Toward Disease-Modifying Therapies in Type 1 Diabetes: Focus on Teplizumab

Abstract

The worldwide incidence of type 1 diabetes continues to rise at an alarming rate. One hundred years after the introduction of insulin, the long-entertained hope of moving from symptomatic treatment to disease-modifying therapies is finally taking shape with regulatory approval of teplizumab to delay the onset of stage 3 disease. Here we review teplizumab’s mechanism of action, setting it against the background of emerging disease-modifying therapies for clinical practice, in language accessible to practicing clinicians. A clinical diagnosis of type 1 diabetes and insulin dependence results from progressive autoimmune destruction of pancreatic β-cells as part of a complicated dialogue between the immune system and the islet. Infusion with teplizumab, a humanized monoclonal antibody that binds the ε-chain of the T lymphocyte CD3 molecule, delays progression from stage 2 to clinical stage 3 type 1 diabetes by almost 3 years. The mechanism of action of teplizumab involves partial agonistic signaling via CD3/TCR and subsequent deactivation, promoting exhaustion of pancreatic β-cell-reactive CD8⁺ T lymphocytes and induction of regulatory T lymphocytes, thereby restoring self-tolerance. With regulatory approval of this agent, clinical practice has entered a new era for treating people with type 1 diabetes, in which disease modification can become the new standard of care. Implementation of global screening for autoantibodies and dysglycemia is underway, enabling efforts to intervene during asymptomatic stages of the disease before insulin treatment is required.

Graphical Abstract

Introduction

Although disease management tools like glucose monitoring, insulin analogs, and automated insulin delivery systems have altered treatment of type 1 diabetes, the disease remains associated with both acute and chronic complications and significant effects on quality of life (1–3). For example, 20%–50% of people with a new diagnosis of clinical type 1 diabetes (stage 3), particularly children, present with diabetic ketoacidosis, a complication associated with considerable morbidity and mortality (4–6). Despite advancements in symptomatic therapeutic approaches, chronic microvascular complications as well as cardiovascular disease remain prevalent and lead to excess mortality and a shorter lifespan in those living with type 1 diabetes (1). In addition, affected individuals and their families fight a daily struggle to achieve tight blood glucose levels and the desired target of HbA_1c <7% (the benchmark established more than 30 years ago for reducing the risk of chronic microvascular complications), while avoiding the opposing risks of acute hypoglycemia (7,8). The near-complete loss of endogenous β-cell function makes this goal challenging to achieve.

The prevalence of type 1 diabetes is increasing worldwide (9). The greatest increase in incidence is among the very young (ages <5 years) through adolescence, but 50% of people with type 1 diabetes are diagnosed as adults (>18 years), often after misdiagnosis with type 2 diabetes (10). In 2021, among the 8.4 million individuals worldwide living with type 1 diabetes the median age of onset was 29 years (9). That year, modeling to estimate global incidence, prevalence, and mortality suggested that the remaining life expectancy of a 10 year old diagnosed with type 1 diabetes in 2021 largely depended on the geographical area, ranging from a mean of 13 years in low-income countries to 65 years in high-income countries (9). In 2040, the model predicts a 60%–107% increase in prevalent cases in comparison with 2021, with the largest relative increase expected in lower-income countries.

Significant achievements have been made toward the goal of arresting and preventing type 1 diabetes. The appreciation that autoantibodies against β-cell components are highly valuable biomarkers for disease progression led to the insight that type 1 diabetes evolves over a varying period of time, spanning from months to years. Cellular and molecular studies of the mechanisms in preclinical models and people with type 1 diabetes have led to repurposing of drugs to inhibit the autoimmune response that directly causes the disease. Preclinical animal and human islet studies have identified the pathophysiologic dialogue between the autoreactive immune system and the β-cell that has resulted in clinical testing of agents (e.g., baricitinib, verapamil) to prevent this destructive path. In parallel, great progress has been made in delivery of insulin therapy with new formulations, hybrid pumps, and continuous glucose monitors. However, these technologies still do not match the precision in metabolic control that is achievable with endogenous β-cells, and people living with type 1 diabetes are subject to the extremes of metabolic control such as hyper- and hypoglycemia, even when average glycemia is improved.

