Type 1 diabetes (T1D) is an autoimmune disease in which pathogenic lymphocytes target autoantigens expressed in pancreatic islets, leading to the destruction of insulin-producing β-cells. Zinc transporter 8 (ZnT8) is a major autoantigen abundantly present on the β-cell surface. This unique molecular target offers the potential to shield β-cells against autoimmune attacks in T1D. Our previous work showed that a monoclonal antibody (mAb43) against cell-surface ZnT8 could home in on pancreatic islets and prevent autoantibodies from recognizing β-cells. This study demonstrates that mAb43 binds to exocytotic sites on the β-cell surface, masking the antigenic exposure of ZnT8 and insulin after glucose-stimulated insulin secretion. In vivo administration of mAb43 to NOD mice selectively increased the proportion of regulatory T cells in the islet, resulting in complete and sustained protection against T1D onset as well as reversal of new-onset diabetes. The mAb43-induced self-tolerance was reversible after treatment cessation, and no adverse effects were exhibited during long-term monitoring. Our findings suggest that mAb43 masking of the antigenic exposure of β-cells suppresses the immunological cascade from B-cell antigen presentation to T cell–mediated β-cell destruction, providing a novel islet-targeted and antigen-specific immunotherapy to prevent and reverse clinical T1D.
mAb43, a zinc transporter 8 (ZnT8)–specific monoclonal antibody, provides lasting protection against autoimmune diabetes in NOD mice.
The in vivo islet specificity of mAb43 enables targeted therapy to enhance safety.
High-affinity binding of mAb43 to the extracellular surface of ZnT8 shields β-cells from antigenic exposure.
β-Cell masking results in a localized increase in regulatory T cells in the pancreatic islet.
Prolonged mAb43 treatment clears destructive insulitis, preserves β-cell mass, and reverses seroconversion of insulin autoantibodies in NOD mice.
The immunological processes bridging the masking of specific cell-surface autoantigens and the broad suppression of polyspecific T-cell autoimmunity are still unclear.
Introduction
During the preclinical phase of type 1 diabetes (T1D), four dominant autoantigens (insulin, GAD65, IA-2, and zinc transporter 8 [ZnT8]) are targeted by pathogenic lymphocytes (1). Cell-surface autoantigens and MHC-restricted T-cell epitopes are the first molecular markers that autoreactive lymphocytes encounter when recognizing β-cells. Because of unsuccessful attempts to immunolabel live human islet cells with autoreactive human sera in the past (2), cell-surface autoantigens have not been considered viable targets for immunotherapy. However, these early methods were designed to immunolabel static cell-surface autoantigens, whereas ZnT8, a dynamic autoantigen linked to the recycling of insulin secretory granules, can only be immunolabeled on the β-cell surface after glucose stimulation and endocytosis arrest (3,4). Our recent studies have detected ZnT8 autoantibodies targeting extracellular epitopes in patients with new-onset T1D (5–7), indicating that ZnT8 is a unique cell-surface marker for β-cell recognition by antigen-specific autoreactive B cells.
B cells are efficient antigen-presenting cells that rely on antigen uptake through the B-cell receptor (BCR) (8). ZnT8 has a Y-shaped dimeric structure similar to an immunoglobin (9), making it naturally prone to multivalent BCR ligation for ZnT8 uptake and processing at the B-cell synapse (10). B-cell antigen presentation is a crucial T1D-precipitating event in vivo; it involves cross-presentation of autoantigens to cognate CD8+ T cells (11) and promotes intraislet CD8+ cytotoxic T-cell survival (12). Masking ZnT8 from BCR recognition could directly block ZnT8 uptake by B cells. Furthermore, insulin and other secretory granule-derived autoantigens may be colocalized with ZnT8 at specific sites of exocytosis on the β-cell surface (13). They could be cocaptured with ZnT8 by autoreactive B cells as bystander antigens to induce polyspecific T-cell responses (14). Thus, binding of antibody and ZnT8 at the β-cell surface may mask multiple β-cell autoantigens, leading to a broader suppression of autoreactive B cells in the early stages of the immunological cascade before it gains enough momentum to precipitate cytotoxic T cells as the final mediator of β-cell destruction (15).
