Breakdown and Repair of Peripheral Immune Tolerance in Type 1 Diabetes

Abstract

Failures in peripheral immune tolerance mechanisms create a permissive environment for autoimmune diabetes initiation and disease progression. Biomarker analyses provide tools that allow recognition of this loss of tolerance, reflecting a serial acquisition of pathogenic characteristics causally linked to islet β-cell dysfunction and death. Autoimmune effector cell activation and expansion, ineffective immune regulation, and tissue response to injury during active disease each represent challenges to homeostasis; however, they also represent targets for therapeutic intervention, with the potential for restoration of tolerance. Limited success in recent clinical trials demonstrates that tolerance in type 1 diabetes (T1D) is achievable, but currently occurs in few subjects and is not durable in most. Combining therapeutic agents to rebuild multiple immune components to restore tolerance, particularly addressing both effector and regulatory T-cell dysfunction, is needed.

Perspectives on immune tolerance in type 1 diabetes (T1D) reflect our understanding of key checkpoints in pathogenesis that determine susceptibility, disease progression, and resistance to therapy. At each of these three steps, immunological outcomes either promote immune activation or immune regulation, and the balance defines whether tolerance prevails. By viewing these checkpoints from the perspective of genetic enablers (susceptibility), environmental triggers (progression), and homeostatic controllers (including response to therapy), a framework emerges that provides a useful paradigm for aligning stages of T1D pathogenesis with prospects for restoration of homeostasis and immune tolerance.

SUSCEPTIBILITY

The principal enablers of T1D are genetic traits that generate a potential for aberrant autoreactivity to islet autoantigens. Most of the genes known to be associated with T1D are involved in immune responses (Concannon et al. 2009), and the two most important of these genes, HLA and INS, are directly mechanistically linked to establishing a peripheral immune repertoire with potential autoreactivity: The HLA genetic association is attributable to specific HLA class II genes, notably HLA-DQB1*0302, which acts as a restriction element for autoreactive CD4 T-cell recognition through presentation of islet-derived peptides binding to either cis- (DQA1*0301/DQB1*0302) or trans- (DQA1*0201/DQB1*0302) class II HLA-DQ8 molecules encoded by particular HLA-DR4 or HLA-DR3/4 haplotypes (Nepom and Erlich 1991; Nepom and Kwok 1998; Chow et al. 2019). The INS genetic association with T1D is attributable to allelic variation in the INS promoter region, which corresponds to differential expression of proinsulin in thymic tissue (Pugliese et al. 1997), and this correlates with the frequency of peripheral insulin-specific CD4 T cells, in which lower thymic expression of INS corresponds to higher autoreactive potential (Durinovic-Belló et al. 2010), presumably due to inefficient negative selection during T-cell maturation.

Notably, neither of these genetic traits, or even the combination of the two, is fully penetrant for T1D clinical outcome. Nondiabetic individuals with these same genes also have a peripheral repertoire populated by autoreactive CD4 T cells with the same specificity, indicating that while these genes confer the potential for disease, they are not sufficient. In this sense, the susceptibility genes are mechanistically responsible for creating a potentially autoreactive host environment, fertile ground for subsequent immune activation events via environmental triggers that enable disease progression. Precedent for this type of environmental interaction with genetic susceptibility is found in other immunologic diseases, which show the types of triggering events that can initiate a break in tolerance. For example, in rheumatoid arthritis, the combination of a history of smoking in individuals carrying the HLA-DR0401 susceptibility genes synergizes for the development of anti-ACPA autoantibodies, a precursor to progressive autoimmune rheumatic disease, apparently due to the up-regulation of PAD enzymes in the lung from chronic smoke irritation (Klareskog et al. 2011; Gravallese and Firestein 2023). Another example is seen in the food allergy to peanuts, in which a specific HLA-DQA allele and a genetic variant of MALT1 are associated with sensitization and progression to allergy, respectively (Huffaker et al. 2023). Intriguingly, in this latter example, the same HLA association modifies the likelihood that oral antigen immunotherapy will be successful, consistent with the notion that a very specific antigen–HLA interaction leads to antigen presentation for immune responses that can be pathogenic or tolerogenic, depending on clinical context (Kanchan et al. 2022).

