REVIEWS Combination immunotherapies for type 1 diabetes mellitus Paolo Pozzilli, Ernesto Maddaloni and Raffaella Buzzetti Abstract | Immunotherapies for type 1 diabetes mellitus (T1DM) have been the focus of intense basic and clinical research over the past few decades. Restoring β‑cell function is the ultimate goal of intervention trials that target the immune system in T1DM. In an attempt to achieve this aim, different combination therapies have been proposed over the past few years that are based on treatments tackling the various mechanisms involved in the destruction of β cells. The results of clinical trials have not matched expectations based on the positive results from preclinical studies. The heterogeneity of T1DM might explain the negative results obtained, but previous trials have not addressed this issue. However, novel promising combination therapies are being developed, including those that couple immunomodulators with drugs that stimulate β‑cell regeneration in order to restore normoglycaemia. This strategy is an encouraging one to pursue the goal of finding a cure for T1DM. This Review summarizes the available data about combination immunotherapies in T1DM, particularly addressing their clinical importance. The available data supporting the use of registered drugs, such as proton pump inhibitors and incretin-based agents, that have been shown to induce β‑cell regeneration will also be discussed. Pozzilli, P. et al. Nat. Rev. Endocrinol. advance online publication 17 February 2015; doi:10.1038/nrendo.2015.8
Introduction
Department of Endocrinology and Diabetes, University Campus Bio-Medico, Via Álvaro del Portillo 21, Rome 00128, Italy (P.P., E.M.). Department of Experimental Medicine, “Sapienza” University, Viale Regina Elena 324, Rome 00161 Italy (R.B.). Correspondence to: P.P. [email protected]
Type 1 diabetes mellitus (T1DM) is an autoimmune condition characterized by the destruction of β cells and an inability to regenerate β cells.1 Several native antigens and antigens that are modified after translation have been identified as potential therapeutic targets,2–7 which has led to the development of immune intervention strategies that either tackle the immune dysregulation8,9 or induce specific immune tolerance to such antigens.10–12 Combining both approaches has been tested in patients with recent-onset (usually 70–80% of β cells are lost or not functioning.16 The process leading to β‑cell destruction starts long before the clinical onset of the disease; several factors involved in this process have been targeted in the attempt to find a cure for T1DM.17 Regardless of the molecules and pathways used as targets for therapies, β‑cell destruction is the key event in the development of T1DM; therefore, the ultimate goal of intervention trials is to preserve (and ideally restore) β‑cell mass and function.
NATURE REVIEWS | ENDOCRINOLOGY
ADVANCE ONLINE PUBLICATION | 1 © 2015 Macmillan Publishers Limited. All rights reserved
REVIEWS Key points ■■ As several mechanisms lead to β‑cell failure in type 1 diabetes mellitus (T1DM), strategies combining different molecules could target different pathways involved in the pathogenesis of T1DM ■■ Different combination therapies have been tested in preclinical studies with encouraging results, but results from clinical trials have not matched expectations ■■ Heterogeneity in the age of patients with T1DM and the low rate of human β‑cell regeneration are possible explanations for the negative results of clinical trials that should be addressed in future studies ■■ Proton pump inhibitors and incretin-based agents can stimulate β‑cell regeneration, so future trials in patients with T1DM should investigate their potential in combination with immunomodulatory compounds ■■ The issue of patient heterogeneity can be addressed by designing trials evaluating the relative effectiveness of interventions in specific subgroups of patients
Box 1 | Heterogeneity of β‑cell status and immune intervention in T1DM ■■ T1DM is an autoimmune disease in which the death of β cells is the final event in the onset of hyperglycaemia. At the time of diagnosis, ~70–80% of β cells are lost; however, the latest evidence shows that this percentage varies considerably among patients30,34 ■■ Several factors, including BMI, HbA1c, age at onset and genetic susceptibility, affect the degree of β‑cell function at diagnosis. These factors have been recognized as key determinants of the rate of β‑cell loss in the following years ■■ As these factors vary among patients, the natural history of the disease also varies, as does the response to immune intervention ■■ The heterogeneity of T1DM means that an improved selection of the patients that are included in immune-intervention trials is required, taking into account the residual β‑cell function measured by levels of C‑peptide, age, BMI and levels of HbA1c Abbreviation: T1DM, type 1 diabetes mellitus.
