelles that remodel insulin granule cargo into pathogenic epitopes. These include free insulin B‐chain peptides, hybrid insulin peptides, post‐translationally modified insulin and C‐peptide epitopes, and stress‐induced insulin sequence variants such as insulin B‐peptide C19S. Because several crinosome‐associated epitopes are poorly represented in the thymus, crinophagy may create a peripheral antigen repertoire that permits escape from central tolerance and activation of autoreactive T cells in islets and draining lymphoid tissues. This review discusses the physiology of β‐cell granule turnover, the mechanisms of crinosome‐associated neoantigen generation, and their implications for T1D pathogenesis, biomarkers, and therapeutic targeting.

1. Introduction: β‐Cells Are Uniquely Positioned to Generate Neoantigens

Type 1 diabetes (T1D) is a chronic autoimmune disease in which autoreactive lymphocytes, particularly T cells, progressively target pancreatic β‐cells, resulting in loss of insulin production and lifelong dependence on exogenous insulin [1, 2, 3, 4, 5]. This selective tissue destruction occurs in a cell type with an unusually specialized secretory identity. Pancreatic β‐cells produce large amounts of insulin in response to glucose, thereby coordinating systemic glucose uptake and maintaining blood glucose homeostasis. In doing so, β‐cells devote a substantial fraction of their protein synthesis capacity to insulin production and maintain thousands of insulin secretory granules, each densely packed with insulin molecules [6].

This secretory specialization places β‐cells under a unique proteostatic burden. Insulin must be synthesized, folded, processed, condensed into dense‐core granules, stored, released, and degraded with high fidelity. Because β‐cells are exposed to persistent demands for insulin synthesis and secretion, they are especially susceptible to endoplasmic reticulum (ER) stress [7, 8, 9, 10]. This stress may shape antigen processing and presentation by changing the intracellular availability of β‐cell proteins, modifying enzymes, and peptide products [11]. Consistent with this idea, recent work in human islets has shown that ER and inflammatory stress induce strong cell type‐specific transcriptional responses, particularly in β‐, α‐, and ductal cells [12].

A second feature that distinguishes β‐cells is the relationship between secretion and intracellular disposal. Each β‐cell stores insulin within approximately 5000–10,000 or more insulin secretory granules, yet acute glucose stimulation releases only a small fraction of this total pool, roughly 1%–2% of all granules (Figure 1) [13]. The remainder must be stored, mobilized later, or degraded. One pathway for degradation is crinophagy, a lysosomal process found in secretory cells, especially exocrine, endocrine, and neuroendocrine cells. In crinophagy, unreleased secretory vesicles fuse directly with lysosomes, leading to rapid digestion and recycling of vesicle contents (Figure 2A) [14]. In β‐cells, dense‐core granules can fuse with lysosomes to generate crinosomes, compartments that degrade excess or aged insulin granule cargo [15, 16].

FIGURE 1.

Insulin dense‐core granule biogenesis, maturation, storage, and glucose‐stimulated secretion in pancreatic β‐cells. (1) Preproinsulin is synthesized in the endoplasmic reticulum (ER), where the signal peptide is removed and proinsulin undergoes folding and disulfide bond formation. (2) Proinsulin traffics through the Golgi apparatus and is packaged into immature dense‐core granules. (3) During granule maturation, acidification activates processing enzymes that cleave proinsulin into insulin and C‐peptide. Mature insulin molecules condense into zinc‐containing dense cores within secretory granules. (4) Mature granules are stored in intracellular pools, with a subset docked and primed for exocytosis at the plasma membrane and a larger reserve pool distributed throughout the cytoplasm. (5) Acute glucose stimulation triggers rapid exocytosis of docked granules, producing the first phase of insulin secretion. (6) Sustained glucose stimulation recruits reserve granules to the plasma membrane, where they undergo exocytosis and support the second phase of insulin secretion. Abbreviations: ER, endoplasmic reticulum. Figure was prepared in BioRender. Wan, X. (2026), https://BioRender.com/qmap9ee.

FIGURE 2.

Crinophagy as a source of neoantigens in type 1 diabetes. Proposed model linking β‐cell crinophagy to the generation of neoantigens and autoimmune T cell activation in type 1 diabetes (T1D). (A) Under physiological conditions, excess, aged, or immature insulin granules undergo crinophagy through direct fusion with lysosomes, forming crinosomes that degrade and recycle granule contents. (B) During β‐cell stress, including inflammatory cytokine exposure, metabolic stress, endoplasmic reticulum (ER) stress, and oxidative stress, crinophagic pathways may be altered through changes in granule trafficking, lysosomal delivery, peptide modification, or protease activity. These changes may increase degradation of granule proteins and diversify the peptide repertoire generated within β‐cells. (C) Crinosome‐associated processing may generate multiple classes of neoantigens, including free insulin peptides, hybrid insulin peptides (HIPs), post‐translationally modified (PTM) epitopes, sequence variants (e.g., C19S insulin), and defective ribosomal products (DRiPs). (D) Peptides represented during thymic antigen presentation contribute to central tolerance through clonal deletion or regulatory T cell (Treg) differentiation. In contrast, neoepitopes that are absent or poorly represented in the thymus may be presented in peripheral tissues by antigen‐presenting cells (APCs), resulting in activation, proliferation, and effector differentiation of autoreactive T cells that contribute to β‐cell destruction. Solid arrows indicate experimentally supported mechanisms, whereas dashed arrows indicate proposed mechanisms that remain under investigation. Abbreviations: APC, antigen‐presenting cell; DRiP, defective ribosomal insulin gene product; ER, endoplasmic reticulum; HIP, hybrid insulin peptide; IFNγ, interferon‐γ; IL‐1β, interleukin‐1β; MHC‐II, major histocompatibility complex class II; mTEC, medullary thymic epithelial cell; PTM, post‐translational modification; ROS, reactive oxygen species; TCR, T cell receptor; Treg, regulatory T cell; TNFα, tumor necrosis factor‐α; T1D, type 1 diabetes. Figure was prepared in BioRender. Wan, X. (2026), (https://BioRender.com/a6r7seo) is licensed under CC‐BY 4.0.