Pathophysiology of Type 1 Diabetes: An Autoimmune Disease With a Role for the β-Cell

In 1986, George S. Eisenbarth formulated a conceptual framework describing the progressive development of type 1 diabetes through various phases. According to this model, a combination of genetic predisposition and putative environmental triggers act in concert to initiate the onset of β-cell autoimmunity (11). With nearly 80 associated genetic regions, type 1 diabetes is classified as a polygenic disease, with approximately 50% of the heritability residing within the HLA class II region (12). The association with HLA class proteins, whose function is to present peptides to and activate T cells, implies a role for autoreactive T cells in the disease pathophysiology. Questions remain with respect to the environmental factor(s) (e.g., virus, dietary agents) that may contribute to the development of type 1 diabetes, either as primary triggers or factors that influence the rate of progression to overt disease (13).

The presence of autoantibodies directed against distinct β-cell antigens, including insulin, the 65-kDa form of glutamic acid GAD (GAD65), the phosphatase IA-2, and Zinc transporter 8 (ZnT8), has been used to identify individuals who will progress to clinical type 1 diabetes. A combination of the presence of these autoantibodies and metabolic testing with oral glucose tolerance tests has enabled definition of different stages of type 1 diabetes (14). The presence of two or more autoantibody types with normoglycemia is defined as stage 1 disease, whereas stage 2 type 1 diabetes is associated with dysglycemia. The clinical diagnosis of type 1 diabetes represents stage 3 (15) (Fig. 1). The rate of progression from the early, presymptomatic stages of type 1 diabetes (stages 1 and 2) to clinical stage 3 type 1 diabetes correlates with increasing numbers of antibodies, their affinity, their type or sequence of appearance, and younger ages of seroconversion (16,17). Most of these concepts are based on studies of relatives of people living with clinical type 1 diabetes, but other data indicate that the risk for progression is similar in multiple autoantibody-positive individuals in the general population without an affected relative. Thus, worldwide initiatives of screening for the presence of early, presymptomatic stages of type 1 diabetes by presence of autoantibodies are targeting family members as well as the general population (18,19).

Figure 1.

Natural history of type 1 diabetes (T1D) according to the time course of the disease. Graph shows the progressive decline of β-cell function over time. In individuals with genetic predisposition, the presence of islet autoantibody indicates the transition to the presymptomatic phase of type 1 diabetes. Loss of β-cell function past a threshold adequate for maintenance of normal glucose regulation leads to the symptomatic phase of type 1 diabetes where insulin replacement therapy is needed. Approaches to disease modification at different stages of type 1 diabetes, and stages for which clinical trials have shown efficacy in disease modification, are also indicated. Figure was created with BioRender (biorender.com).

Autoantibodies represent reliable biomarkers of initiation of an immune attack against the β-cell. However, data from the most widely studied spontaneous animal model for type 1 diabetes, the nonobese diabetic (NOD) mouse, suggest that while B lymphocytes may serve an important function as antigen-presenting cells, the antibodies that are produced by autoreactive B cells and plasma cells do not directly mediate β-cell destruction (20,21). In contrast, T lymphocytes, typically all carrying the CD3 surface marker, appear to be the direct executioners of β-cells (22,23). Observations from the analysis of pancreas specimens obtained from deceased people with type 1 diabetes through the Network for Pancreatic Organ donors with Diabetes (nPOD) (https://www.jdrfnpod.org/) program confirm the notion of T lymphocytes being most proximally responsible for β-cell killing (24,25).

Investigations in NOD mice and tissues from nPOD and other collections of human pancreases have implicated other immune cell subsets in disease (25) (Fig. 2). For example, tissue-resident macrophages or dendritic cells are among the first cells to infiltrate the islets in NOD mice, followed by adaptive immune cells (T and B cells). They can present proteins from damaged β-cells and activate T cells (26). T follicular helper (Tfh) cells enable presentation of antigen to T cells by B cells in the lymph node (27). CD4⁺ T cells, that recognize antigens shed from damaged β-cells and presented by class II HLA molecules will produce cytokines, soluble mediators that enhance the function of other T cells and also have direct damaging effects on β-cells (28). CD8⁺ T lymphocytes appear to be the direct effectors of β-cell destruction (28). They target autoantigens that can be presented by HLA class I molecules on the β-cell surface and can mediate cell killing through different mechanisms including cell-mediated cytotoxicity but also produce cytokines with direct β-cell-damaging effects. Regulatory T cells (Tregs) regulate (i.e., dampen) immune responses and ensure proper peripheral self-tolerance, overseeing the resolution of immune responses through a variety of mechanisms (29). The regulatory function of Tregs is implemented via cell-surface inhibitory molecules such as CTLA-4, production of immunosuppressive soluble cytokines like IL-10, and possibly by utilizing IL-2 that is needed for survival by effector cells. Whereas no data suggest that people with type 1 diabetes have decreased or dysfunctional Tregs, these cells and their products are interesting therapeutic targets for restoring the immune balance and arresting progression of the autoimmune β-cell destruction. Several other immune cell subsets such as natural killer (NK) cells, as well as neutrophils, have been detected within pancreatic islets and may also be involved in disease, but at present their role is unclear.