As the preservation of β cells is the primary efficacy outcome of intervention trials in T1DM, a sensitive and specific biomarker of β‑cell status should be used. MRI and nuclear imaging techniques have been proposed as potential new approaches to evaluate β‑cell mass in clinical trials.18–20 However, these techniques are still being developed for this application and have not yet been standardized such that they could be used to measure efficacy in intervention trials. If we cannot easily measure β‑cell mass in vivo, we can make it measurable by evaluating β‑cell function.21 Although β‑cell function is influenced by several factors not related to the amount of surviving β cells (for example, drugs that interfere with their function), the estimation of their capacity to secrete insulin is probably more clinically relevant than the assessment of β‑cell mass.22
C-peptide levels and β‑cell decline The measurement of C‑peptide levels in the serum of patients with T1DM has been used to investigate changes in β‑cell function since the 1970s.23 Indeed, the measurement of stimulated C‑peptide under standardized conditions is recognized by the American Diabetes Association as the most suitable primary outcome for clinical trials investigating how to preserve β cells. 24 Both the 2 h mixed-meal tolerance test (MMTT) and the glucagon-stimulation test (GST) have been used in
international clinical trials to estimate the C‑peptide reserve. These two tests work through different ways to stimulate β cells to produce insulin. MMTT is better at mimicking the physiology of meal ingestion,25,26 whereas GST is not influenced by other factors that could be impaired in T1DM such as the incretin system.27 Several studies have highlighted the relevance of residual C‑peptide in patients with T1DM, as patients with higher levels have improved metabolic control and a lower rate of complications than patients with low levels of C‑peptide. For example, in the Diabetes Control and Complication Trial, participants with stimulated C‑peptide levels >0.2 nmol/l at baseline had reduced HbA1c levels, reduced rates of hypoglycaemia and fewer end-organ complications. 28,29 These results indicate that the severity of T1DM increases progressively as C‑peptide levels decrease. Due to the relevance of C‑peptide levels in the assessment of residual β‑cell function, several studies have investigated the determinants of baseline C‑peptide at diagnosis as well of the decline in levels of C‑peptide following diagnosis. As expected, C‑peptide levels were higher in patients diagnosed by surveillance than in those diagnosed at the clinical onset of the disease. 30 C‑peptide levels were also higher in patients diagnosed after puberty than in those diagnosed before puberty.31,32 Moreover, HbA1c and BMI were inversely associated with fasting levels of C‑peptide at diagnosis.33 β‑cell function rapidly declines around the time of diagnosis and the slope of β‑cell loss becomes gentler as the disease progresses. 30 However, stimulated C‑peptide levels remain at or above 0.2 nmol/l (which is the level identified as the clinically relevant threshold for C‑peptide response) in some patients for years to decades after diagnosis, with the duration depending on genetic and clinical features.30 Data from TrialNet studies revealed that >90% of patients with T1DM have detectable levels of C‑peptide 2 years after the diagnosis and that 11% of patients had no notable fall from baseline by 2 years.34 In 2014, a large European multicentre longitudinal study showed an age-dependent decline in β‑cell function (as measured by fasting levels of C‑peptide) in patients with T1DM.35 Taken together, these data suggest that the velocity and severity of β‑cell decline in T1DM is influenced by several factors, including age at diagnosis, BMI, the degree of metabolic control and genetic predisposition.36–39 The nature of these factors and the interactions between them differ for each patient (contributing to the heterogeneity of the disease 40), which indicates that trials should be developed to test personalized approaches to the cure of T1DM that depend on baseline C‑peptide levels and on determinants of β‑cell loss. As the knowledge of the natural history of β‑cell decline is critical to design and interpret trials in T1DM, future research should aim to improve the description of these determinants. Indeed, patient heterogeneity can considerably influence the efficacy of immune intervention trials, as demonstrated by the effect of different HLA genotypes on the clinical outcomes of immunotherapy (Box 1).41