At this initial stage of disease susceptibility, predisposition is hard-wired into an individual's background, and is challenging to target for tolerogenic therapies. In the NOD mouse, which displays a class II molecule (IAg7) structurally similar to DQ8, the introduction of a modified class II transgene (Singer et al. 1998) or selective breeding with nonsusceptible haplotypes in heterozygous animals (Wicker et al. 1995) is sufficient to prevent spontaneous progression to diabetes. HLA-DQ0602 may have a similar role in protection from T1D in humans, since it is negatively associated with disease incidence, particularly in children (Kockum et al. 1995). Distinct physicochemical properties of the class II molecule encoded by this protective allele have been described that suggest a potential role in enhanced negative selection of T cells (Ettinger et al. 2000). However, feasible ways to manipulate the islet peptide–HLA interaction to recapitulate a tolerance outcome that alters the frequency or fate of the peripheral autoreactive CD4T cells are daunting and will require novel forms of intervention. Potential examples include attempts to interfere with the DQ8–peptide interface (Ostrov et al. 2018) or to enhance thymic negative selection directly via specific antigen exposure (Atibalentja et al. 2009), but there are relatively few therapeutic efforts to tolerize at this very early step in pathogenesis compared to a multitude of therapeutic initiatives aimed at the next step, when activation, expansion, and regulation of the immune response coalesce in a progressive autoreactivity that results in clinically evident disease.

PROGRESSION

Since autoreactive T cells with islet specificity are found in all genetically susceptible individuals, what are the characteristics that represent a loss of tolerance leading to disease progression in some, but not all? Which of these offer opportunities or specific molecular targets for tolerogenic therapy?

One fundamental observation was the finding that the age of onset and rate of progression to T1D can be markedly different among identical twins (Redondo et al. 2008). This simple but significant concept indicates that control of disease progression is a potent homeostatic system and provides a rationale for comparisons between T1D and genetically matched non-T1D subjects to identify indicators and potential control points for tolerance. These types of studies identify numerous T-cell characteristics, which correspond to a loss of tolerance, such as the acquisition of markers indicating in vivo activation (e.g., CD38; Yang et al. 2017) or chronic activation (e.g., Kv1.3; Beeton et al. 2006), an apparent deficit in functional regulatory T cells (Long et al. 2010; Hull et al. 2017) or the inability of effector T cells to respond to regulation (Schneider et al. 2008), expansions in potentially inflammatory subsets (Arif et al. 2017), oligoclonal expansions of autoreactive cells (Cerosaletti et al. 2017), or the prevalence of stem cell memory T cells (Abdelsamed et al. 2020). Studies focused on a similar stem-like CD8T cell in the NOD mouse model of autoimmune diabetes have identified a distinct TCF1^hi phenotype with a key role as a self-renewing progenitor cell for generating and replenishing effector autoimmune T cells responsible for β-cell destruction (Gearty et al. 2022).

Each of these T-cell characteristics represent properties of autoreactive immune activity that occur during loss of tolerance, and therefore they are potentially useful biomarkers for monitoring disease status. In other cases, it is the loss of a T-cell phenotype, such as T-cell exhaustion, that indicates an effector response associated with more rapid disease progression (Wiedeman et al. 2020). This constellation of autoimmune characteristics presents both an opportunity and a challenge. The opportunity is to use the properties of the acquired marker(s) as therapeutic targets, for example using depletional agents directed to acquired activation markers, providing support for Tregs through cell therapy or IL-2 administration, enhancing T-cell exhaustion (e.g., with teplizumab, a Mab against CD3; Long et al. 2016), deviating immune programming away from Th17 using anti-IL-6 or anti-IL-6R, etc. As discussed in the next section, many of these therapeutic strategies have been advanced into clinical trials, but successful tolerance has been elusive, with modest success in a limited number of individuals. Each of these features of disease progression is likely a downstream consequence of autoreactive activation, making successful tolerance somewhat refractory to single drug strategies without addressing root causes or addressing multiple targets.