Recent immunopeptidomic studies have shifted the conceptual role of crinophagy from a purely homeostatic degradation pathway to a potential antigen‐editing compartment [17, 18]. In this model, insulin granule cargo is remodeled within crinosomes into islet‐enriched epitopes relevant to T1D. These epitopes include hybrid insulin peptides (HIPs), post‐translationally modified insulin and C‐peptide epitopes, and insulin peptides carrying single‐residue sequence changes [17, 18, 19, 20, 21, 22]. Several of these epitopes are recognized by autoreactive T cells with pathogenic potential (Figure 2D) [19, 20, 21, 23]. In the NOD mouse model, β‐cell crinophagic granules contain an epitope repertoire that is more diverse than the insulin repertoire detected in thymus, suggesting that T cells recognizing islet‐restricted epitopes may escape thymic deletion [17].

In the context of neoantigen generation in T1D, there has been substantial evidence documenting the crucial role of ER stress and inflammation (i.e., inflammatory cytokines) in promoting neoantigen generation in β‐cells. In line with this, crinophagy appears to be particularly associated with stress‐ and inflammation‐induced neoantigens. Our earlier work showed that inducing ER stress in mouse islets significantly increased the presence of the 6.9HIP, which is formed via the fusion of C‐peptide and islet amyloid polypeptide (IAPP), in the crinophagic granules [24]. This idea has gained further support from the identification of an insulin B‐chain peptide in which cysteine 19 is replaced by serine, referred to as C19S. C19S‐containing insulin peptides were detected in both mice and humans, and were enriched in the crinosome fraction of β‐cells exposed to ER stress and different inflammatory cytokines including TNFα, IFNγ, and IL‐1β. Moreover, C19S‐specific CD4+ T cells expanded at diabetes onset in individuals with T1D and persisted through disease progression [18]. The abundance of this neoepitope increased with inflammation and oxidative stress, linking the islet microenvironment to neoantigen generation [18]. Together, these findings suggest that crinophagy is not only a physiological mechanism for insulin granule quality control, but also a potential source of neoantigens that can promote autoimmune responses in T1D. We propose that crinophagy creates a peripheral β‐cell antigen repertoire that differs from the repertoire represented during thymic selection, thereby linking normal granule turnover to autoimmune recognition.

2. β‐Cell Insulin Dense‐Core Granules and the Physiology of Crinophagy

2.1. Insulin Dense‐Core Granule Biogenesis and Maturation

Insulin granule biogenesis begins with glucose‐regulated transcription and translation of preproinsulin mRNA [25, 26, 27]. Preproinsulin contains an N‐terminal signal peptide that targets the nascent polypeptide to the ER, where it is translocated across the ER membrane and the signal peptide is removed [26, 27, 28, 29]. In the ER, proinsulin folds with assistance from chaperones and forms the disulfide bonds required for native insulin structure [28, 30]. Only properly folded proinsulin exits the ER and traffics through the Golgi apparatus to the trans‐Golgi network, where packaging into immature dense‐core granules begins (Figure 1) [30, 31, 32, 33].

During granule maturation, immature dense‐core granules acidify, which activates proteases such as PC1/3 and PC2, as well as carboxypeptidase E. These enzymes cleave proinsulin into insulin and C‐peptide. Insulin then condenses with zinc and calcium into hexameric structures that form the electron‐dense granule core [6, 30]. Mature granules are stored in functionally distinct pools. A minority of granules are docked at or near the plasma membrane and are poised for rapid exocytosis upon glucose stimulation, whereas most granules remain deeper in the cytoplasm and must be mobilized during sustained secretory demand (Figure 1) [34].

2.2. Dense‐Core Granule Fates: Exocytosis or Crinophagy

Once formed, insulin granules have two broad fates: regulated exocytosis or intracellular degradation. Acute glucose stimulation releases only a portion of docked granules, corresponding to a small fraction of the total granule pool [34]. Continued insulin secretion then depends on recruitment and trafficking of reserve granules to the plasma membrane [30]. Intracellular granules that remain indefinitely must eventually be turned over. Current evidence suggests that aged granules, nascent granules under nutrient or metabolic stress, and immature granules can all enter lysosomal degradation pathways, including crinophagy (Figure 2B) [15, 35, 36, 37, 38, 39].