2 | ADVANCE ONLINE PUBLICATION
www.nature.com/nrendo © 2015 Macmillan Publishers Limited. All rights reserved
REVIEWS β-cell damage and β‑cell regeneration A dynamic and intense turnover process replaces damaged or lost β cells. During fetal development, this process is the result of differentiation of stem cells or progenitor cells, whereas after birth, new β cells are produced by replication of pre-existing β cells.42 The dysfunctional autoimmune response in T1DM means that T cells mediate the destruction of β cells. This response disrupts the regeneration mechanism, which means that the β‑cell proliferative capacity is not able to replace the cells lost. As the cellular and humoral immune changes associated with T1DM persist for decades after the diagnosis, the imbalance between β‑cell loss and regeneration is relevant even in patients with long-standing disease.43 Many theories have been developed about T1DM aetiology. These include molecular mimicry that might be due to dietary antigens or viral cross-reaction towards the β cells,44–46 involvement of the gut microbiome47,48 and defects in the expression of antigens on β cells in the pancreatic lymph nodes and thymus. Collectively, these events could determine the loss of immunological tolerance to β cells with consequent inexorable impairment of the mass and function of β cells. T lymphocytes, B lymphoc ytes and macrophages can be found in the islet infiltrate.49 This autoimmune response is made up of many phases and the trigger events have yet to be elucidated. However, CD4+ CD25+ FOXP3+ regulatory T (TREG) cells are known to be key for the maintenance of a noninflamed environment, as they suppress prolifer ation and the cytokine response of immune cells such as CD4+ CD8+ T cells, natural killer cells, natural killer T cells, B cells, dendritic cells and macrophages.50 This action still remains crucial for suppressing inflammation and many studies are focusing on this mechanism in an effort to restore the balance between TREG cells and T effector cells (TEFF) and the reconstitution of immuno logical tolerance to pancreatic autoantigens to avoid β‑cell death.
Combination therapy Immune monotherapies Numerous strategies have been developed to block the autoimmune-mediated β‑cell destruction; these strat egies are generally targeted at preserving C‑peptide levels and increasing the number of β cells with the aim of restoring β‑cell function. Previous studies have focused on different monotherapies, including non-antigen- specific immunomodulators (such as CTLA‑4 Ig,9 IL‑1 receptor antagonist 51 and anti-CD20 monoclonal antibody [mAb]) that aim to expand the TREG cell population8 and antigen-specific immunotherapies that have a high specificity for β‑cell autoimmunity.52 The Immune Tolerance Network initiated the AbATE trial, which used an anti-CD3 mAb (teplizumab) to halt β‑cell destruction and to induce tolerance in patients with T1DM.53 In this trial, two 14‑day courses of teplizumab (one at the beginning of the study and the other after 12 months) resulted in all patients achieving levels of HbA1c
Introduction
Department of Endocrinology and Diabetes, University Campus Bio-Medico, Via Álvaro del Portillo 21, Rome 00128, Italy (P.P., E.M.). Department of Experimental Medicine, “Sapienza” University, Viale Regina Elena 324, Rome 00161 Italy (R.B.). Correspondence to: P.P. [email protected]
Type 1 diabetes mellitus (T1DM) is an autoimmune condition characterized by the destruction of β cells and an inability to regenerate β cells.1 Several native antigens and antigens that are modified after translation have been identified as potential therapeutic targets,2–7 which has led to the development of immune intervention strategies that either tackle the immune dysregulation8,9 or induce specific immune tolerance to such antigens.10–12 Combining both approaches has been tested in patients with recent-onset (usually 70–80% of β cells are lost or not functioning.16 The process leading to β‑cell destruction starts long before the clinical onset of the disease; several factors involved in this process have been targeted in the attempt to find a cure for T1DM.17 Regardless of the molecules and pathways used as targets for therapies, β‑cell destruction is the key event in the development of T1DM; therefore, the ultimate goal of intervention trials is to preserve (and ideally restore) β‑cell mass and function.
