Evidence for C-Peptide as a Validated Surrogate to Predict Clinical Benefits in Trials of Disease-Modifying Therapies for Type 1 Diabetes Free

Type 1 diabetes is a chronic autoimmune disease that targets destruction of pancreatic β-cells, the only cells in the body that manufacture and secrete insulin. While this autoimmunity selectively targets β-cells, other hormone-producing cells in pancreatic islets (e.g., α-cells and δ-cells) may also be affected, contributing to metabolic dysregulation. Autoimmune destruction of β-cells in type 1 diabetes is initially subclinical, usually developing over a period of months or years, until insulin secretion can no longer maintain normal glucose homeostasis. At clinical presentation, most affected individuals are dependent on exogenous insulin administration but can produce endogenous insulin, measured as C-peptide (for the connecting peptide cleaved from proinsulin) at levels that have been associated with protection from secondary end-organ complications, such as retinopathy and nephropathy (1,2). However, the loss of β-cell function continues after diagnosis, and even in adults, where disease progression is slower, those not initially requiring insulin become insulin dependent within 3–5 years (3). Insulin production can be detected even 50 years after diagnosis in as many as 67% of patients, but the clinical significance of very low levels of residual insulin production in patients who have been treated long-term with exogenous insulin has not been established (4). Understanding these relationships is important for evaluating therapies that are aimed at preserving β-cell function in the new-onset setting.

The current approach to diabetes management (other than islet or pancreas transplantation) involves lifelong insulin replacement. While significant advances have been made in methods for insulin delivery, measurement of glucose levels, and development of analog insulins, current insulin therapy is neither physiologic nor glucose responsive, and it poses significant risks to recipients (5,6), including hypoglycemia, which can be serious or fatal. Suboptimal insulin therapy leads to hyperglycemia and diabetic ketoacidosis (DKA) and, chronically, long-term end-organ complications, such as retinopathy, kidney failure, neuropathy and vascular disease with limb loss, and heart failure (7). Uptake of new technologies in practice remains limited, the optimal utilization remains complex and elusive, and most patients still fail to achieve treatment targets as recommended by the American Diabetes Association (8). Globally, estimates for a 10-year-old child who develops type 1 diabetes show a 24-year gap in remaining life expectancy compared with a child without type 1 diabetes; losses are even greater in lower-income countries (9).

Insulin replacement and associated management technologies do not affect the underlying autoimmune disease process. Disease-modifying therapies have been identified that can change the trajectory of the disease but do not restore the β-cells already lost. Thus, interest has focused on using these therapies early in the disease course, when functional β-cells are still present, and identifying the clinical significance of residual β-cell function that can be preserved by newer therapies. In this Perspective, we summarize key clinical study data supporting acceptance of stimulated β-cell function (i.e., C-peptide responses) as the fundamental measure of disease progression in type 1 diabetes and, thus, an appropriate surrogate outcome to assess efficacy of disease-modifying therapies, given the observed associations between C-peptide responses, HbA_1c, insulin use, hypoglycemia, and long-term complications across different study designs.

β-Cell function is routinely measured via the C-peptide response to a standardized oral glucose tolerance test (OGTT) or mixed-meal tolerance test (MMTT). C-peptide is a direct, quantitative, and specific measure of glucose-dependent insulin secretion from β-cells. Cleaved from proinsulin by prohormone convertase enzymes within the β-cell secretory granules, the connecting peptide (C-peptide) is released from the linking insulin A- and B-chains and cosecreted in an equimolar ratio with the now-active hormone insulin (10).

Measurement of C-peptide allows for feasible and specific assessment of endogenous insulin production, even in individuals receiving exogenous insulin (which lacks C-peptide). C-peptide is exclusively cleared by the kidneys from the peripheral circulation at a constant rate, unlike insulin, which is metabolized in the liver and cleared by the kidneys (10). C-peptide is stable in specimens prepared from peripheral blood for clinically feasible storage times at multiple temperatures (11), and harmonization efforts for quantitative assays for C-peptide that can be standardized against international reference materials are ongoing (12).