There have been many analyses that use epidemiologic or observational investigational tools to study the apparent breakdown of tolerance in early T1D. Particularly notable are longitudinal cohort studies monitoring genetically susceptible children who acquire evidence of immune activation usually documented by the appearance of islet autoantibodies (Ziegler et al. 2003; Lynch et al. 2018; Vatanen et al. 2018; Krischer et al. 2022). One of the recurring observations is the association of various enteroviral or rotaviral infections with disease progression (Hyöty et al. 1995; Honeyman et al. 2010; Honkanen et al. 2017). While often interpreted as evidence for molecular mimicry between viral and islet antigens, alternative explanations such as generalized inflammation from tissue cell death (Horwitz et al. 1998), indirect effects mediated through altered microbiome, or interference with systemic regulatory responses are also reasonable possibilities. Another type of induced breakdown of tolerance leading to autoimmune diabetes is seen occasionally in cancer patients treated with immune checkpoint inhibitors (ICIs), which directly demonstrates a role for T-cell costimulation checkpoints in maintaining homeostasis, although the comparability of ICI-induced diabetes and T1D has been questioned (Perdigoto et al. 2022; Quandt et al. 2024).

The time between the initial production of autoantibodies against islet antigens and the onset of dysglycemia is highly variable, a window of opportunity not only for intervention to restore tolerance but also a time frame in which disease amplifiers may arise that escape tolerogenic mechanisms. In this context, T-cell recognition of posttranslationally modified islet proteins (PTM antigens) is particularly interesting. Proposed to be products of transpeptidation reactions in islet granules (Wang et al. 2019; Reed and Kappler 2021) or defective ribosomal translation (Kracht et al. 2017) or enzymatic modification (McGinty et al. 2014), the presence of activation and memory properties of peripheral T cells reactive to PTM antigens (Nguyen et al. 2023) indicates an intriguing amplification mechanism associated with disease progression consistent with the hypothesis that a lack of thymic negative selection has led to a peripheral repertoire biased toward de novo activation, moving the immune response in a pathogenic direction. Whether these PTM modifications actually initiate disease is not clear, but since islet stress appears to foster transpeptidation and enzymatic PTM, it is plausible that a neoantigenic T-cell response to PTM antigens acts as an important amplifier of anti-islet immunity that can potentially bypass and overcome underlying tolerance mechanisms, thereby contributing to disease progression.

Direct comparisons between these immune parameters associated with T1D progression in humans with the disease course in murine models of autoimmune diabetes are rare, in part due to the differences between cells recovered from peripheral blood in humans versus lymph nodes and spleen in rodents. In addition, major differences (von Herrath and Nepom 2009) between the most common murine model (NOD strain of mice) and human pathophysiology, such as a strong gender effect on penetrance in NOD but not humans and accumulation of peri-insulitic infiltrates of lymphocytes immediately before rapid destruction of islets in NOD but not seen in the nPOD catalog of human pancreatic pathology specimens, are two such examples. Thus, although there may be a few cases of human T1D that are pathophysiologically similar to the NOD mouse, they are certainly not common, reflected in the large list of therapeutic interventions that interfere with disease progression in the NOD but have failed in human trials, including heat shock proteins, various insulin-related antigens, immune adjuvants, and others (Shoda et al. 2005). In spite of these differences, some general concepts from mouse models regarding mechanisms of disease remain useful for translational applications. One such concept is the importance of genetic background for influencing tolerance. Although the disease in the NOD mouse is highly penetrant, extensive backcrossing onto other strains has identified multiple disease-associated genetic loci (Wicker et al. 2005), most of which provide relative protection against spontaneous diabetes. Examples that have translated well to humans include IL-2 and CTLA4, where relevant pathways were implicated in murine models and conform to regulatory pathways in human immunobiology that influence the progression of T1D. In this way, efforts to understand disease protective mechanisms in mice are candidates for investigation in humans. Another concept that translates well between murine and human diabetes is the generation of PTM antigens in the islet. As discussed above, multiple types of PTM are associated with T1D progression, and this understanding developed in human diabetes as a direct consequence of the discovery of transpeptidase-generated “hybrid antigens” in the BDC2.4 model of murine NOD diabetes (Baker et al. 2019).