Figure 2.

The immunopathogenesis of type 1 diabetes. Antigen-presenting cells (APC), including dendritic cells and B cells, interact with CD4⁺ and CD8⁺ T lymphocytes, facilitating autoantigen presentation. This process drives the activation of β-cell-specific autoreactive T lymphocytes. Furthermore, the exposure of B lymphocytes to β-cell antigens leads to the production of autoantibodies against β-cell proteins, serving as a clinical biomarker for islet autoimmunity. Additionally, antigen-presenting cells can induce CD4⁺ Tregs, which contribute to preventing the attack of autoreactive CD8⁺ T cells on β-cells. Figure was created with BioRender (biorender.com).

A large body of work points toward a role for the β-cell itself in its own demise (recently reviewed in more detail in 25,30–33). When exposed to inflammation or other stressors (e.g., a viral infection), β-cells manifest aberrations in protein folding, autophagy, splicing, and both enzymatic and biochemical posttranslational modifications, as well as an upregulation of compensatory molecular signaling pathways such as endoplasmic reticulum stress or cellular senescence (34). These changes can exacerbate β-cell dysfunction and/or accelerate apoptosis, as well as lead to chemokine release, HLA class I overexpression, and the formation and presentation of neoantigens such as hybrid peptides or peptides with posttranslational modifications, that may promote or even expand the scope of an immune attack on β-cells. More than 60% of the type 1 diabetes candidate genes are expressed in the β-cell and activated by cytokine exposure, and potentiate further production of inflammatory cytokines and chemokines, as well as yield direct cellular toxicity (35,36). Markers of β-cell stress and dysfunction precede and predict the onset of clinical diabetes in many individuals with earlier stages of type 1 diabetes (37–39). Other islet cells, as well as islet vasculature and innervation, have also been implicated in type 1 diabetes pathophysiology. Interestingly, even the exocrine pancreas may be involved, since patients with type 1 diabetes and their relatives show reduced pancreatic volumes (25,40,41).

The Path to Disease-Modifying Therapies in Type 1 Diabetes

Most clinical trials have been carried out in people with type 1 diabetes at the time of or soon after the diagnosis with stage 3 type 1 diabetes, as they present with metabolic symptoms, in contrast to those in earlier stages—not yet insulin dependent, and for their identification autoantibody screening is needed (42). The end points of these trials have included stimulated C-peptide levels (a proxy for β-cell function), as well as HbA_1c and other glycemic control measures.

Given their central role from preclinical and other studies, the majority of therapeutic trials to date have focused on targeting immune cells with the objective of arresting the ongoing immune response leading to near complete β-cell loss (42). Evidence that interfering with T lymphocytes could alter the course of type 1 diabetes in humans originated in the 1980s, when immune suppressive agents like cyclosporin, a calcineurin inhibitor that can block cytokine production and curtail T cell activity, was shown to delay progression of β-cell destruction and even lead to clinical remission (43). Although the treatment showed some benefits in reducing exogenous insulin dependence, continuous treatment was needed and the effects were only temporary. Furthermore, treatment was associated with significant side effects, including nephrotoxicity.

Since then, other systemic immunomodulators have been tested in people recently diagnosed with stage 3 type 1 diabetes. With several, effects on β-cell preservation have not been shown, while a clear effect has been seen with others (44). For example, a short intervention with low doses of the polyclonal anti-thymocyte globulin (ATG) (doses from 0.5 to 2.5 mg/kg i.v. over 2 days) showed reproducible preservation of β-cell functional mass in young individuals in several trials (45–47). Continuous therapy with monoclonal golimumab (an anti-TNFα antibody) subcutaneously or the oral JAK inhibitor, baricitinib, slowed the decline of functional β-cell mass, as indicated by higher stimulated C-peptide levels together with lower doses of insulin needed to maintain tight glycemic control (48,49). However, with both of these agents, which target cytokines and immune signaling, efficacy waned after discontinuation. More recently, treatment with ustekinumab, which blocks p40, a component of IL-12 and IL-23 (cytokines that lead to the development of pathologic T cells), yielded improvement in stimulated C-peptide levels after 1 year of treatment (50).