C-peptide measurement during OGTT standardized to 1.75 g/kg (up to 75 g) glucose is currently the preferred assessment for stimulated β-cell function in type 1 diabetes before clinical presentation to identify glucose intolerance or dysglycemia (13,14). MMTTs have been used to assess β-cell reserve in patients after diagnosis, because the mixed-nutrient stimulus results in a more physiologic stimulation of insulin secretion (15) and is achieved with a lower glucose challenge (up to 50 g carbohydrate), and classification of the level of dysglycemia against the reference standards of impaired glucose tolerance and clinical diabetes is not relevant once the diagnosis is established. In established diabetes, MMTT area under the curve (AUC) and peak C-peptide measurements each reflect β-cell secretory capacity (2). The C-peptide response to MMTT is highly reproducible (16), thus estimating the residual β-cell function that is important for regulation of glucose homeostasis (2,17).

Collectively, multiple physiologic, biometric, and practical qualities make C-peptide an ideal target of measurement suitable for use as a specific and reliable biomarker for β-cell function (18).

Disease-modifying therapies, by definition, have a mechanism of action targeted to modification of the underlying autoimmune disease process in type 1 diabetes. In contrast, glycemic outcomes are heavily affected by advances in glycemic management, such as the use of continuous glucose monitors (CGM) and insulin pumps, which now provide automated insulin delivery based on CGM sensor glucose data. To support acceptance by regulators, payors, and prescribers, outcomes in product development studies must be considered clinically meaningful by both regulators and clinicians. Generally accepted outcomes that fulfill this criterion are well established (level 3 and level 2 hypoglycemia events [7], microvascular complications, HbA_1c, etc.). However, designing studies to achieve statistical significance with these end points is usually not feasible in patients with new-onset type 1 diabetes, where level 3 and level 2 hypoglycemia events are uncommon, and microvascular and end-organ complications take years to develop. Measurements of end points based on end-organ complications also can be affected by therapies that may be introduced as part of medical practice, given the time needed to accrue these events. Hence, powering trials to meet traditional efficacy end points in a population with new-onset type 1 diabetes would require prohibitively large numbers of enrolled subjects and/or prohibitively long follow-up.

Studies of control groups from clinical trials have enabled robust models of predicted C-peptide decline in the first year after diagnosis of type 1 diabetes; an individual’s C-peptide decline in the first year of stage 3 disease can be estimated based on age and baseline C-peptide at diagnosis (19,20). This modeling may support efficient design for clinical trials in new-onset type 1 diabetes.

Data from various clinical trials have addressed the significance of preserved or restored β-cell function. These clinical trials include longitudinal studies, cross-sectional studies, interventions with immune therapies to preserve β-cell function, and replacement of β-cells with islet or pancreas transplantation. These trials, involving different patient groups, data acquisition, and interventions, have all identified clinically significant benefits from preservation or restoration of β-cell function measured from stimulated C-peptide concentrations. Here, we summarize these data.

The landmark Diabetes Control and Complications Trial (DCCT) was a 10-year randomized study of intensive versus conventional insulin therapy in 1,441 type 1 diabetes patients, and it demonstrated that intensive insulin therapy significantly reduces the risk of microvascular complications (21). This study established HbA_1c as a clinically meaningful, validated surrogate outcome measure for efficacy assessment of glucose-lowering therapies (22). The DCCT demonstrated significant benefits of intensive insulin therapy but also showed that intensive treatment resulted in a significant risk of hypoglycemia.

Subsequent analysis demonstrated that residual β-cell function, measured as MMTT-stimulated C-peptide at 90 min, changed this relationship (21). Individuals on intensive insulin therapy with higher stimulated C-peptide levels at study entry (200–500 pmol/L [responders]) achieved benefits of reduced complications (especially reduced retinopathy progression) while also having lower rates of hypoglycemia than intensive therapy participants, who had stimulated C-peptide <200 pmol/L at study entry (nonresponders) (Table 1). During 1983–1989, 855 subjects with type 1 diabetes of 1–5 years’ duration were enrolled, of whom 412 were randomized to intensive therapy. Over 9 years of follow-up, with declining sample size owing to staggered entry, C-peptide responders maintained a lower HbA_1c than C-peptide nonresponders. After adjusting for baseline HbA_1c, responders in the intensive therapy group had a 65% lower risk of severe hypoglycemia than nonresponders. Furthermore, these C-peptide responders had a 50% lower risk of retinopathy through the first 4 years of follow-up compared with nonresponders (Table 1).