Granule age is one determinant of degradative fate. Aged insulin granules display reduced glucose‐stimulated, microtubule‐dependent mobility and are disposed of within actin‐positive multigranular bodies, consistent with lysosomal disposal [35]. Conversely, newly formed granules can be diverted to lysosomes under nutrient deprivation or metabolic stress [36, 37]. Starvation‐induced and stress‐induced nascent granule degradation suggest that β‐cells can acutely suppress secretion and reroute newly produced insulin cargo to lysosomes when secretory output would be maladaptive. Immature granules containing proinsulin can also be degraded after SNARE‐facilitated fusion with lysosomes, providing a mechanism to regulate excess proinsulin and maintain intracellular insulin homeostasis [15]. Currently, there is no consensus on whether all of these lysosome‐directed pathways should be classified as crinophagy or whether the term should be reserved for the degradation of mature insulin granules. For the purposes of this review, we consider these processes together because they share the common feature of lysosome‐mediated degradation of insulin granules. An unresolved question is whether crinophagy selectively samples particular granule populations. Current evidence indicates that aged granules, immature granules, and stress‐diverted nascent granules can all enter lysosomal degradation pathways, yet it remains unclear whether these granule populations converge on identical crinosomes or are processed through distinct degradative routes. Granule age, secretory history, oxidative damage, and maturation status may all influence granule selection. If different granule populations contain distinct insulin processing intermediates or undergo different biochemical modifications before lysosomal fusion, selective crinophagy could substantially influence the peptide repertoire available for antigen presentation. At present, however, little is known regarding the molecular determinants that govern granule selection during crinophagy.

Crinophagy differs from macroautophagy [40]. In macroautophagy, a double‐membrane autophagosome forms around cytoplasmic material and then fuses with acidified lysosomes. In crinophagy, dense‐core granules fuse directly with lysosomes to form crinosomes [14]. This distinction matters because crinophagy may be selectively modulated without broadly inhibiting macroautophagy, although available pharmacological tools can affect multiple lysosomal pathways [14, 17]. The extent to which inflammation, ER stress, and metabolic stress change the rate, cargo selectivity, or antigenic output of β‐cell crinophagy remains incompletely understood.

2.3. Crinophagy Regulation

Direct studies of β‐cell crinophagy regulation remain limited. Under homeostatic conditions, crinophagy contributes to the degradation of aged granules and immature granule cargo, thereby helping regulate proinsulin and insulin content [15, 35]. When regulated secretion is impaired, β‐cells can maintain hormone storage levels by increasing granule degradative pathways, including crinophagy and related lysosomal routes [41]. This compensatory response supports the idea that crinophagy is integrated into the broader balance among insulin biosynthesis, storage, secretion, and disposal.

Pharmacological evidence also suggests partial overlap between pathways that regulate crinophagy and other lysosomal processes. Chloroquine, an inhibitor commonly used to interfere with autophagosome‐lysosome fusion and lysosomal function, reduces the number of crinosome‐like vesicles in NOD β‐cells [17]. Steroid hormones may also affect crinophagy: progesterone has been reported to increase crinophagic degradation, whereas corticosterone decreases it in isolated islets [42]. Despite these observations, the molecular rules that determine which granules of different ages or maturation states are selected for crinophagy remain poorly defined. Identifying those rules is central to understanding whether crinophagy samples granule cargo randomly or generates a biased antigenic repertoire.

2.4. Lysosomal Machinery Regulating Crinophagy

Recent studies have begun to identify the molecular machinery that mediates crinophagy in pancreatic β‐cells. Unlike macroautophagy, which requires de novo autophagosome formation, crinophagy proceeds through direct fusion of dense‐core granules with lysosomes [43]. This process involves a specialized SNARE complex consisting of VAMP4 on insulin granules and the lysosomal/endosomal SNAREs syntaxin‐7 (STX7), syntaxin‐8 (STX8), and VTI1B [15]. Genetic disruption of VAMP4 impairs granule delivery to lysosomes, increases intracellular insulin content, and disrupts normal insulin homeostasis, establishing direct membrane fusion as a central mechanism of β‐cell crinophagy [15]. In parallel, lysosomal trafficking proteins also contribute to granule turnover. Rab7‐interacting lysosomal protein (RILP), acting downstream of Rab7 and interacting with the granule‐associated GTPase Rab26, promotes lysosomal degradation of proinsulin‐containing granules and limits insulin secretion [44]. Together, these studies suggest that granule trafficking, membrane fusion, and lysosomal maturation represent coordinated regulatory checkpoints governing crinophagy.

Although these molecular components have now been identified, how they are regulated during inflammation, oxidative stress, or autoimmune diabetes remains largely unknown. Whether inflammatory cytokines alter SNARE complex assembly, Rab GTPase activity, lysosomal positioning, or granule selection has not been systematically investigated. Defining these regulatory mechanisms will be important for understanding whether β‐cell stress alters only the quantity of crinophagy or fundamentally changes the antigenic products generated within crinosomes.

3. Autoantigens, Thymic Tolerance, and the Emergence of Islet‐Restricted Epitopes

3.1. Classical Native β‐Cell Autoantigens

The canonical β‐cell autoantigens in T1D include insulin/proinsulin, GAD65, IA‐2/ICA512, and ZnT8, with additional targets such as chromogranin A, IAPP, and C‐peptide identified through T cell and immunopeptidomic studies [11, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56]. Among these antigens, insulin/proinsulin is especially central. Insulin B‐chain epitopes are disease‐driving in NOD mice, and insulin remains one of the earliest and most β‐cell‐specific targets of autoimmunity [45, 57, 58, 59].

A notable feature of this antigen set is its connection to β‐cell secretory biology. Insulin, C‐peptide, IAPP, chromogranin A, IA‐2/ICA512, phogrin/IA‐2β, and ZnT8 are all linked to dense‐core granules or to their maturation, storage, and regulated secretion [6, 54]. This localization suggests that granule biogenesis, turnover, and lysosomal degradation are not merely housekeeping processes; they may also shape which β‐cell peptides become immunologically visible [11].