In the current study, we used a cell-surface ZnT8 antibody (mAb43) to mask β-cells in NOD mice, a model of spontaneous autoimmune diabetes. mAb43 is a ZnT8-boosted autoantibody that possesses a significantly stronger binding affinity to ZnT8 than natural ZnT8 autoantibodies by several orders of magnitude (4). The binding of mAb43 can effectively mask the entire extracellular surface of ZnT8, competitively obstructing the recognition of β-cells by polyclonal autoantibodies from patients with T1D and corresponding BCRs (4). Moreover, ZnT8 is a tissue-specific autoantigen expressed solely in the pancreatic islet (16,17). Systemically administered mAb43 results in specific mAb43 accumulation within pancreatic islets and persistent ZnT8 binding in vivo (4). However, the exact in vivo consequences of β-cell masking by mAb43 are still unknown. The current study evaluated the long-term efficacy and safety of repeated mAb43 dosing in NOD mice.
Research Design and Methods
Animals
NOD/ShiLtJ (NOD), C57BL/6J (B6), and B6.BKS(D)-Leprdb/J (B6 db) mice were purchased from The Jackson Laboratory. A total of 201 NOD female mice were randomly distributed into various treatment cohorts. A small number of NOD mice were bred in house as detailed in Supplementary Fig. 1. All mouse procedures were approved by the Institutional Animal Care and Use Committee for animal welfare of Johns Hopkins University.
Cell Line
Human EndoC-β1H cells were cultured as described previously (18).
Antibody Production
Mouse mAb43 from a hybridoma clone was produced in an animal origin–free and serum-free medium and purified by size-exclusion high-performance liquid chromatography in PBS under sterile conditions. An isotype control antibody (mouse immunoglobulin G2b [IgG2b]) was purchased from Bioxcell and repurified as above. mAb43-mScalet was produced as described previously (4).
Immunofluorescence Labeling, Imaging, and Data Analysis
Live dispersed human pancreatic islet cells from a 42-year-old female donor (RRID: SAMN34130383) were stimulated with 20 mmol/L glucose at 37°C, chilled on ice, and then immunolabeled with mAb43 (1:50 = 20 μg/mL) and an Alexa 647–labeled secondary antibody (1:200), counterstained with DAPI, and imaged on a confocal microscope. To image the colocalization of ZnT8 and insulin on the β-cell surface, live EndoC-β1H cells in a monolayer cell culture were coimmunolabeled using an anti-insulin antibody (1:50) with mAb43 (1:50) or anti-CD71 (1:50), otherwise using an identical protocol as above. To examine the effect of mAb43 blockade, mAb43 (1:50) was added to the cell culture medium during glucose stimulation. Immunofluorescence was quantified using ImageJ. Immunofluorescence labeling of CD4+FoxP3+CD25+ regulatory T cells (Tregs) was performed on paraformaldehyde-fixed frozen sections of NOD pancreata.
mAb43 Uptake in Human Islet Cells
Human islets, isolated from nondiabetic cadaveric pancreata by Prodo Laboratories, were exposed to the mScarlet-tagged monoclonal antibody (mAb) in the islet culture medium up to 60 h and monitored over time for intraislet mScarlet fluorescence. Next, the islets were paraformaldehyde fixed, and the intracellularly trapped mAb was coimmunolabeled with insulin and glucagon for individual islet cell imaging in intact islets.
mAb43 Administration
mAb43 at 5 mg/kg, isotype-matched mouse IgG2b at 5 mg/kg, or PBS was administered weekly to female NOD mice by intravenous injections. The dosing schemes and numbers of mice in various treatment groups are detailed in Results. For new onset diabetic mice, mAb43 or isotype control was adminstered daily for 5 days followed by weekly dosing as described above.