Perhaps the highest utility of murine models is in generating concepts that, although differing in detail from human immunology, nevertheless suggest novel pathways and strategies with tolerogenic potential. Anti-CD3 monoclonal antibody therapy for T1D was initially developed based on efficacy in the NOD mouse (Herold et al. 1992; Chatenoud et al. 1994), but the resulting high degree of durable efficacy is not fully recapitulated in humans, where the response is more heterogeneous, and a minority of treated subjects show long-term clinical remission (Perdigoto et al. 2019). Another example is the report that genetic deletion of regulatory T cells in mice is sufficient to allow spontaneous polyclonal self-reactive follicular T cells to expand (Lee et al. 2023), reiterating the importance of FOXP3-positive regulatory T cells to control autoimmunity (Spence et al. 2018; Sakaguchi et al. 2023) and lending support to efforts to increase robust regulatory T cells as a tolerogenic, or at least homeostatic maintenance, therapy in humans (Dong et al. 2021). Another example that led to a direct translational strategy was the demonstration in the NOD mouse that the therapeutic benefit of anti-CD3 given in a suboptimal dose could be enhanced by coadministration of an oral bacterial carrier encoding IL-10 alongside proinsulin (Cook et al. 2020), an experiment that was recapitulated in a phase 1 clinical trial (Mathieu et al. 2024).

Two of the important features of T1D disease progression that are difficult to recapitulate in murine models are the complexity of an outbred population and diverse environmental elements including microbiome, diet, and infections that influence immunologic development and maturation. An overly literal translation of murine models, particularly the NOD mouse, can lead to viewing T1D as a single disease pathway, overlooking the heterogeneous nature of immune responses that occur in people. In this regard, lessons can be learned from other autoimmune diseases, such as systemic lupus erythematosus and rheumatoid arthritis, which differ from T1D in many respects but which show how diverse immune responses and environmental interactions can converge to manifest as particular diseases, and in the process identify therapeutic targets. For example, a general theme in lupus pathogenesis is an inappropriate response to intra- or extracellular nucleic acids that trigger intracellular signals and sensors leading to exuberant immune activation, which, when accompanied by a failure of regulation, leads to progressive immune-mediated pathology (Accapezzato et al. 2023). In rheumatoid arthritis, an important causal pathway appears to be an inappropriate response to pathogens that elicit neutrophil TRAPs involving citrullination defense mechanisms that, when accompanied by a failure of regulation, lead to immune responses to PTM antigens and progressive pathology (Gravallese and Firestein 2023). The analogous paradigm in T1D allows for diverse triggering events from environmental stimuli in the context of genetic susceptibility, which elicit exuberant immune activation when there is a failure of regulation, a simple recipe for loss of tolerance and immunopathology that converges in a final common pathway of islet cell destruction. From this perspective, peripheral tolerance mechanisms that protect from tissue damage represent the last bulwark against disease progression, in which therapeutic restoration of tolerance requires not only boosting these regulatory elements but also negating the exuberant autoimmune responses, in parallel.