In addition to agents affecting immune cells, drugs that can affect β-cells have been used in two other contemporary trials. With treatment with the calcium channel blocker verapamil, which inhibits TXNIP (an inducer of β-cell apoptosis), improvement in C-peptide was shown in pilot studies and in a large randomized trial (51,52). In a combination trial, an anti–IL-21 antibody, a cytokine produced by Tfh and other cells, was tested with or without the GLP-1 receptor agonist liraglutide, and improvement was seen in C-peptide and clinical parameters at 1 year (53).

Unfortunately, to date, regulatory approval has not been granted for any of these interventions, and a major hurdle on the path to approval is the hesitance of regulatory bodies worldwide to recognize stimulated C-peptide as a valid end point in phase 3 development. Whereas the academic world has yielded an array of data, e.g., from the islet transplantation field, on C-peptide reflecting functional β-cell mass, to date, regulators continue to demand glycemic end points impacted by multiple variables, like HbA_1c. Particularly in a world of automated insulin delivery systems, showing significant differences in glycemic end points in the first years after clinical diagnosis of type 1 diabetes has proven too high a demand for most disease-modifying therapy manufacturers, as study of very large populations (>1,000 people) would be required for an adequately powered trial.

Teplizumab: The First Approved Disease-Modifying Therapy in Type 1 Diabetes

Delaying or preventing a diagnosis of stage 3 type 1 diabetes offers an end point that has clear clinical benefit. To date, only one intervention, the humanized anti-CD3ε monoclonal antibody (mAb) teplizumab, has achieved this goal. Teplizumab was approved as the first disease-modifying therapy by the U.S. Food and Drug Administration in 2022 and more recently in several countries worldwide.

History and Clinical Results

Teplizumab, originally termed hOKT3γ1(Ala-Ala), is a humanized version of the first murine mAb, OKT3, with two point mutations in the Fc portion of the Ig molecule to prevent Fc receptor binding (54). OKT3 was used to treat allograft rejection in kidney transplant patients, but there was significant toxicity because of cytokine release syndrome, which was thought to be due to cross-linking of the bivalent mAb between the FcR and CD3 on T cells. To address these issues, the molecule was modified. Another humanized nonmitogenic variant, known as ChAglyCD3 or otelixizumab, was developed through introduction of a single mutation in the Fc portion, which prevents glycosylation and inhibits Fc receptor binding.

The availability of safe humanized FcR nonbinding anti-CD3 monoclonal antibodies paired with preclinical studies suggesting efficacy led to clinical trials exploring these immunotherapeutic strategies in type 1 diabetes (54,55). The humanized anti-CD3ε mAb teplizumab was studied in several investigator-initiated clinical trials and 2 phase 3 trials.

In 2002, Herold et al. (56) conducted the first clinical trial, known as Teplizumab Study 1, and investigated teplizumab’s efficacy in individuals with newly diagnosed stage 3 type 1 diabetes. The results revealed improved C-peptide response with reduced HbA_1c levels and insulin requirements for at least 2 years. Subsequently, other clinical trials with otelixizumab or teplizumab have been carried out (summarized in Table 1).

Table 1.

Overview of completed and ongoing clinical trials with humanized anti-CD3 monoclonal antibodies in type 1 diabetes