Diabetes Monitoring
Mice were monitored weekly for 16-h fasting blood glucose (FBG) levels. Diabetes onset was defined by two consecutive FBG readings >250 mg/dL.
β-Cell Mass Measurements
Pancreata were paraformaldehyde fixed and paraffin embedded. Five pancreatic sections of 5-μm thickness across the mouse pancreas were immunohistochemically stained for insulin with an eosin counterstain. The percentage of the pancreas that was positive for insulin staining was measured on digital pathology images and then multiplied by pancreas weight to assess β-cell mass.
Insulitis Scoring
Histological assessment of insulitis was performed on hematoxylin-eosin–stained pancreatic sections on a scale of 0–3, such that 0 indicates no infiltrate; 1, peri-islet infiltrate; 2, mild intraislet infiltrate (<50%); and 3, severe intraislet infiltrate (>50%).
Immunophenotyping
Pancreatic islets, pancreatic lymph nodes (PLNs), and spleens were isolated, dispersed into single cells, and suspended at a concentration of 1 × 106 cells/mL. The single-cell suspension was then stimulated by phorbol-12-myristate-13-acetate (10 ng/mL) and ionomycin (250 ng/mL) followed by brefeldin (3.0 μg/mL) inhibition of cytokine secretion. Next, cells were sequentially incubated with Zombie NIR dye (1:200), Fc-γ receptor blocker (1:50), and monocyte/macrophage blocker (1:20) and then immunolabeled for cell-surface markers (1 μg antibody per million cells), followed by brief fixation and permeabilization before immunolabeling for intracellular FoxP3. The list of fluorophore-conjugated primary and isotype control antibodies is provided in the Supplementary Methods. Immunolabeled cells were acquired on an Aurora 4L spectral flow cytometer and analyzed using FlowJo. Reference spectra for individual fluorophores were generated using microspheres labeled with the same primary antibodies. The spectral data were unmixed, with dead cells excluded, and then gated for individual markers using isotype controls to set threshold levels for nonspecific immunolabeling. Cell populations expressing individual markers were calculated as overall percentages of total CD4+ or CD19+ cells or lymphocytes.
GSIS
Mouse islets were isolated and cultured as described previously (4). Size-matched islets were exposed to mAb43 (10 μg/mL), mAb20 (10 μg/mL), or mouse IgG2b (10 μg/mL) in the presence or absence of a cytokine cocktail consisting of mouse interleukin-1β (0.5 ng/mL), tumor necrosis factor-α (1 ng/mL), and interferon-γ (10 ng/mL). These islets were incubated for 24 h, and then basal insulin secretion, glucose-stimulated insulin secretion (GSIS) with 20 mmol/L glucose and 50 μmol/L IBMX, and total cellular insulin content were measured in triplicate using a homogeneous time-resolved fluorescence immunoassay (19).
Glucose Tolerance Testing
B6 db or NOD mice were fasted for 4 h, and blood glucose level was measured before and after peritoneal glucose injection (2 mg/kg) at 15, 30, 60, and 120 min. Sera were collected for insulin measurement before and after glucose injection at 0, 2, 8, 15, and/or 30 min.
Statistical Analysis
Data were analyzed using R (version 4.3.1) and are expressed as mean ± SE or median (interquartile range) for continuous variables or as count and percentage for categorical variables. P values ≤0.05 were considered statistically significant. R codes and outputs are available in the Supplementary Results.
Data and Resource Availability
All data generated during this study are included in the published article and its online supplementary files. Reagents generated are available upon reasonable request. Johns Hopkins University requires material transfer agreements for all inbound or outbound research materials.