RESTORATION OF HOMEOSTASIS

Does the loss of tolerance in T1D indicate defects in immune regulation or defective overactivation? As summarized above, biomarker, phenotypic, and functional studies indicate that both are closely linked to disease progression. Indeed, a large body of evidence from clinical trials and observational studies indicates that therapeutic interventions that address only one but not the other are insufficient to restore tolerance (Huffaker et al. 2021; Nepom 2022). One of the notable features of immunomodulatory clinical trials in T1D is that multiple therapeutic strategies have shown similar transient clinical outcomes. For example, partial preservation of C-peptide secretion occurs following various anti-T-cell (e.g., teplizumab, alefacept, ATG), anti-B-cell (e.g., rituximab), anti-cytokine (e.g., golimumab), or costimulatory blockade (e.g., abatacept) therapies (Herold et al. 2013; Orban et al. 2014; Pescovitz et al. 2014; Rigby et al. 2015, 2023; Haller et al. 2018). Relapse, as measured by the rate of C-peptide loss, occurs in most subjects for each of these agents. In the clinical trials of teplizumab and alefacept, however, between 10% and 30% of treated subjects displayed a more durable response, and focusing on the immune system characteristics in those individuals provides additional lessons for understanding peripheral tolerance.

A key observation is that therapies directed against T cells, unless they spare the regulatory T-cell compartment, are insufficient for tolerance when therapy is discontinued. This has actually been appreciated since the initial studies of cyclosporine as a potential therapy in T1D, which showed durable remission in a minority of treated subjects (Papoz et al. 1990) and is reinforced by the description of recurrent T1D in subjects who received pancreatic transplants, in spite of receiving potent anti-rejection immunosuppression (Burke et al. 2015). Both teplizumab and alefacept are more selective with respect to T-cell inhibition, and although each was initially developed with depletional or antagonistic activity expected, it is now recognized that their unanticipated agonist activity-inducing exhaustion-like properties in effector T-cell populations are the dominant phenotypic correlate of clinical response (Linsley and Long 2019; Diggins et al. 2021). Importantly, neither of these drugs reduce regulatory T-cell numbers or function, and thus they somewhat fortuitously impact both effector and regulatory arms of the immune system in a favorable direction.

Although a tolerogenic strategy of ablating, deviating, or anergizing the effector response while at the same time boosting regulatory responses seems like a simple equation, there are significant challenges ahead. One issue is recognition of the importance of heterogeneity in the early stages of the pathogenic sequence leading to T1D. As discussed above, it is likely that multiple pathways are variably activated in the progression to a pathogenic state of immune activation, raising the possibility that the type of tolerogenic therapy may need to be tailored for targeting particular pathways in different individuals (Linsley et al. 2021). The concept of disease “endotypes” is intended in part to address this issue (Battaglia et al. 2020), and may help explain why fairly small numbers of patients respond to different particular types of immune therapies such as anti-cytokines, costimulatory blockade, B-cell depletion, etc. A second issue is concern about the durability of regulatory responses that are based on immunologic plasticity—the concept that the immune response is well adapted to change phenotype in response to external cues. For example, the demonstration that IL-6 exposure strongly influences the development of Th17 cells versus Treg cells suggested that tissue inflammation could be a strong barrier to enhancing regulation (Bettelli et al. 2008), a concept that was validated in preclinical model systems documenting the evolution of “ex-Treg” cells from a regulatory to a potentially pathogenic phenotype (Zhou et al. 2009). This type of reprogramming of regulatory cells toward pathogenic characteristics is an issue that clouds prospects for regulatory cell therapy, and may require combination therapy with anti-inflammatory agents, deletion or inhibition of effector responses, and/or manipulation of the regulatory cells to be refractory to reprogramming, perhaps requiring gene editing or molecular engineering strategies.