Study	Drug, dose	Trial identifier,* status	Inclusion criteria	Clinical outcome	Phase
Teplizumab Study 1 (56)	Teplizumab, 23 mg	n/a, completed	Stage 3; <6 weeks diagnosed, age 7–30 years, stimulated C-peptide >0.2 nmol/L	Better C-peptide response at 1 and 2 years, less insulin requirement	2a
Treatment With hOKT3γ1(Ala-Ala) in T1DM (DAIT identifier ITN007AI [NDB01]) (80)	Teplizumab, 38 mg	NCT00806572, terminated (adverse events related to drug lot)	Stage 3; <6 weeks diagnosed, age 7–30 years, stimulated C-peptide >0.2 nmol/L	Better C-peptide responses, less insulin requirement	2
AbATE study (DAIT identifier ITN027AI) (59)	Teplizumab, 17 mg × 2 at entry and month 12	NCT00129259, completed	Stage 3; <8 weeks diagnosed, age 8–30 years, autoantibody positive	Better C-peptide response at 2 years, less insulin requirement	2
Delay (61)	Teplizumab, 17 mg	NCT00378508, completed	Stage 3; 4–12 mos after diagnosis, age 8–30 years	Better C-peptide response at 1 year, less insulin requirement	2
Protégé study (62)	Teplizumab, 2.4, 2.9, or 9.0 µg/m2 at entry and month 12	NCT00385697, completed	Stage 3; <12 weeks from diagnosis, age 8–35 years	Better C-peptide response at 1 and 2 years, less insulin requirement in subgroups	3
Protégé Encore study (74,81)	Teplizumab, daily intravenous dosing for 14 days at entry and 6 mos	NCT00920582, completed	Stage 3; <12 weeks diagnosed, age 8–35 years	Better C-peptide; discontinued prematurely	3
TN10 (64)	Teplizumab, 14 days of daily intravenous dosing at entry	NCT01030861, completed	Stage 2; age 8–45 years, relatives of individuals with T1D	Delay in diagnosis of stage 3 T1D, improved C-peptide response	2
PROTECT (63)	Teplizumab, intravenous dosing for 12 days at entry and 6 mos	NCT03875729, completed	Stage 3; <12 weeks diagnosed, age 8–17 years	Better C-peptide response, less insulin requirement	3
European otelixizumab trial (57)	Otelixizumab, 48 mg	EudraCT 2006-00357911, completed	Stage 3; <4 weeks diagnosed, age 12–39 years, autoantibody positive	Better C-peptide response, less insulin requirement in subgroups	2
DEFEND-1 (58)	Otelixizumab, 3.1 mg	NCT00678886, completed	Stage 3; <12 weeks diagnosed, age 12–45 years, autoantibody positive, stimulated C-peptide >0.2 nmol/L	No improvement in C-peptide response or insulin use	3
DEFEND-2 (82)	Otelixizumab, 3.1 mg	NCT01123083, completed	Stage 3; <12 weeks diagnosed, age 12–17 years, autoantibody positive, stimulated C-peptide >0.2 nmol/L	No improvement in C-peptide response or insulin use	3
Belgian Diabetes Registry study (83)	Otelixizumab, cumulative dose 9, 18, 27, or 36 mg i.v. over 6 days	NCT02000817, completed	Stage 3; <6 weeks diagnosed, age 16–27 years, autoantibody positive, stimulated C-peptide >0.2 nmol/L	No improvement in C-peptide response or insulin use	1, 2a
TN10 Extension	Teplizumab, intravenous dosing daily for 12 days	NCT04270942, completed	Stage 3; prior TN10 participants	No data	2
PETITE-T1D	Teplizumab, 14 days of daily intravenous dosing at entry	NCT05757713, ongoing—fully recruited	Stage 2; age <8 years	No data	4

n/a, not available; DAIT, National Institute of Allergy and Infectious Diseases Division of Allergy, Immunology, and Transplantation; mos, months; EudraCT, European Union Drug Regulating Authorities Clinical Trials Database; PETITE-T1D, Teplizumab in Pediatric Stage 2 Type 1 Diabetes; T1D, type 1 diabetes; TN10, Anti-CD3 Mab (Teplizumab) for Prevention of Diabetes in Relatives At-Risk for Type 1 Diabetes Mellitus.

*Trial identifiers are clinical trial reg. nos., ClinicalTrials.gov, unless otherwise indicated. Adapted from Daifotis et al. (81).

In individuals with recent-onset stage 3 type 1 diabetes, a short course of otelixizumab increased C-peptide levels at 6 months (57). However, there were safety concerns in this study, one of which was Epstein-Barr virus (EBV) reactivation in 75% of patients. To eliminate this, investigators for a phase 3 trial, Durable Response Therapy Evaluation for Early or New-Onset Type 1 Diabetes (DEFEND), dramatically reduced the dose of otelixizumab. Unfortunately, with the markedly reduced dose, the trial did not meet its end point and the agent did not progress to confirmatory studies (58).

Based on the success of “Study 1,” a further study (Autoimmunity-blocking Antibody for Tolerance in Recently Diagnosed Type 1 Diabetes [AbATE)]) was conducted to evaluate teplizumab’s efficacy in people with stage 3 type 1 diabetes given two courses—at diagnosis and after 1 year (59). In AbATE, 2-year C-peptide levels were significantly better and insulin requirements and HbA_1c levels were reduced in treated versus control study participants. For determination of whether teplizumab would show efficacy after the new-onset period (i.e., >100 days following diagnosis), in the Delay study the drug was tested in people who had been diagnosed in the past 4–12 months and still maintained clinically significant β-cell function, i.e., a stimulated C-peptide level of at least 0.2 pmol/mL (60). As in people with new-onset stage 3 type 1 diabetes, teplizumab yielded C-peptide preservation, but with less effect on the other clinical outcomes, possibly related to an imbalance in the HbA_1c levels in the study arms at entry (61).