Results
Masking Antigenic Exposure of ZnT8 and Insulin
Glucose stimulation can induce the appearance of ZnT8 on the cell surface of human and rodent β-cell lines as well as primary mouse islet cells (3,4). Immunolabeling of live primary human islet cells with mAb43 also revealed discrete dots or granule-like clusters on the cell surface (Fig. 1A). Double immunolabeling of human EndoC-β1H cells with anti-insulin and mAb43 or anti-CD71 antibody demonstrated partial colocalization of insulin with ZnT8, but not with CD71 (Fig. 1B). In contrast, mAb20, a ZnT8-specific mAb targeting a cytosolic epitope (20), did not immunolabel ZnT8 on the β-cell surface, as its epitope was inaccessible to the extracellular mAb20 (Fig. 1B). Because CD71 is not linked to the insulin secretory pathway, its immunofluorescence was used as an independent in-cell reference to normalize the surface intensity of insulin immunolabeling (Fig. 1C). Notably, extracellular mAb43 significantly decreased the intensity of cell-surface insulin immunofluorescence compared with mAb20 (Fig. 1D), demonstrating that mAb43–ZnT8 binding masks ZnT8 on the β-cell surface and concomitantly reduces the antigenic exposure of insulin, which remains localized to exocytotic sites on the cell-surface membrane (13).
mAb43 Uptake in Human Islet Cells
Cell surface–bound mAb43 underwent rapid internalization at 37°C. A comparison of live human islets exposed to mScarlet-tagged mAb43 and mAb20 demonstrated selective islet uptake of mAb43-mScarlet over mAb20-mScarlet, indicating receptor-mediated mAb43 endocytosis (Fig. 1E). The intracellularly trapped mAb43-mScarlet was confirmed by anti-IgG immunolabeling of paraformaldehyde-fixed islets. Coimmunolabeling for insulin and glucagon revealed the presence of mAb43-mScarlet in both α- and β-cells (Fig. 1F).
mAb43 Treatment Prevents Spontaneous T1D in NOD Mice
Weekly mAb43 administration to mice at a dose of 5 mg/kg can maintain β-cell masking with no detectable off-target biodistribution (4). In the current study, weekly mAb43 injection was initiated at 10 weeks of age, and FBG levels were monitored weekly to track disease development (Fig. 2A). NOD mice receiving either isotype-matched IgG2b or PBS vehicle showed no differences in glycemic profile (Fig. 2B) or diabetes incidence (P = 1.0 by Kaplan-Meier survival analysis with log-rank test). Consequently, mice from the isotype and PBS groups were combined into a single control group. A total of 64 female NOD mice were assigned to the mAb43 treatment group, whereas the control group consisted of 40 female mice. Subsets of mice from each group were randomly selected for terminal assessments at 20, 25, 30, and 35 weeks of age, as well as for evaluation of treatment reversal after discontinuing mAb43 injections at different time points (Fig. 3A). Throughout the study to 35 weeks of age, all mice in the mAb43 treatment group maintained euglycemia, except for one mouse that experienced a hyperglycemic episode from weeks 21 to 25 (Fig. 2C). However, this mouse was still classified as nondiabetic because its FBG subsequently returned to <250 mg/dL (Fig. 2A). In contrast, all mice in the control group developed diabetes by 35 weeks of age (Fig. 2C), underscoring the complete protection provided by mAb43 treatment.
Sustained T1D Protection by Continuous mAb43 Dosing
We continuously monitored five NOD mice to assess long-term mAb43 protection with a delayed treatment initiation at 14 weeks of age (Fig. 2D). This later intervention was inadequate to prevent the onset of acute T1D in one of the five mice. However, the remaining four mice maintained euglycemia until 40 weeks of age, whereas all mice in the control group (n = 13) developed diabetes (Fig. 2E and F). After treatment cessation at 40 weeks of age, all four mAb43-treated mice remained nondiabetic until 75 weeks of age (Fig. 3A right). These findings suggest that mAb43 treatment provided prolonged but not indefinite protection after treatment cessation.