However, defining an autoantigen by its source protein is insufficient. Central tolerance and peripheral activation are determined by the specific peptide–MHC complexes that are generated and presented [60, 61]. Thus, the key issue is not only whether insulin or another β‐cell protein is expressed in the thymus or islet, but whether the relevant disease‐associated peptide is generated in the same compartment and context [17, 19].

3.2. Central Tolerance and Gaps in Thymic Antigen Representation

Thymic insulin expression provides a paradigm for central tolerance to β‐cell antigens [62, 63, 64]. However, thymic expression of a source protein does not ensure thymic presentation of every disease‐relevant peptide [65]. This limitation is especially important when epitopes depend on β‐cell‐specific granule processing, stress chemistry, post‐translational modification, peptide fusion, or crinosome formation. Insulin is expressed in a subset of medullary thymic epithelial cells, and human thymic insulin expression is influenced by INS VNTR alleles associated with T1D susceptibility [61, 66, 67].

The distinction between protein expression and peptide presentation is illustrated by insulin B‐chain epitopes. Presentation of intact insulin by MHC class II in the NOD thymus preferentially generates the insulin B‐chain 13‐21 epitope (B:13‐21), whereas APCs exposed to extracellular free insulin peptides, an event that takes place in pancreatic islets, can not only present B:13‐21 but also the insulin B‐chain 12‐20 epitope (B:12‐20), which only differs from B:13‐21 by one amino acid shift [58, 68]. This difference shows that distinct antigens and processing pathways can generate distinct peptide–MHC complexes from the same source protein. The C‐peptide epitope CP2 encoded by Ins2 is also spontaneously presented in NOD medullary thymic epithelial cells, but presentation of the epitope CP1 encoded by Ins1 was similar to background levels, further illustrating that thymic presentation includes some, but not all, insulin‐derived peptides [17].

3.3. Unique Neoepitopes Identified in Islet β‐Cells

The islet antigen repertoire contains canonical insulin peptides that are also represented in the thymus [11]. However, islet β‐cells and islet APCs can display peptides that are absent or poorly represented in the thymic repertoire [17]. These include processing‐dependent insulin B‐chain variants, proinsulin peptides fused to other β‐cell proteins, post‐translationally modified or chemically remodeled insulin epitopes, and crinosome‐associated peptides (Figure 2C) [19]. The mechanisms that generate these neoepitopes are discussed in detail in Section 4.

3.4. Implications for Escape From Central Tolerance

The emergence of islet‐restricted epitopes provides a plausible explanation for how autoreactive T cells can evade central tolerance despite thymic expression of insulin. While antigen expression in the thymus shapes tolerance to insulin, physiological mechanisms in the pancreatic islet can generate epitopes that are absent, rare, or differently processed in the thymus [17, 66, 67]. As a result, neoepitopes encountered primarily in the pancreas, pancreatic lymph nodes, or peripheral blood after β‐cell degranulation may activate T cells that escaped thymic deletion or Treg selection. These cells can then expand under inflammatory conditions and contribute to β‐cell destruction [18, 21, 69, 70]. In this sense, crinophagy may transform insulin from a thymically expressed self‐protein into a peripheral autoantigen by generating peptides that were not efficiently sampled during central tolerance [17].

4. Crinosomes as β‐Cell Antigen‐Editing Compartments

4.1. Lysosomal Processing of Insulin Granule Cargo

Lysosomes are acidified vesicles containing hydrolases, including cathepsins, that degrade vesicular and endocytic cargo [14]. Upon fusion between the insulin granule and lysosomes, the granule cargo is exposed to these lysosomal enzymes and is rapidly degraded and recycled (Figure 2A). Crinosome‐dependent antigen generation is chloroquine sensitive. Short‐term chloroquine treatment reduces crinosome‐like vesicles by approximately half in NOD mouse β‐cells and reduces presentation of the insulin epitopes B:12–20 and B:13–21 from the crinosome‐enriched fraction [17]. These findings suggest that the lysosomal processing of insulin granule cargo contributes to the generation of disease‐relevant β‐cell antigens [71]. Importantly, the antigenic products of crinophagy are not expected to be identical to intact insulin or to peptides generated by conventional APC processing of insulin protein [72]. Instead, crinosomes may provide an acidic, enzyme‐rich environment in which insulin granule cargo is fragmented, modified, or recombined.

4.2. Catabolized Insulin‐Derived Epitopes in Crinosomes

The insulin B‐chain 9‐23 peptide was one of the earliest and most intensively studied diabetogenic insulin epitopes in NOD mice [58]. Subsequent work revealed that T cells can recognize two overlapping peptide registers within this region: the 12‐20 register, referred to as B:12‐20, and the 13‐21 register, referred to as B:13‐21 [58, 68, 73, 74]. In type A antigen presentation, insulin‐reactive T cells recognize B:13‐21 generated when APCs process intact insulin protein, whereas in type B antigen presentation, T cells recognize B:12‐20 derived from free or denatured insulin peptides [58].

This distinction has major implications for tolerance. In type A antigen presentation, T cells recognize insulin processed through conventional APC pathways, in which internalized insulin is partially degraded and loaded onto MHC class II molecules within late endosomal compartments [68, 75]. In type B antigen presentation, T cells instead recognize exogenous insulin peptides that can bind MHC class II molecules in contexts not reproduced by intact insulin processing [68]. Because intact insulin is presented in the thymus, whereas free peptides may be enriched in islets, T cells activated by type B presentation can escape thymic deletion and become activated in the target tissue [68, 76].