Figure 1 shows the accumulation of defects in immune tolerance that cascade during the course of T1D disease onset, and indicates opportunities for targeted tolerance therapies. While genetic predisposition for autoreactivity is not amenable to direct revision, inadequate tolerance functions attributed to those genes are potential therapeutic targets. Preventing loss of tolerance at a very early stage may involve modifying environmental stimuli or remodeling the developing immune repertoire to interfere with autoantigen sensitization of naive autoreactive T cells. NOD mouse colonies can be protected from autoimmune diabetes using dietary, microbiome, or housing changes, observations that have been recognized for decades, providing rationale for efforts to find ways to modify those types of environmental stimuli in susceptible populations (Chapman et al. 2012). More targeted approaches that aspire to directly modify the selection and expansion of autoreactive T cells by modifying the peptide–MHC interaction during T-cell development are also potential ways to create an immune system refractory to breakdown of tolerance, by lowering the peripheral population of autoantigen-specific T cells and thus creating a higher barrier to initial immune activation. The evolution from T1D predisposition to progression is marked by activation and expansion of immune reactivities to islet antigens, offering a variety of opportunities for restoration of tolerance. Indeed, it is not the specific recognition of autoantigens that dictates pathogenic outcomes, it is the type of immune sequelae that follow such recognition events, with a multitude of nonpathogenic possibilities. Antigen recognition in a format that encourages deletion, deviation, anergy, or exhaustion of effector T cells is, therefore, a primary goal of many therapeutic strategies. Several tolerization strategies use different ways of delivering autoantigens for antigen presentation, considering localization to pancreata or lymph nodes, use of various or multiple islet antigens, and linking such autoantigens to ligands or functional tethering molecules to favor immune deletion or deviation of responses, for example by targeting antigen presentation in the liver or via tolerogenic dendritic cells. Since this phase of developing autoimmunity is characterized by antigen determinant spreading, it is also a time in which therapies designed to interfere with intercellular collaboration, such as T-cell interactions with B cells and/or with dendritic cells, are most promising. In particular, addressing both T-cell and B-cell tolerization may be most effective to interfere with this early expansion of antigen breadth, potentially inhibiting the spreading of autoreactivity to other islet antigens and PTM antigens. Recent studies of disease-associated B cells in T1D identify peripheral anergic autoreactive B cells (B_ND) that are potentially tolerant, although this appears to be reversible in a subset of cells (B_ND2) that directly contribute to pathogenesis (Stensland et al. 2023). Insight into the role of these cells in antigen presentation to autoreactive T cells could enable targeted therapeutics to help sustain a tolerogenic response. Interfering with stress pathways that favor PTM formation, potentially including inhibition of PAD enzymes (Barasa and Thompson 2023), may also have a role at this stage of the disease. These types of therapeutic strategies interfering with antigen-specific effector responses may be difficult later in disease progression, a time when multiple islet antigens are recognized as targets and when tissue inflammation is robust.

Figure 1.

The serial loss of tolerance in type 1 diabetes (T1D): opportunities for intervention. The establishment and perpetuation of defective peripheral immune tolerance in T1D is a cumulative process, illustrated as an evolution in three phases. First is the generation of a peripheral immune T-cell repertoire that is prone to autoimmunity, in which the predisposition for a loss of tolerance is genetically controlled and reflected in an increased frequency of autoreactive naiveperipheral T cells. Expansion of these autoreactive cells can be a result of antigen-specific or nonspecific environmental factors that represent a transition from a potentially autoimmune to an autoimmune-prone T-cell repertoire. Next is the activation of these autoreactive cells, which occurs in the peripheral immune system coincident with a failure of regulatory constraints, the “second hit” in unleashing pathogenic effector immune responses. In this phase, the spreading of the immune response occurs through the recruitment of B cells driving autoantibody production and additional antigen presentation, expansion to additional tissue antigens, and recognition of neoantigens generated from posttranslationally modified target proteins. Islet cell damage is accompanied by a β-cell stress response that further amplifies immune activity. A final phase is the confluence of multiple activated immune cell types and inflammatory pathways, which, in the absence of effective counterregulation, reflects unrestrained pathogenic autoimmunity without tolerogenic barriers. Immune defects that accumulate during this serial loss of tolerance identify pathways and targets for potential therapy designed to restore homeostasis. Opportunities for this type of mechanism-informed therapeutic intervention strategy are indicated in the lower panels, either preventing initial loss of tolerance, restoring tolerance capacity through rebalancing of regulatory and effector compartments, or replacing tolerogenic potential through multilineage immune reconstruction. As discussed in the text, a combination of therapeutic strategies is needed to address the extent and diversity of failed tolerance mechanisms during these phases of disease.