Reversal of mAb43 Protection After Treatment Cessation
To further assess the reversibility of mAb43 therapy, we continuously monitored FBG after discontinuation of mAb43 treatment at 20, 25, 30, or 35 weeks of age (Fig. 3A). Over time, all mice gradually developed diabetes, with varying lag times to disease onset (Fig. 3B). The complete reversal of the mAb43 protective effect indicated a causal relationship between mAb43 treatment and diabetes prevention.
mAb43 Treatment Preserves β-Cell Mass
We conducted immunohistochemistry quantification of insulin staining on cross-sections of NOD pancreata from mice in the treatment and control groups at different time points, with or without diabetes (Fig. 4A). All mAb43-treated mice remained nondiabetic and displayed significantly higher levels of β-cell mass compared with diabetic mice at all ages.
Prolonged mAb43 Treatment Reduces Insulitis
The extent of insulitis reflects autoimmune progression. Grading hematoxylin-eosin–stained pancreatic sections revealed two distinct patterns of lymphocytic infiltration in 20-week-old diabetic and nondiabetic mice in the control group (Fig. 4B). The higher or lower levels of insulitis severity corresponded to NOD subtypes with acute or progressive T1D onset (21). The age-matched mAb43-treated group contained both subtypes and thus exhibited median insulitis severity between that observed in controls with acute or progressive onset (Fig. 4B). Notably, extending mAb43 treatment from 20 to 30 weeks progressively reduced insulitis severity, culminating in a significant clearance of destructive infiltrates by 35 weeks of age (Fig. 4B). Moreover, mice with later diabetes onset at 42–47 and 75 weeks of age had an insulitis rebound (Fig. 4B); however, insulitis in these diabetic mice was less severe and associated with higher residual β-cell mass levels compared with that in control mice with diabetes at 20 weeks of age (Fig. 4A).
Seroreversion of IAAs
Serum insulin autoantibodies (IAAs) are serological biomarkers for insulin-specific autoimmunity. Consistent with a significant clearance of destructive insulitis at 35 weeks of age (Fig. 4B), the same nondiabetic mice exhibited complete IAA negativity (Fig. 4C). This shift to IAA negativity in conjunction with the preservation of β-cell mass (Fig. 4A) is considered a reliable predictor for the induction of insulin-specific immune tolerance in NOD mice (22). Nonetheless, the establishment of this immune tolerance was only temporary, being lost after mAb43 treatment cessation (Fig. 3B). Furthermore, all mice with established diabetes were IAA negative (Fig. 4C), indicative of a significant reduction in insulin secretion that led to a decrease in the stimulation of the insulin-specific B-cells to produce IAAs.
Reversal of New-Onset T1D
We further examined the efficacy of mAb43 therapy in two different diabetic states. First, we examined diabetes reversal in newly diagnosed NOD mice during later stages of adulthood. Accordingly, we delayed diabetes onset with mAb43 treatments from 10 weeks of age followed by discontinuation of mAb43 treatments between 20 to 35 weeks of age (Fig. 3A) and then resumed mAb43 or isotype treatment within 48 h of diabetes onset. There was an average 13-week interval between the cessation of mAb43 pretreatment and the initiation of mAb43 retreatment (Fig. 3B). Given the short half-life of mAb43 pancreatic retention (6 days) (4), mAb43 carryover from the pretreatment was negligible. To avoid the confounding effect of insulin on β-cell survival (23), exogenous insulin was not administered to diabetic mice. All mice receiving mAb43 exhibited a progressive decline in FBG, which eventually stabilized at various levels (Fig. 5A). Nine of 10 mice achieved remission, with one remaining mildly diabetic. In contrast, all control group mice remained severely diabetic until their death (Fig. 5B). Kaplan-Meier survival analysis showed a significant increase in survival compared with the control group (Fig. 5C). Next, we investigated diabetes reversal in young NOD mice without mAb43 pretreatment (Fig. 5D–F). Despite differing ages at diabetes onset, isotype treatments revealed no significant difference in postonset survival between diabetic mice with and without mAb43 pretreatment (P = 0.6 by Kaplan-Meier survival analysis with log-rank test). This absence of significant effects on postonset disease progression was consistent with the inability of mAb43 pretreatment to establish permanent immune tolerance (Fig. 3B). All young diabetic mice experienced a gradual reduction in FBG levels with mAb43 treatment, but this fell short of an achievement of remission before death (Fig. 5D). Nevertheless, mAb43 treatment significantly extended postonset survival compared with isotype treatment in the control group (Fig. 5F). These findings suggest mAb43 therapy is more effective in reversing delayed autoimmune diabetes in older mice.