Among subjects randomized to standard therapy, responders also experienced lower retinopathy rates and significantly lower hypoglycemia rates than nonresponders; however, differences were smaller and of shorter duration than in the intensive therapy group. Over 7 years of follow-up, after adjusting for baseline HbA_1c, a strong inverse association was observed between higher entry stimulated C-peptide and lower HbA_1c, and this association was similar at every year of follow-up (23).

In subsequent follow-up of the DCCT cohort (Epidemiology of Diabetes Interventions and Complications [EDIC] study 2015–2017), stimulated C-peptide levels in 944 surviving participants (average diabetes duration of 35 years) were measured using a new high-sensitivity assay (24). C-peptide secretion at this point was substantially lower. Nonetheless, >10% of EDIC participants had detectable stimulated C-peptide (assay lower limit of quantification 3 pmol/L), and participants with residual C-peptide >30 pmol/L had substantially lower risk of hypoglycemia than nonresponders despite similar HbA_1c levels and comparable insulin use (Table 2). Additional analyses (23) explored the relationships between the quantitative level of baseline C-peptide and outcomes in the intensive therapy group. These analyses demonstrated that while preservation of stimulated C-peptide at ≥200 pmol/L has clinically beneficial outcomes, so does an increase in the concentration of C-peptide across the range of values.

In summary, the long-term follow-up afforded by the DCCT/EDIC indicates that higher entry levels of C-peptide (at 1–5 years postdiagnosis) are associated with improved HbA_1c and substantially reduced long-term complications during 7 years of follow-up, as well as reduced hypoglycemia rates with up to 35 years of follow-up. The result from the DCCT/EDIC is an adjusted-within-cohort comparison rather than comparison with a matched control or a randomized control group from the same sample; hence, additional data from C-peptide preservation or restoration studies support these conclusions (see islet transplant and TOMI-T1D data below).

The Scottish Diabetes Research Network Type 1 Bioresource Study (SDRNT1BIO) prospectively examined the relationship of baseline C-peptide levels with glycemic outcomes and microvascular complications among 6,076 adults with type 1 diabetes in Scotland who were monitored for an average of 5.2 years (25). Time points in this study assessed random nonfasting C-peptide using an immunoassay (26) with a lower limit of detection of 3 pmol/L.

The authors reported cross-sectional analysis between baseline C-peptide and history of severe (level 3) hypoglycemia in the preceding year and baseline retinopathy status as well as the prospective relationship between baseline C-peptide and incident hospitalization for hypoglycemia and incident retinopathy (Table 3) (25). Both baseline and prospective analyses showed strong relationships between nonfasting C-peptide and these clinical outcomes (Table 3 and Fig. 1). In the prospective analysis, after adjusting for covariates of sex, BMI, age at diagnosis, and duration of disease in a linear regression model, nonfasting C-peptide was strongly and significantly associated with reductions in incidence of both retinopathy and hypoglycemia.

Effects on the risk of serious hypoglycemic episodes were detectable at lower levels of C-peptide, and the relationship was continuous down to the lower limit of detection of the test. The relationship of incident retinopathy to nonfasting C-peptide levels at baseline was approximately linear, with no evidence of threshold effect. Participants with C-peptide levels of 30–200 pmol/L had about one-half the incidence of admissions for severe hypoglycemia and about two-thirds the incidence of retinopathy compared with those with undetectable C-peptide.

Taken together, this large cohort with >5 years’ average follow-up provides important real-world data of the association between C-peptide and clinical outcomes in the broader population outside of clinical studies.

Transplantation of a whole pancreas or isolated pancreatic islets provides robust evidence for C-peptide as a biomarker of β-cell function that predicts clinically meaningful benefits. These transplants can restore islet function in patients who previously had very low or undetectable stimulated C-peptide levels (27).

In the phase 3 CIT-07 trial of the of islet-alone transplantation and the phase 3 CIT-06 trial of islet-after-kidney transplantation, both from the Clinical Islet Transplantation (CIT) Consortium, restoration of β-cell function yielded substantial and durable improvement in HbA_1c, CGM measures of average glucose levels, glucose variability, time-in-target-range glycemia, frequency of serious, clinically important hypoglycemia, and frequency of severe hypoglycemia events (SHEs) (28,29). In long-term follow-up of both studies (CIT-08), protection from severe hypoglycemia was maintained in >90% of participants up to 8 years posttransplant, with ∼80% maintaining MMTT-stimulated C-peptide ≥100 pmol/L (mean C-peptide >1,300 pmol/L between 3 and 5 years posttransplant) with reduced insulin requirements and >25% of recipients maintaining insulin independence (30).