In the phase 3 Protégé trial three dosing regimens of teplizumab were evaluated (14 days with full dose, 14 days low dose, and 6 days full dose) in 8- to 35-year-old individuals with new-onset stage 3 type 1 diabetes (62). The study was conducted in the U.S., Eastern Europe, and India. However, in the study a composite measure was used as the primary end point (i.e., the frequency of patients with HbA_1c <6.5% who were using <0.5 units/kg/day insulin), with failure to meet the primary end point at 12 months. There were significant geographic differences in participant C-peptide levels, ICA512/IA-2 antibody positivity, and HbA_1c levels. The post hoc analysis showed that C-peptide levels were significantly better than with placebo at 12 and 24 months in participants receiving the full 14-day drug course. At year two, there continued to be improvement in C-peptide responses (61). Interestingly, the treatment effects were more pronounced in younger (8–11 years) individuals, those recruited from the U.S., and those who received treatment within 6 weeks after clinical diagnosis, suggesting greater benefit in patients who still have sufficient β-cell mass at the time of therapy.

Subsequently, in another phase 3 trial, Provention Bio’s Type 1 Diabetes Trial Evaluating C-Peptide with Teplizumab (PROTECT), conducted between 2019 and 2023, individuals with new-onset stage 3 type 1 diabetes aged 8–18 years were enrolled, with a primary end point of higher 78-week stimulated C-peptide levels in comparison with placebo (63). C-peptide levels were once again significantly improved in drug-treated versus placebo-administered participants. Investigators were instructed to treat study participant glucose levels to American Diabetes Association target HbA_1c levels (HbA_1c <7%), and insulin doses needed to achieve this criterion were lower in the drug-treated group throughout the study. A noninterventional extension study is ongoing (clinical trial reg. no. NCT04598893, ClinicalTrials.gov).

The major breakthrough in the field came as methods for identifying asymptomatic people in early-stage type 1 diabetes were advancing. The data arising from studies in people with stage 3 type 1 diabetes treated with teplizumab raised the question as to whether the intervention would delay or prevent progression to clinical, stage 3 type 1 diabetes. Groups around the world showed that in relatives who had antibodies to two autoantigens along with early signs of dysglycemia (i.e., stage 2 type 1 diabetes), the risk for progression to clinical disease in 5 years was ∼75% (14). With this understanding the Type 1 Diabetes TrialNet consortium (https://www.trialnet.org/) conducted a study to test whether a single course of teplizumab would delay or prevent clinical diagnosis of stage 3 type 1 diabetes in individuals with stage 2 disease (64). The study showed that a single 14-day treatment delayed the median time to clinical diagnosis by close to 3 years (65). In a follow-up analysis, 28 of 44 drug-treated individuals only progressed to clinical stage 3 type 1 diabetes after 5 years and 12 drug-treated individuals had not progressed to stage 3 type 1 diabetes 5 years after therapy (66). These observations on delay in progression to stage 3 type 1 diabetes, together with data from 375 treated individuals in five clinical trials with 1,500 patient-years of follow-up, were in support of the safety of the drug and the consistent improvement in C-peptide with reduced insulin requirements (63).

Mechanism of Action and Safety of Teplizumab: What Clinicians Should Know

Anti-CD3 antibody was shown to reduce hyperglycemia in mice with nonautoimmune diabetes administered low-dose streptozotocin and more extensively in the NOD mouse model, where there are many similarities to human disease (54). Interesting data on timing and dosing and the importance of the type of antibody (Fc receptor binding vs. nonbinding) were also identified. These data suggest that lower doses and interventions closer to the moment of hyperglycemia (in stages where autoimmune destruction of β-cells is already ongoing) are more effective than interventions with higher doses or interventions in younger mice before an active autoimmune response is present (55,67). Studies in the NOD mouse have also pointed to clues as to the mechanism of action of these anti-CD3 antibodies.