Selective interference with autoreactive effector cells, however, is insufficient for tolerance in the context of progressive autoimmunity, as evidenced by recurrent islet cell loss after discontinuation of depletional therapies, discussed in the previous section and summarized elsewhere (Smilek et al. 2014; Nepom 2022). Targeting effector responses during this phase of immune expansion and disease progression must be accompanied by efforts to enhance regulatory immune systems essential for durable immune homeostasis. Generating or reinforcing regulatory immune activity in the context of T1D progression, however, is challenged by several conundrums: (1) is it sufficient to boost regulatory T-cell responses to a particular islet antigen, or is it necessary to have regulatory specificity for multiple antigens; (2) will naturally occurring regulatory T cells, if boosted in numbers, be sufficient to control autoimmunity, or is it necessary to engineer stability and functional robustness into such cells; (3) will the islet microenvironment inflammatory profile of cytokines and chemokines obviate desired regulatory activity; and (4) is it necessary to use sequential therapies, first abrogating pathogenic effector responses to have a hospitable opportunity for therapies that support regulatory dominance? These are fundamental questions that lack suitable preclinical T1D models, and pose a challenge to conventional clinical trial designs. Novel “umbrella” or other adaptive trial designs (Woodcock and LaVange 2017) may be helpful for testing various options in an efficient manner, and/or a multidrug approach (Skyler 2015), perhaps modeled after the decades-old transformational “reverse the pyramid” strategy in rheumatoid arthritis (Wilske and Healey 1989), in which therapy starts with multiple agents combined, followed by selective withdrawal of individual components to assess durable tolerance.

The last panel in Figure 1 addresses the daunting task of reestablishing tolerance in the context of a raging autoimmune response, in which there are multiple types of activated immune cells, tissue inflammation, and innate immune catalysts all simultaneously converging on β-cell damage and destruction. Regenerating insulin secretory capacity, tissue integrity, and establishing a regulatory environment in the islets is likely to require a comprehensive approach to therapy that silences immune system effector responses, supports β-cell recovery to minimize a tissue stress response, and creates a robust regulatory network to prevent recurrent pathology. Notably, although some antigen-specific T-cell effectors and regulatory cells and some autoantibody specificities have been well-studied, very little is known about the tissue response in vivo. In particular, the role of the tissue environment in driving the pathogenesis of T1D and resistance to tolerization therapy may be critical and may differ among individuals due to underlying genetic variation or β-cell fitness. The findings in T1D clinical trials that anti-TNF monoclonal antibody therapy, but not anti-IL-1 or anti-IL-6R monoclonals, result in partial improvements in β-cell function is currently a conundrum (Moran et al. 2013; Greenbaum et al. 2021; Rigby et al. 2023) but an indication that there are opportunities to target both innate and adaptive immune pathways and explore potential synergistic effects. Even less is known about alternative stress-induced responses, including the neuronal interface with islet cells, that may prove to have a role in perpetuating disease or the tissue response to therapy.

CONCLUDING REMARKS

Immune tolerance in T1D is achievable, as evidenced by well-documented examples of the successful rescue of β cells in the early stages of the disease, with durable retention of islet function. When this occurs in the context of immune therapies such as teplizumab and alefacept, it is accompanied by functional and phenotypic changes in the peripheral immune response that are likely responsible for success. In other words, changes in peripheral immunity reflect and functionally arrest the pathogenic loss of β-cell function. However, this desirable result currently occurs infrequently, with most therapeutic outcomes being short-lived and not having durable effects on disease progression. Better understanding of the relationship between tissue-localized immune responses in the islet and those that are measured in the periphery may help improve tolerization success, which likely will require insight into innate immune pathways and other non-T-cell components perpetuating autoimmunity. Expanding the frequency of success and addressing the widespread clinical need requires using an entire toolbox of immune tolerance mechanisms and exploring novel combinations of therapeutic targets, simultaneously abrogating pathogenic effector responses and boosting homeostatic regulatory immunity.