Histological Characterization of Diabetes Reversal
All diabetic mice, regardless of mAb43 treatment, displayed significantly reduced islet mass compared with nondiabetic mice (Fig. 5I vs. Fig. 4A). In isotype-treated mice, most remaining islets at 3 weeks postonset lacked insulin immunostaining, whereas mice treated with mAb43 and in remission retained a substantial proportion of insulin-positive islets at 12 weeks postonset (Fig. 5G). This preservation of β-cell mass resulted in a noteworthy improvement in blood glucose clearance compared with that in isotype-treated diabetic mice (Fig. 5K). Ki67 immunolabeling revealed proliferating β-cells within the islets (Fig. 5H), although their low frequency suggested minimal contribution to islet expansion. The expanded insulin-positive areas in Fig. 5G were a result of surrounding immune cells that internalized secreted insulin from β-cells. Insulitis persisted in all mice examined, but mAb43 treatment promoted nondestructive peri-islet infiltrates (Fig. 5J). These findings indicate that mAb43 therapy preserves residual β-cell mass and function in the presence of ongoing insulitis.
Islet-Resident Treg Induction
The immunoregulatory effects of mAb43 on insulitis were investigated ex vivo using pancreatic sections from 20-week-old nondiabetic mice that received weekly injections of either mAb43 or isotype control, starting at 10 weeks of age (n = 2 × 10). Anti-CD4, FoxP3, and CD25 coimmunolabeling revealed no significant differences in the numbers of islet-residing CD4+ T cells between the mAb43 and control groups (Fig. 6A and B). However, mAb43 treatment significantly increased the number of CD4+FoxP3+CD25+ T cells within the islets (Fig. 6B). Notably, FoxP3 and CD25 immunofluorescence exhibited a high degree of colocalization in all FoxP3+ islets. As a result, CD4+FoxP3+CD25− T cells were not detected (Fig. 6B). Thus, mAb43 treatment increased the Treg population within islet infiltrates.
Islet-Localized Tolerogenic Immune Responses
The broader immunoregulatory effects of mAb43 were further examined using high-dimensional multiplexing flow cytometry to phenotype tissue-resident lymphocytes isolated from 20-week-old nondiabetic mice subjected to mAb43 or isotype treatment from 10 weeks of age (n = 2 × 7). mAb43 did not significantly alter the percentages of CD19+ B cells, CD4+ T cells, or CD8+ T cells within the lymphocyte populations residing in islets, PLNs, or spleens (Fig. 6C–E). Similarly, mAb43 treatment had no significant impact on the percentages of CD19+CD40+, CD19+CD80+, or CD19+MHCII+ cells in these tissues (Fig. 6F–H). Although major immune cell subsets and their phenotypes remained largely unaffected, mAb43 treatment resulted in a shift in the islet-residing CD4+ T cells toward higher FoxP3 expression levels (Fig. 6I), leading to a significant increase in the percentage of CD4+FoxP3+ T cells (Fig. 6J). In contrast, the percentage of CD4+FoxP3+ T cells in the PLNs and spleen remained unchanged (Fig. 6I and J). Therefore, the immunophenotypic changes induced by mAb43 were primarily confined to the CD4+FoxP3+ pool in pancreatic islets, shifting the islet lymphocyte infiltrates toward a tolerogenic phenotype.