In a nationally funded islet transplant program in the U.K., higher MMTT-stimulated C-peptide levels were associated with increasingly improved benefits, including additional clinical outcomes (31). Increasing β-cell function across predefined stimulated C-peptide groups was associated with reduced insulin dose, HbA_1c, mean glucose (low [<200 pmol/L], 10.7; excellent [>1,000 pmol/L], 7.5 mmol/L), and glucose SD (low, 4.4; excellent, 1.4 mmol/L). Moreover, highly statistically significant continuous associations were found between stimulated C-peptide and mean interstitial glucose (lower by 2.5% per 100 pmol/L higher C-peptide), glucose SD, time outside glucose target range, and measures of hyperglycemia/hypoglycemia risk. Nevertheless, only islet transplant recipients with excellent (stimulated C-peptide >1,000 pmol/L) islet graft β-cell function were afforded insulin independence, suggesting a minimum threshold of reserve capacity for insulin secretion required to maintain near-normal glycemic control without exogenous insulin.

A recent analysis from the Collaborative Islet Transplant Registry (CITR), involving 677 islet transplant recipients with baseline C-peptide <100 pmol/L followed for a median of 5 years after transplantation, applied receiver operating characteristics analysis to determine thresholds of fasting and MMTT-stimulated C-peptide, with and without correction for glucose, that best predicted absence of severe hypoglycemia, HbA_1c ≤6.5%, and insulin independence (key data are shown in Table 4) (32). Consistent with previous studies discussed above, stimulated C-peptide as low as 120 pmol/L predicted absence of SHEs (C-peptide <100 pmol/L was assessed as negative). Higher stimulated C-peptide (800 pmol/L) predicted HbA_1c ≤6.5%, while even higher stimulated C-peptide (>970 pmol/L) predicted insulin independence. Absence of SHEs is especially notable in this population, since at baseline a recent history of SHEs is typically a requirement for transplant eligibility. Note that as shown in Table 4, thresholds for fasting levels can also be determined (usually two- to threefold lower than stimulated levels) but come at some cost in predictive value, and, in general, fasting levels are less sensitive to change than stimulated responses.

Other studies of islet transplant recipients have found additional clinical benefits from this intervention. A prospective crossover cohort study of 45 islet transplant recipients, comparing islet cell transplantation with intensive medical therapy, found that transplant recipients experienced significantly slower declines in renal function (e.g., glomerular filtration rate decrease −1.42 vs. −4.79 mL/min/1.73 m²/year; Ρ < 0.0001) and less progression of retinopathy (e.g., 0/51 vs. 10/82 eyes with any progression of diabetic retinopathy; Ρ < 0.01) versus individuals receiving intensive insulin therapy while on the transplant waiting list (33). In a separate 5-year prospective longitudinal study of 21 islet transplant recipients, significant improvements were observed over the 5-year posttransplant period in measures of sensory nerve function; for example, baseline versus 5 years’ posttransplant sensory nerve assessments included median conduction velocity 47.5 versus 41.2 m/s (P < 0.01) and median action potential 5.5 versus 11.8 mV (also P < 0.01) (34). Nerve function improvements correlated with overall improvement in median fasting/stimulated C-peptide from 0 at baseline to 298/894 pmol/L at 5 years (P < 0.0001).

Islet transplantation is also associated with significant reductions in diabetes distress and fear of hypoglycemia and increased self-assessment of personal well-being (29,35). These improvements are not significantly associated with insulin independence, indicating that even when exogenous insulin is still required, restored β-cell function is associated with benefits that are meaningful to transplant recipients.

In summary, these successful transplant interventions provide evidence of a prospective association between MMTT-stimulated C-peptide and clinical benefits (improved glycemic control and insulin independence at the higher end of the restored C-peptide and protection from hypoglycemia at the lower end) as well as providing estimates of threshold levels for these effects (Table 4). Furthermore, these data provide information on quality of life and add to the evidence from the DCCT/EDIC on reduction in long-term complications.

Multiple disease-modifying therapies aimed at altering autoimmunity have been investigated for potential benefits in preserving endogenous β-cell function in those with new-onset type 1 diabetes (36). To date, at least nine disease-modifying therapies tested in phase 2 trials and one phase 3 trial, with different drug targets and thus different mechanisms of action, have shown significantly increased stimulated C-peptide levels versus placebo-treated subjects after 12–24 months of intervention (37–48). However, these phase 2 trials were not designed with the power or follow-up duration to conclusively report on secondary outcomes, such as HbA_1c or insulin use (typically, <100 subjects were enrolled per trial) (Table 5).