Crinosomes provide a physical compartment that may help explain the source of such free insulin peptides. Recent work found that B:12‐20 is enriched in crinosomes, whereas B:13‐21 is enriched in conventional dense‐core granules [17]. Moreover, medullary thymic epithelial cells and thymic APCs preferentially present B:13‐21, while dispersed islet cells and islet APCs present both B:13‐21 and B:12‐20 [17]. These observations support a model in which B:12‐20 presentation depends on peripheral or islet peptide availability rather than canonical thymic insulin processing (Table 1).

TABLE 1.

β‐cell neoantigens potentially linked to granule turnover and crinosome‐associated processing in T1D.

Neoantigen/epitope	Source protein(s)	Proposed mechanism of generation	Evidence linking to crinophagy a	T‐cell recognition	Human evidence	References
Insulin B‐chain 12‐20 (B:12‐20)	Insulin	Generated from granule/crinosome‐associated insulin processing	Enriched in crinosome fractions; reduced by chloroquine; preferentially represented in islets versus thymus	CD4+ T cells	Indirect; related insulin peptide–MHC‐II complexes can be detected outside islets	[17, 57, 68, 71, 74, 76, 77, 78]
Insulin B‐chain 13‐21 (B:13‐21)	Insulin	Conventional processing of intact insulin protein in APCs	Enriched in dense‐core granules rather than crinosomes; preferentially represented in thymic processing	CD4+ T cells	Conserved insulin epitope in mice and humans	[17, 57, 68, 73, 75]
Hybrid insulin peptides (HIPs)	Insulin/proinsulin fused with IAPP, chromogranin A, or other granule peptides	Peptide fusion through proteolytic transpeptidation or spontaneous peptide ligation	Chloroquine reduces HIP abundance in crinosome‐enriched fractions	CD4+ T cells	HIP‐reactive T cells detected in individuals with T1D	[17, 20, 21, 69, 79, 80, 81, 82]
C19S insulin neoepitope	Insulin B‐chain	Stress‐associated cysteine‐to‐serine transformation in insulin B:9–23 region	Enriched in crinosome fractions	CD4+ T‐cell	Identified in mouse and human islets and detected in the peripheral human MHC‐II peptidome after metabolic stimulation	[18, 83]
Oxidatively modified insulin peptides	Insulin	Oxidative post‐translational modification under inflammatory or oxidative stress	Direct crinosome linkage is not established	CD4+ and CD8+ T cells	Autoantibody and T‐cell responses from people with T1D	[84, 85, 86]
Deamidated insulin peptides	Insulin and other secretory proteins	Tissue transglutaminase‐mediated deamidation and stress‐induced modifying enzyme activity	Direct crinosome linkage is not established	CD4+ T cells	Human T‐cell recognition reported. But glutamine deamidation did not enhance responses in one study	[84, 87, 88, 89, 90, 91, 92]
Insulin B:9‐23–C‐peptide fusion peptide	Proinsulin	Likely through transpeptidation in lysosomal compartments	Direct crinosome linkage is not established	CD4+ T‐cell	Human T‐cell relevance reported	[93]
Insulin defective ribosomal products (DRiPs) a	Alternative insulin translation product	Alternative translation of insulin from second ORF	Direct crinosome linkage is not established	CD8+ T cells	Human T1D relevance reported	[94, 95]

^a Evidence for direct crinosome involvement varies by epitope class; several entries are included as granule‐associated or stress‐associated comparator neoantigens rather than confirmed crinosome products.

4.3. Crinophagy May Favor the Generation of Hybrid Insulin Peptides

Crinosomes contain an acidic, protease‐rich environment that differs fundamentally from both secretory granules and the endoplasmic reticulum. Lysosomal acidification activates multiple cathepsins and hydrolases capable of extensive peptide cleavage, trimming, and remodeling. Similar lysosomal environments support antigen processing of professional APCs, where peptide generation is strongly influenced by local pH, protease abundance, and peptide residence time. Although β‐cell crinosomes have not yet been characterized to the same extent, they potentially represent a comparable biochemical environment capable of reshaping peptide repertoires rather than simply degrading proteins to amino acids. Future proteomic characterization of crinosome‐associated proteases and peptide intermediates will be important for understanding how this compartment contributes to neoantigen diversity.

Hybrid insulin peptides (HIPs) form when insulin or proinsulin fragments covalently fuse with other β‐cell granule peptides, such as IAPP or chromogranin A [20, 51]. HIPs provide a striking example of how the dense‐core granule pathway can generate antigenic structures not encoded as contiguous sequences in the genome. One mechanism involves cathepsin D, an aspartic protease that can drive HIP formation through proteolytic transpeptidation [79]. HIPs may also form spontaneously when peptides containing an aspartic anhydride intermediate react with the N‐termini of other β‐cell peptides [80]. The acidic environment of β‐cell dense‐core granules or crinosomes may favor such reactions [17].

Crinosomes may therefore facilitate HIP formation because they combine high peptide concentration, acidity, and proteolytic enzyme activity [17]. Supporting this idea, chloroquine treatment decreases overall HIP abundance in crinosome‐enriched fractions [17]. Functionally, HIPs are highly relevant because they can act as strong agonists for pathogenic CD4+ T cell clones in T1D [20, 21, 81, 82]. Together, these findings suggest that peptide fusion may connect β‐cell secretory granule biology to the generation of autoreactive T‐cell targets, while leaving open which granule or lysosomal compartments are most important in vivo.