Lack of Detectable mAb43 Effects on β-Cell Functions
To investigate the direct impact of mAb43 therapy on the insulin secretory function of β-cells, we assessed insulin secretion and production using isolated islets from B6 mice, which were free from the confounding effects of insulitis. Although proinflammatory cytokine exposure increased basal insulin secretion regardless of mAb43 treatment (Fig. 7A), mAb43 alone or in combination with cytokines did not induce significant differences in GSIS or total insulin content within the islets (Fig. 7B and C). This observation is supported by data from young NOD mice before developing overt insulitis, where mAb43 injections from 3 to 6 weeks of age did not affect in vivo GSIS compared with isotype-treated mice (Fig. 7D). In contrast, mAb43 treatment of adult diabetic NOD mice preserved in vivo insulin secretion against persistent insulitis (Fig. 7E). Moreover, in an obesity-induced diabetes mouse model, weekly administration of mAb43 or isotype control to male B6 db mice from 6 to 17 weeks of age did not induce any signs of insulitis (Fig. 7F) or antibody-dependent cellular cytotoxicity, as evidenced by the absence of significant differences in FBG levels (Fig. 7G) and glucose tolerance tests at 10 and 15 weeks of age (Fig. 7H and I). In summary, our findings indicate that mAb43 therapy preserves β-cell function in the face of insulitis. The absence of antibody-dependent cellular cytotoxicity may be attributed to rapid internalization of surface-bound mAb43 in vivo.
Discussion
The experiments described herein demonstrate the efficacy of mAb43 in preventing and reversing new-onset T1D in NOD mice. These findings suggest that mAb43 acts on critical steps in autoimmune responses leading to T cell–mediated destruction of β-cells. In this process, autoantigens on the β-cell surface could act as molecular attractants, driving antigen-specific homing of autoreactive B cells to the islets. Infiltrating B cells can then acquire cell-surface autoantigens, generating T-cell epitopes for presentation to autoreactive T cells, thereby directing their homing to the islets and differentiation into cytotoxic effectors (24). Of note, autoreactive B cells are present in a substantial proportion of the circulating B-cell population in healthy individuals (25). ZnT8 is a major membrane-bound autoantigen (26). Surfaced ZnT8 may facilitate BCR cross-linking to gather ZnT8 into the B-cell synapse, leading to efficient ZnT8 uptake and processing and presentation of ZnT8-derived T-cell epitopes (Fig. 6K). Indeed, a ZnT8186–194 peptide is one of the dominant T-cell epitopes driving islet homing of pathogenic CD8+ T cells in human T1D (27).
The mechanism of mAb43 actions involves masking ZnT8 from B-cell recognition (Fig. 6L), preserving the pool of naive ZnT8-reactive B cells that can interact with cognate CD4+ T cells to generate inducible Tregs in the islet (28). This notion is in line with an increased local Treg number within the islet of mAb43-treated NOD mice (Fig. 6A and B). Tregs play a crucial role in suppressing cytotoxic CD8+ T-cell activation and promoting the regulatory milieu by secreting proinflammatory interleukin-10 (29). Recently, the US Food and Drug Administration approved the first-ever T1D-modifying therapy, teplizumab, an anti-CD3 mAb that suppresses immune responses through transient depletion of T cells and modulation of the CD3/T-cell receptor complex (30). In NOD mice, a side-by-side comparison between teplizumab and mAb43 revealed that both therapies effectively reverse new-onset T1D (31), but only mAb43 exhibited a safety profile applicable to chronic use.