TOMI-T1D (Trial Outcome Markers Initiative in Type 1 Diabetes) is an ongoing JDRF-funded platform project incorporating patient-level data previously collected in 21 randomized controlled trials (RCTs) of disease-modifying interventions, including 2,901 subjects within 100 days of diagnosis (49). The platform includes models for six parameters: C-peptide AUC during a 2-h MMTT, glucose AUC during a 2-h MMTT, insulin use over time, HbA_1c, frequency of level 3 hypoglycemic events (severe cognitive impairment or requiring external assistance), and frequency of either level 2 (blood glucose <54 mg/dL) or level 3 hypoglycemia. The long-term aim of the project is to develop a clinical trial simulation platform to model the progression of new-onset type 1 diabetes (49). Combined analysis of 8 of the 21 studies that achieved their primary end point demonstrated that preservation of C-peptide data in prospective studies reduced HbA_1c up to 24 months from diagnosis (Fig. 2A). Reductions in insulin use were less marked (perhaps reflecting less aggressive management in these older studies), but improvements in insulin dose–adjusted A1C were highly significant. Across all 21 studies, stimulated C-peptide levels both at study entry and at 1 year were strongly associated with reduced level 2/3 hypoglycemia rates (Fig. 2B). Thus, in this meta-analysis, subjects treated prospectively with multiple interventions designed to preserve residual β-cell function, acting via different biochemical targets and using multiple mechanisms of action, experienced improved clinical end points (HbA_1c and insulin dose–adjusted A1C) proportionate to the degree of stimulated C-peptide preservation. Correlation between C-peptide preservation and reduction in hypoglycemia was also confirmed.

Disease-modifying therapies could modify the autoimmune disease process in type 1 diabetes, thus preserving β-cell function. Given the continued decline of β-cell function over time in type 1 diabetes, the greatest opportunity to intervene with disease-modifying therapies exists early in the autoimmune disease process. However, development of disease-modifying therapies is limited by the tools available to assess efficacy. Clinical trial end points accepted as surrogates for clinical benefit in diabetes have been validated for glycemic control agents, but compared with C-peptide, these end points are less sensitive to the impact of disease-modifying therapies. This is especially the case in new-onset type 1 diabetes studies where end-organ complications and severe hypoglycemia events are infrequent, and glycemic control metrics are subject to confounding by intensive insulin therapy. This confounding has become more apparent in recent RCTs in which the use of advanced insulin delivery technology and CGM has increased (46,48). All these metrics therefore require impractically large and/or long trials to observe therapeutic impacts from disease-modifying therapies in new-onset type 1 diabetes. For commercial entities considering development of new therapies, it is impractical to rely on clinical benefits that require a 10- to 20-year follow-up to wait for complications. Disease-modifying therapy trials therefore require an end point that can identify clinically meaningful impacts in early-stage disease, and in a relatively small population, given the limited number of people developing new-onset type 1 diabetes at any one time. C-peptide is well established as a biomarker of β-cell function (18) but has not been universally accepted as a surrogate end point for clinical benefits of therapeutic interventions.

The studies summarized above provide complementary data from different study designs, each adding to the other to create a consistent and robust picture (Table 6). The long-term follow-up data from the DDCT/EDIC cohort demonstrate associations between higher initial C-peptide and reduced HbA_1c in the medium term as well as fewer long-term complications and less hypoglycemia in subsequent years. The SDRNT1BIO data extend these associations from clinical trials to a real-world population showing reductions in hypoglycemia, DKA risk, and retinopathy associated with higher C-peptide levels at study entry. Nonetheless, these are associations, and showing a causative link requires prospective intervention studies. The data from islet transplantation provide this information while also indicating likely thresholds for different clinical benefits, with stimulated C-peptide levels >800 pmol/L for improved glycemic control and >970 pmol/L for insulin independence (Table 4), although this will vary widely between individuals due to differences in insulin sensitivity. Stimulated levels as low as 120 pmol/L were associated with reduced hypoglycemia in the CITR, and this result was consistent with benefits in terms of hypoglycemia observed at ≥30 pmol/L in both SDRNT1BIO and DCCT/EDIC. Finally, circumventing the challenges of limited sample size in trials powered on C-peptide, the TOMI-T1D meta-analysis has demonstrated robust and proportionate reductions in HbA_1c in the 2 years after diagnosis in prospective clinical interventional trials of C-peptide preservation therapies.