4.4. Post‐Translationally Modified C‐Peptide and Insulin Epitopes

Post‐translationally modified neoepitopes may arise when β‐cell stress changes local enzymatic activity, redox chemistry, peptide charge, MHC binding, or TCR recognition. However, the evidence varies by antigen and modification, and some proposed modifications remain debated [84, 85, 86, 87, 88, 89, 90, 91, 96].

Tissue transglutaminase can generate negatively charged peptides through deamidation, improving binding of some islet autoantigen peptides to human HLA‐DQ molecules [87, 92]. ER stress in human β‐cells can increase the activity of modifying enzymes, including tissue transglutaminase, and generate modified epitopes recognized by CD4+ T cells from patients with T1D [84]. Oxidative chemistry can also create post‐translationally modified insulin peptides capable of stimulating CD4+ and CD8+ T cells from people with T1D [85, 86]. At the same time, not every modification increases immunogenicity. C‐peptide itself is an autoantigen in humans: CD4+ T cells from more than 60% of individuals with recent‐onset T1D recognize C‐peptide epitopes [88]. However, one study found that glutamine deamidation of C‐peptide did not enhance CD4+ T‐cell responses [89]. Therefore, PTM neoepitopes should be discussed as a heterogeneous category in which each modification must be evaluated experimentally.

The relationship between PTM generation and crinophagy remains unclear. ER stress and oxidative stress clearly influence β‐cell antigenicity, but their effects on crinophagy rate, crinosome composition, and PTM formation within crinosomes are not fully defined. Because crinosomes contain acidic hydrolases and degraded insulin products, they are plausible sites for antigen remodeling; however, direct evidence connecting specific PTMs to crinosome processing is still limited.

4.5. Sequence Variants and Noncanonical Insulin Products

Noncanonical insulin products arise through mechanisms outside standard proinsulin processing, including alternative translation, defective ribosomal products, altered peptide boundaries, and stress‐driven single‐residue transformations [18, 93, 94, 95]. Human insulin mRNA contains an alternative open reading frame encoding a defective ribosomal insulin gene product, or insulin‐DRiP, that is highly immunogenic and targeted by T cells in T1D [94]. Inflammation during insulitis can affect translation fidelity, and IFNγ increases recognition of insulin‐DRiP‐derived antigen by pathogenic CD8+ T cells [95].

Another example of altered insulin processing is an insulin B‐chain 9‐23 peptide extended at the C‐terminus by fusion with the C‐peptide of proinsulin. This chimeric peptide can induce CD4+ T cell responses as strong as, or stronger than, responses induced by mutated insulin peptides and is thought to arise from transpeptidation in lysosomal compartments [93].

The recently described C19S insulin neoepitope represents a distinct mechanism. C19S involves replacement of cysteine 19 in the insulin B‐chain 9‐23 region with serine, without a corresponding mutation in the insulin gene [18]. Its abundance increases with oxidative stress and inflammation, suggesting that stressed islets generate qualitatively different insulin antigens. This microenvironment‐driven single‐residue transformation is conserved in mice and humans and is enriched in crinosomes. Despite weaker binding to mouse MHC class II, the C19S neoepitope induces a stronger T‐cell response than the native B‐chain 9‐23 epitope [18]. In contrast, a recent study utilizing molecular dynamics simulation suggests the C19S neoepitope binds to human HLA‐DQ8 with stronger affinity than the native epitope [83]. It remains unclear whether inflammation and oxidative stress transform insulin peptides before crinosome entry, within crinosomes, or in APCs that acquire insulin‐derived material.

4.6. Crinophagy Positioned to Generate Neoantigens

Taken together, these findings suggest that crinophagy is uniquely positioned to generate neoantigens because it concentrates several antigen‐remodeling mechanisms within a single intracellular compartment. Unlike conventional antigen processing pathways, crinosomes combine high concentrations of insulin granule cargo with an acidic, protease‐rich environment capable of extensive peptide cleavage and remodeling. This environment can generate free insulin peptides such as B:12‐20, promote peptide fusion events that give rise to HIPs, and potentially facilitate post‐translational modifications or stress‐associated chemical transformations that alter antigenicity. Crinosomes may also preferentially degrade aged, immature, or stress‐diverted granules, thereby enriching for proteins and peptides that have experienced altered processing or cellular stress. As a result, crinophagy is not simply a degradative pathway but a potential antigen‐editing compartment capable of producing peptide repertoires that differ from those generated by conventional APC processing or represented during thymic selection. By creating novel epitopes from insulin, crinophagy provides a plausible mechanistic link between normal β‐cell granule turnover and the emergence of autoreactive T cell targets in T1D.

However, current evidence does not establish whether crinophagy is a primary driver of disease initiation, an amplifier of ongoing autoimmunity, or a physiological pathway that becomes immunologically consequential only in the context of genetic susceptibility and islet inflammation. Because crinophagy contributes to normal insulin granule quality control under homeostatic conditions, it is unlikely to be intrinsically pathogenic. Instead, disease relevance may emerge when inflammatory, oxidative, or metabolic stress alters the quantity or quality of epitopes generated within crinosomes, thereby reshaping the β‐cell antigen repertoire encountered by the immune system. In this framework, crinophagy may function less as a trigger of autoimmunity and more as a mechanism through which β‐cell physiology influences antigen availability and immune recognition.