Discontinuing mAb43 treatment caused the halted autoimmunity progression toward T1D to resume, even in elderly NOD mice (Fig. 3). This finding underscores the necessity for ongoing mAb43 treatment to achieve sustained efficacy. The chronic use of mAb43 may alter ZnT8-mediated zinc accumulation in the insulin secretory granule, where zinc is cocrystallized with insulin to form the insulin-dense core (32). Our data show that exposing mouse islets to mAb43 does not alter GSIS or islet insulin content ex vivo (Fig. 7). Previous work has demonstrated that ZnT8-knockout mice maintain normal insulin biosynthesis and GSIS despite the absence of zinc accumulation in insulin secretory granules (33). Additionally, ZnT8 haploinsufficiency in NOD mice protects against β-cell dysfunction by increasing mitochondrial respiration (34). Furthermore, downregulation of ZnT8 expression in human EndoC-β1H cells provides protection against inflammatory stress by reducing the burden of unfolded protein response to ZnT8 misfolding (19). In humans, ZnT8 haploinsufficiency has been associated with protection against type 2 diabetes, without adverse effects (35,36). The lifelong protective effects of ZnT8 downregulation demonstrate the safety of sustained ZnT8-suppressive therapy in humans (37). These findings suggest that ZnT8-specific mAb43 therapy is a viable long-term solution to prevent or reverse new-onset T1D.
T1D in humans exhibits clinical heterogeneity associated with genetic variations, diverse autoimmune responses, and environmental influences. Previous preclinical findings from inbred NOD mice have shown limited success when translated into clinical trials (38). However, mAb43 therapy stands out as a categorically distinct modality from all established lymphocyte-targeted immunotherapies. Unlike other approaches, mAb43 does not directly modify lymphocytes but rather preserves β-cells in the presence of insulitis. mAb43 binds equally to human and mouse ZnT8 on highly conserved and membrane-flush cell-surface epitopes (4), ensuring the preservation of mAb43-mediated β-cell masking regardless of differences between human and mouse immune systems. Importantly, ZnT8 is a major target of pathogenic CD8+ T cells in human T1D (27,39), further supporting the clinical relevance of ZnT8-masking therapy. Islet autoantibodies maternally transferred to the fetus seem to be protective against future islet autoimmunity in children born to mothers with T1D (40–42). This protective role of human islet autoantibodies is mirrored by the protective effects of mAb43, functioning as a cell-surface autoantibody to provide long-term β-cell protection without adverse effects.
Article Information
Acknowledgments. The authors thank Xiaoling Zhang from the Ross Flow Cytometry Core, Johns Hopkins University School of Medicine, for assistance in flow cytometry data acquisition and analysis.
Funding. This study was supported by National Institutes of Health (NIH) grants R01DK125746 (D.F.), P30DK116073 (L.Y.), R01DK110183 (M.L.G.), R01DK135688 (M.N.P.), and R01DK084171 (G.W.W.). The Aurora 4L spectral flow cytometer was funded through NIH grant S10OD026859. The Zeiss confocal microscope was funded through NIH shared instrumentation grant S10OD016374. Human pancreatic islets (RRID: SAMN34130383) were provided by the National Institute of Diabetes and Digestive and Kidney Diseases–funded Integrated Islet Distribution Program (RRID: SCR _014387) at City of Hope (NIH grant U24DK098085).
Duality of Interest. No potential conflicts of interest relevant to this article were reported.
Author Contributions. D.K., Z.G., A.W.M., and D.F. were responsible for immunophenotyping and data interpretation. D.K., D.C.S., and G.W.W. performed mouse experiments. D.K. and M.L.G. performed immunohistology. Z.G. was responsible for antibody production and β-cell experiments. S.Y. was responsible for biostatistics and data deposit. L.Y. performed insulin autoantibody measurement. C.S. and M.N.P. performed human islet immunolabeling and imaging. D.F. was responsible for conceptualization, experimental design, and manuscript writing. All authors read and edited the manuscript. D.F. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
This article contains supplementary material online at https://doi.org/10.2337/figshare.25254211.
This article is featured in a podcast available at diabetesjournals.org/diabetes/pages/diabetesbio.