As mentioned above, teplizumab specifically binds to the ε-chain of the CD3 complex found on the surface of CD4⁺ and CD8⁺ T cells, transmitting signals to the cells causing transient activation and internalization of the receptor complex (64,68,69). The strength of the signal that is delivered via the CD3 complex is an important determinant of the fate of the T lymphocytes. With the first few doses of teplizumab, there is a transient reduction in the circulating lymphocytes that resolves spontaneously despite the continued course dosing. The reduced circulating lymphocyte levels are not likely to be due to depletion of T lymphocytes given that they resolve spontaneously and rapidly with continued drug administration. Rather, preclinical data indicate that the circulating lymphocyte reduction reflects egress from the circulation to the gut wall. Consistent with this, in a preclinical model there is a link of drug efficacy to the microbiome (70–72). The absence of true lymphocyte depletion is clinically relevant, since this mechanism of action differs from depleting monoclonal antibodies, such as CAMPATH 1H (CD52), which cause profound, prolonged depletion. Indeed, the transient reactivation of EBV that occurs in some EBV-seropositive patients treated with teplizumab resolves without the need for antiviral treatments and the clinical experience has not suggested an increased risk of infections, even when teplizumab was given during the 2019 coronavirus disease pandemic (63).

Of interest, the partial agonist signal induces T lymphocytes with regulatory properties and reduces the functions of others. In preclinical studies, anti-CD3 antibodies induced Foxp3⁺ TGFβ-producing Tregs. Also, in effector T lymphocytes, the activation signal that is delivered by binding to the CD3 complex is followed by a state of cellular exhaustion, a dysfunctional state evidenced by reduced levels of cytokines and changes in ligands needed for cell maintenance and growth, like the IL-7 receptor (65,66,68,73). Teplizumab-treated people with the lowest levels of IL-7 receptor expression after treatment have the most robust responses to the drug. Possibly as a result of these effects of teplizumab, the typical expansion of autoantigen-reactive CD8⁺ T cells that occurs in people with early-stage preclinical type 1 diabetes is prevented (65) (Fig. 3). Interestingly, changes induced by the 2-week treatment with teplizumab are found 18 months after treatment and may account for the prolonged effects of treatment (>10 years) in some individuals. However, not all cells are affected equally, since autoantigen-reactive T cells are reduced in frequency but immune responses to viral pathogens are maintained. Collectively, these and other data indicate that the drug mechanism does not involve prolonged immune suppression but, rather, is more specific in its effects, without requiring continuous treatment.

Figure 3.

Hypothesized mechanism of action of teplizumab. Illustration of the proposed mechanism underlying the tolerogenic effect of teplizumab. The binding of anti-CD3 mAb to the CD3/TCR complex activates the complex, modulating expression of CD3 but also delivering a partial agonist signal to T cells. This partial activation signal renders T cells unresponsive to their cognate antigen (i.e., partial exhaustion). In addition, the signal may include a regulatory phenotype and reduce the expression of the receptor for IL-7, a cytokine that is involved in CD8⁺ T cell maintenance and expansion. The effects of the drug are selective, since CD8⁺ T cells that are responsible for virus recognition recover function, while those that recognize autoantigens, and are thought to be involved in the pathogenesis, may be affected for extended periods of time without continuous drug administration. Figure was created with BioRender (biorender.com).

The safety of teplizumab has been evaluated in >700 people with various stages of type 1 diabetes, age 8–45 years (74). Serious adverse events were reported in 12.4% and 8.2% of people in the teplizumab and control groups, respectively, and involved gastroenteritis, cellulitis, pneumonia, abscess, and sepsis, but most (76.7%) were considered unrelated to drug treatment. Table 2 describes the most common adverse events thought to be drug related and reported during teplizumab trials in type 1 diabetes. Apart from the mechanism-based transient lymphopenia, the most common side effects that differ in frequency in comparison with control patients are rash (in ∼35% of people), transient increases in liver function tests, and low-grade cytokine release with the initial doses of the drug. This has been treated with ibuprofen, antihistamines, and acetaminophen. Infection rates were similar between teplizumab-treated and control participants. As noted, transient EBV and cytomegalovirus reactivation is more frequent following teplizumab treatment than with control but resolves without the need for antiviral treatments. An observational patient registry has been created to address long-latency risk (clinical trial reg. no. NCT06481904, ClinicalTrials.gov), and this long-term follow-up will be critical for continued assessment of risk versus benefits as increasing numbers of people are treated (75). Because of the immune modulatory effects, the timing of vaccinations needs to be addressed prior to administration of teplizumab, with avoidance of administration of inactivated vaccines in the immediate posttherapy period (6 weeks). Also, administration of live vaccines is not recommended during teplizumab treatment or during the first year after therapy.

Table 2.