In the absence of very large, long-term intervention studies in new-onset type 1 diabetes, taken together, these data provide compelling evidence that measuring C-peptide for β-cell function represents a validated surrogate end point in type 1 diabetes to support risk-benefit assessments for medical product development.

Although therapeutic options have improved for type 1 diabetes, there are still significant unmet needs to improve care and outcomes that can be addressed by disease-modifying therapies. However, for optimal benefit, disease-modifying therapies need to be used in early-stage disease, and the clinical efficacy end points traditionally used in type 1 diabetes studies are ill-suited to assessments in this setting due the low rates of hypoglycemia and long-term complications at this stage and the reduced power afforded by the limited size of the new-onset disease population. In contrast, changes in C-peptide represent a highly appropriate and relevant end point for disease-modifying therapies in those recently diagnosed with type 1 diabetes. The extensive body of evidence summarized here indicates that the association between stimulated C-peptide and clinical outcomes, including glycemic control and rates of hypoglycemia and long-term complications, is a consistent finding using multiple approaches with various study designs (e.g., longitudinal, cross-sectional, and interventional). This provides strong support for the acceptance of stimulated C-peptide as a validated surrogate end point for clinical trials of disease-modifying therapies in people with type 1 diabetes. Such acceptance will allow much more rapid progress in the development of disease-modifying therapies and the transfer of the benefits of this approach to patients.

Acknowledgments. We thank Michelle A. Cissell and JDRF staff member Emily DiLossi for editorial review and document quality control. We thank JDRF staff members Stephen Karpen and Campbell Hutton for critical review of the manuscript.

Duality of Interest. E.L. is a former employee of Regeneron Pharmaceuticals and owns company stock. C.J.G. reports hosting a clinical trial site for Provention Bio’s Phase 3 Trial Evaluating Teplizumab in Patients With Recent-Onset Type 1 Diabetes (PROTECT) studies, has received in-kind research support for investigator-initiated studies from Bristol-Meyers Squibb and Takeda, and serves on the independent data monitoring committee for the Verapamil SR in Adults With Type 1 Diabetes (Ver-A-T1D) trial. H.M.C. reports grants or contracts from IQVIA, JDRF, the Chief Scientist Office, Diabetes UK, Medical Research Council (UK Research and Innovation), and the EU Commission; honoraria for a speaker’s bureau for Novo Nordisk and an educational event for Medscape; participation on a data safety monitoring board or advisory board for Novo Nordisk and for Bayer AG; and stock or stock options in Bayer AG and Roche Pharmaceuticals. J.S.S. reports having been an advisor to 4immune Therapeutics, Abvance Therapeutics, Precigen ActoBio, Adocia, Altheia, Arecor, AstraZeneca, Avotres, Bayer, COUR Pharmaceuticals, Cue Biopharma, Dance Biopharm/Aerami, Dexcom, Diasome Pharmaceuticals, Enthera, Imcyse, IM Therapeutics, Kriya Therapeutics, Novo Nordisk, Oramed, Orgenesis, Provention Bio, Sanofi, Signos, Vertex Pharmaceuticals, and Viacyte, has stock or stock options in 4immune Therapeutics, Abvance Therapeutics, Aerami, Applied Therapeutics, Avotres, Dexcom, IM Therapeutics, Oramed Pharmaceuticals, Orgenesis, and Signos, is a member of the board of directors of Applied Therapeutics, and is chair of the Strategic Advisory Board of the EU INNODIA consortium. M.R.R. reports consulting fees from Sernova Corp. and Vertex Pharmaceuticals and research support from Dompe Farmaceutici S.p.A. K.C.H. has been a consultant to Sanofi Pharmaceuticals and to Provention Bio and is on the scientific advisory boards of NexImmune and Sonoma Biotherapeutics. He is a coinventor on a patent for teplizumab for delay of type 1 diabetes in at-risk individuals. No other potential conflicts of interest relevant to this article were reported.

Prior Presentation. Parts of this study were presented in oral form at the 19th IDS (Immunology of Diabetes Society) Congress, 23–27 May 2023, Paris, France.