Collectively, the current literature suggests that crinophagy should not be viewed as an isolated mechanism responsible for every β‐cell neoantigen. Rather, crinosomes likely represent one intracellular environment in which multiple antigen‐generating mechanisms converge. Peptide cleavage, peptide fusion, post‐translational modification, oxidative chemistry, and stress‐associated peptide remodeling may occur before granule entry, within crinosomes, or following uptake by professional antigen‐presenting cells. The relative contribution of each pathway probably differs among individual epitopes. Nevertheless, by concentrating dense‐core granule cargo within an acidic lysosomal compartment enriched for proteolytic activity, crinophagy provides a unique environment capable of diversifying the β‐cell peptide repertoire available for immune recognition.

5. T Cell Responses to Crinophagy‐Associated Epitopes

T cell fate depends not only on the source protein but also on where and how the peptide–MHC complex is generated and presented [97]. Shared thymic and islet epitopes may support deletion or regulatory diversion, whereas islet‐restricted epitopes generated by granule processing, crinosomes, or stress chemistry may preferentially activate autoreactive CD4+ T cells that escaped central tolerance (Figure 2D) [17, 18, 68, 76, 98]. This distinction helps connect the biochemical origin of antigenic peptides to their immunological consequences.

In contrast to thymically represented insulin epitopes, islet‐restricted neoepitopes—including HIPs, crinosome‐associated insulin peptides, and C19S insulin are associated with activated, effector, or memory CD4+ T cell phenotypes rather than regulatory diversion [17, 18, 21]. HIP‐reactive T cells can adopt pathogenic effector fates, and C19S‐specific CD4+ T cells expand at diabetes onset and persist with memory features through disease progression [18, 21]. These findings suggest that neoepitope identity can influence not only antigen specificity but also the quality of the T cell response.

β‐cell antigens also need not remain confined to the islet. Granule‐derived peptides and proteins can be captured by islet APCs, transported to pancreatic lymph nodes, or released into circulation, where they are presented by APCs and support priming or expansion of cognate CD4+ T cells [19, 21, 76, 77, 78]. Glucose‐stimulated β‐cell degranulation releases insulin peptide fragments into circulation, and blood leukocytes can recapitulate diabetogenic peptide–MHC‐II complexes displayed in pancreatic islets [77, 78]. Thus, crinophagy‐associated epitopes may influence both local islet inflammation and systemic antigen recognition.

Evidence supporting the relevance of crinophagy‐associated neoantigens is increasingly emerging from human studies (Table 1). Immunopeptidomic analyses of human islets have identified naturally presented insulin‐derived peptides, HIPs, and other noncanonical β‐cell antigens that are recognized by autoreactive T cells, demonstrating that these epitopes are generated in vivo rather than representing artifacts of experimental systems [19, 22, 99, 100, 101, 102]. Human T cell responses against HIPs have been detected in individuals with T1D, and HIP‐reactive CD4+ T cells exhibit pathogenic effector phenotypes associated with disease activity [20, 21]. Post‐translationally modified insulin and C‐peptide epitopes are likewise recognized by T cells and autoantibodies from patients with T1D, supporting a role for stress‐associated antigen remodeling in human disease [84, 85, 88, 90]. Most notably, the recently described C19S insulin neoepitope was identified in both mouse and human islets, and human C19S‐specific CD4+ T cells were found to expand at diabetes onset and persist with memory characteristics throughout disease progression [18]. Autoreactive T cells recognizing insulin, HIPs, and other β‐cell antigens are readily detectable in peripheral blood from individuals with T1D, indicating that disease‐relevant antigen recognition extends beyond the pancreas [103, 104]. Indeed, recent work demonstrated that the C19S insulin neoepitope can be detected in the peripheral human MHC‐II peptidome following metabolic stimulation [18]. Although direct evidence linking each neoepitope class to crinosome processing in human β‐cells remains incomplete, the convergence of human immunopeptidomic, T cell, and biomarker studies supports the broader concept that β‐cell intrinsic antigen remodeling generates clinically relevant autoimmune targets.

6. Open Questions and Future Directions

Despite progress in defining crinophagy‐associated neoantigens, several fundamental questions remain unresolved. These questions extend beyond the identification of individual epitopes and instead concern the broader cell biology of crinophagy, the biochemical mechanisms of antigen remodeling, and the immunological pathways linking β‐cell antigen generation to autoimmune recognition. Addressing these questions will require integrating advances in β‐cell biology, lysosome biology, immunopeptidomics, and antigen presentation.

First, which forms of β‐cell stress favor crinophagy‐derived neoantigen production in vivo? Inflammatory, oxidative, ER, metabolic, and aging‐associated stresses may each alter granule turnover, lysosomal activity, or antigen chemistry. C19S formation demonstrates that inflammation and oxidative stress can enhance insulin neoantigen abundance, but it remains unclear whether these stresses increase crinosome formation, change crinosome cargo, modify insulin before lysosomal entry, or alter processing inside APCs [18].

Second, how are crinosome‐derived peptides presented to T cells? Several nonexclusive routes are possible. Peptides could be loaded directly onto β‐cell MHC class II under inflammatory conditions, released after β‐cell degranulation, transferred from β‐cells to islet APCs, or regenerated inside inflammatory APCs after uptake of insulin granule material [105, 106]. Each route would have different implications for tolerance, tissue specificity, and therapeutic intervention. Defining the relevant pathway for each epitope will require compartment‐specific antigen tracing, immunopeptidomics, and cell‐specific manipulation of antigen‐processing pathways.