Most common drug-related adverse events during teplizumab trials in type 1 diabetes

	Description	Treatment
During administration
Transient lymphopenia	The most common side effect related to teplizumab treatment. The lymphopenia is mechanism based. Average lymphocyte count reaches nadir at day 5 of treatment and returns to baseline levels by week 6	None
Low-grade cytokine release syndrome	Includes rash, headache, nausea, vomiting, and rigors	Ibuprofen, antihistamines, and acetaminophen
Fever	Mostly low grade	Ibuprofen and/or acetaminophen
Rash	Torso, hands, and limbs. Often appears at the end of the first week of infusions.	Antihistamines
Transient increases in liver function tests	May occur from cytokine release. Not common and generally mild	None
Early weeks of follow-up
Rash	Torso, hands, and limbs	Antihistamines
Reactivation of EBV	May be viremia only or symptomatic (viral syndrome)	Symptomatic: clearance occurs spontaneously
Reactivation of CMV	May be viremia or symptomatic (viral syndrome)	Symptomatic: clearance occurs spontaneously

Data are derived from Herold et al. (74). CMV, cytomegalovirus.

There are features of teplizumab that limit its acceptability. The drug is administered intravenously daily for 14 consecutive days at doses of 65 µg/m² on day 1, 125 µg/m² day 2, 250 µg/m² day 3, 500 µg/m² day 4, and 1,030 µg/m² days 5–14 (75,76). In addition, the cost of teplizumab (TZIELD) is very high—greater than that for most biologics. The list price in the U.S. can exceed $200,000. Therefore, insurance assistance is needed for most and out-of-pocket costs vary, especially considering the additional costs associated with laboratories, procurement of longer-term intravenous access, infusion staffing, work leave, and travel and lodging that may be required for the 14-day infusion.

Summary and Future Directions

A growing understanding of type 1 diabetes pathophysiology, as well as the arrival of biomarkers to identify people with type 1 diabetes early in the disease process, at stages before symptoms are present, is opening a path to delaying the progression to clinical type 1 diabetes with disease-modifying therapies. In people with stage 2 or stage 3 type 1 diabetes clinical trials have demonstrated the ability of teplizumab to attenuate disease progression—and, of clinical significance, without prolonged immune suppression. As a result, teplizumab was the first drug approved to delay the development of clinical type 1 diabetes.

However, important questions remain, one of which is whether the drug efficacy is similar across ages. A study is ongoing to look at safety in those <8 years of age, and additional studies will be needed to assess efficacy in older individuals. Furthermore, an unresolved issue is whether baseline differences affect responses, e.g., differences in HLA haplotypes or autoantibodies or other personal factors such as the recipient’s microbiome or previous virus exposures. Also, a direct comparison of one versus two courses of treatment has not been undertaken, so the need for repeated dosing and the potential of a combination of treatments are unexplored areas (77,78). The use of biomarkers to predict or track efficacy of the drug or identification of shared demographic features of those individuals most likely to respond (such as age or rate of decline of C-peptide) may enable precision medicine approaches to treatment. Finally, the full therapeutic potential of modifying disease progression with teplizumab is still unfolding, and more research is needed for understanding long-term benefits. Given that teplizumab is the first drug approved to delay any autoimmune disease, lessons learned may open the door to prevention of other autoimmune conditions through treating the underlying mechanisms before the full onset of clinical disease and before disease-managing therapies, like insulin replacement, are all that can be offered.

Funding Statement

C.M. has served as the vice president and president of the European Association for the Study of Diabetes. E.A.J. receives grant support from Breakthrough T1D, the Marco J. Heidner Charitable Trust, the Cystic Fibrosis Foundation, the National Institute of Diabetes and Digestive and Kidney Diseases/National Institutes of Health (NIH), and the National Institute of Allergy and Infectious Diseases/NIH. E.K.S. has received reimbursement and travel support for lectures from the American Diabetes Association, serves as treasurer of the Immunology of Diabetes Society and an Associate Editor for Diabetes Care, and has grant funding from the NIH, Breakthrough T1D, and The Leona M. and Harry B. Helmsley Charitable Trust. M.A.A. receives grant support from various federal and not-for-profit agencies, has received support for travel from various federal (NIH) or not-for-profit agencies (e.g., Breakthrough T1D, World Diabetes Foundation), and has a leadership role for the American Diabetes Association Scientific Sessions and as a journal editor at Diabetes Care. K.H. has received research support from NIH and Breakthrough T1D. No funding or payment was available to support the writing of this article.