Third, what distinguishes tolerogenic from pathogenic antigen repertoires? Some insulin‐derived peptide–MHC complexes are represented in the thymus and may promote deletion or Treg selection, whereas other peptides appear preferentially in islets or inflammatory peripheral compartments. The determinants of these different outcomes likely include peptide abundance, MHC‐binding register, peptide stability, APC identity, co‐stimulation, tissue inflammation, and timing during disease progression [107]. Understanding these variables will be essential for designing antigen‐specific tolerogenic therapies that do not inadvertently amplify pathogenic responses.

Fourth, can crinosome processing be therapeutically targeted without impairing β‐cell granule homeostasis? Crinophagy has a physiological role in maintaining insulin granule quality and quantity, so broad inhibition could impair β‐cell function. A more attractive strategy may be to target disease‐associated features of crinosome processing, such as stress‐induced peptide chemistry, specific proteases, peptide export, or APC uptake, while preserving basal granule turnover. The ability of chloroquine to reduce crinosome‐like vesicles and diminish crinosome‐associated epitope presentation suggests proof of principle, but more selective approaches will be needed [17].

Finally, crinosome‐derived neoantigens may serve as biomarkers of disease stage or therapeutic response [100, 108, 109, 110, 111, 112]. Studies in NOD mice have shown that β‐cell‐derived peptides and peptide–MHC‐II complexes can disseminate beyond the islet and become detectable in peripheral immune compartments, raising the possibility that disease‐relevant antigen repertoires may be monitored outside the pancreas [77, 78]. Meanwhile, neoepitope‐specific T cells, including C19S‐reactive CD4+ T cells, can persist during disease progression in humans [18]. If crinosome‐associated epitopes reflect β‐cell stress or inflammatory state, they could help identify disease activity earlier than conventional measures. Recent work demonstrates that β‐cell‐derived neoantigens can be detected in the peripheral human MHC‐II peptidome following metabolic stimulation, suggesting that circulating immune cells may provide a window into ongoing β‐cell antigen release [18, 113, 114]. However, one caveat to peripheral detection is that islet‐reactive CD8+ T cells can be enriched in the pancreas of individuals with T1D without being increased in blood, indicating that circulating T cell frequencies may fail to capture tissue‐enriched autoreactive responses [115]. An important next step will be to determine whether comprehensive immunopeptidomic profiling of disease‐associated HLA molecules, particularly HLA‐DQ8 and HLA‐DQ2, can be used to monitor the evolving β‐cell antigen repertoire during disease progression or therapeutic intervention. Future work should determine whether these epitopes and their cognate T cells predict progression, remission, or response to immunotherapy.

7. Concluding Remarks: Crinophagy as a Bridge Between β‐Cell Physiology and T1D Autoimmunity

Crinophagy provides a conceptual bridge between normal β‐cell physiology and autoimmune pathogenesis (Figure 2). At the physiological level, crinophagy helps maintain insulin granule quality and quantity by degrading aged, excess, immature, or stress‐diverted secretory granules. This function is essential for a cell that produces enormous amounts of insulin but releases only a small fraction of its stored granule pool during acute stimulation. Without intracellular recycling pathways, β‐cells would be unable to match insulin biosynthesis and storage to changing metabolic demand.

At the immunological level, the same recycling pathway creates a β‐cell‐specific peptide‐generating compartment. By fusing dense‐core granules with lysosomes, crinophagy exposes insulin, C‐peptide, IAPP, chromogranin A, and other granule‐associated proteins to acidic hydrolases and potentially reactive chemical conditions. The resulting crinosomes can contain degraded insulin products, free peptide fragments, modified peptides, and peptide‐fusion products. These antigenic products differ from intact insulin and may also differ from peptides generated by thymic or conventional APC processing.

At the pathological level, crinophagy may generate neoepitopes that are poorly represented during central tolerance. This peripheral‐thymic mismatch offers a mechanism by which insulin, despite being a thymically expressed self‐protein, becomes a dominant autoantigen in T1D. Crinosome‐associated B:12‐20, HIPs, PTM‐bearing insulin or C‐peptide epitopes, and C19S insulin illustrate how local β‐cell biology can create peptide–MHC complexes capable of activating T cells that escape thymic deletion or regulatory diversion [5, 17, 18, 20, 21, 68].

This model reframes β‐cells as active participants in the generation of their own autoimmune targets. The β‐cell is not simply a passive victim of immune attack; its secretory burden, granule lifecycle, stress responses, and lysosomal degradation pathways help determine which antigens are displayed to the immune system. This perspective does not diminish the importance of genetic susceptibility, MHC‐II binding, APC function, or T cell activation [116]. Rather, it integrates these immune mechanisms with β‐cell intrinsic physiology.

Future studies should define how crinosome formation is regulated, which granule populations enter this pathway, how inflammatory and oxidative stress alter crinosome chemistry, and how crinosome‐derived peptides reach APCs and secondary lymphoid organs. Such work may reveal new biomarkers of β‐cell stress and new therapeutic strategies that reduce pathogenic antigen generation without compromising insulin homeostasis. Ultimately, crinophagy may represent the point at which β‐cell physiology intersects with immune recognition, transforming normal secretory granule turnover into a source of autoimmune neoantigens.

Funding

This work was supported by Juvenile Diabetes Research Foundation United States of America (5‐CDA‐2022‐1175‐A‐N). National Institute of Diabetes and Digestive and Kidney Diseases (R01DK134437). National Institute of Allergy and Infectious Diseases (R01AI162591).

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.