NATURE REVIEWS | ENDOCRINOLOGY
ADVANCE ONLINE PUBLICATION | 1 © 2015 Macmillan Publishers Limited. All rights reserved
REVIEWS Key points ■■ As several mechanisms lead to β‑cell failure in type 1 diabetes mellitus (T1DM), strategies combining different molecules could target different pathways involved in the pathogenesis of T1DM ■■ Different combination therapies have been tested in preclinical studies with encouraging results, but results from clinical trials have not matched expectations ■■ Heterogeneity in the age of patients with T1DM and the low rate of human β‑cell regeneration are possible explanations for the negative results of clinical trials that should be addressed in future studies ■■ Proton pump inhibitors and incretin-based agents can stimulate β‑cell regeneration, so future trials in patients with T1DM should investigate their potential in combination with immunomodulatory compounds ■■ The issue of patient heterogeneity can be addressed by designing trials evaluating the relative effectiveness of interventions in specific subgroups of patients
Box 1 | Heterogeneity of β‑cell status and immune intervention in T1DM ■■ T1DM is an autoimmune disease in which the death of β cells is the final event in the onset of hyperglycaemia. At the time of diagnosis, ~70–80% of β cells are lost; however, the latest evidence shows that this percentage varies considerably among patients30,34 ■■ Several factors, including BMI, HbA1c, age at onset and genetic susceptibility, affect the degree of β‑cell function at diagnosis. These factors have been recognized as key determinants of the rate of β‑cell loss in the following years ■■ As these factors vary among patients, the natural history of the disease also varies, as does the response to immune intervention ■■ The heterogeneity of T1DM means that an improved selection of the patients that are included in immune-intervention trials is required, taking into account the residual β‑cell function measured by levels of C‑peptide, age, BMI and levels of HbA1c Abbreviation: T1DM, type 1 diabetes mellitus.
As the preservation of β cells is the primary efficacy outcome of intervention trials in T1DM, a sensitive and specific biomarker of β‑cell status should be used. MRI and nuclear imaging techniques have been proposed as potential new approaches to evaluate β‑cell mass in clinical trials.18–20 However, these techniques are still being developed for this application and have not yet been standardized such that they could be used to measure efficacy in intervention trials. If we cannot easily measure β‑cell mass in vivo, we can make it measurable by evaluating β‑cell function.21 Although β‑cell function is influenced by several factors not related to the amount of surviving β cells (for example, drugs that interfere with their function), the estimation of their capacity to secrete insulin is probably more clinically relevant than the assessment of β‑cell mass.22
C-peptide levels and β‑cell decline The measurement of C‑peptide levels in the serum of patients with T1DM has been used to investigate changes in β‑cell function since the 1970s.23 Indeed, the measurement of stimulated C‑peptide under standardized conditions is recognized by the American Diabetes Association as the most suitable primary outcome for clinical trials investigating how to preserve β cells. 24 Both the 2 h mixed-meal tolerance test (MMTT) and the glucagon-stimulation test (GST) have been used in
international clinical trials to estimate the C‑peptide reserve. These two tests work through different ways to stimulate β cells to produce insulin. MMTT is better at mimicking the physiology of meal ingestion,25,26 whereas GST is not influenced by other factors that could be impaired in T1DM such as the incretin system.27 Several studies have highlighted the relevance of residual C‑peptide in patients with T1DM, as patients with higher levels have improved metabolic control and a lower rate of complications than patients with low levels of C‑peptide. For example, in the Diabetes Control and Complication Trial, participants with stimulated C‑peptide levels >0.2 nmol/l at baseline had reduced HbA1c levels, reduced rates of hypoglycaemia and fewer end-organ complications. 28,29 These results indicate that the severity of T1DM increases progressively as C‑peptide levels decrease. Due to the relevance of C‑peptide levels in the assessment of residual β‑cell function, several studies have investigated the determinants of baseline C‑peptide at diagnosis as well of the decline in levels of C‑peptide following diagnosis. As expected, C‑peptide levels were higher in patients diagnosed by surveillance than in those diagnosed at the clinical onset of the disease. 30 C‑peptide levels were also higher in patients diagnosed after puberty than in those diagnosed before puberty.31,32 Moreover, HbA1c and BMI were inversely associated with fasting levels of C‑peptide at diagnosis.33 β‑cell function rapidly declines around the time of diagnosis and the slope of β‑cell loss becomes gentler as the disease progresses. 30 However, stimulated C‑peptide levels remain at or above 0.2 nmol/l (which is the level identified as the clinically relevant threshold for C‑peptide response) in some patients for years to decades after diagnosis, with the duration depending on genetic and clinical features.30 Data from TrialNet studies revealed that >90% of patients with T1DM have detectable levels of C‑peptide 2 years after the diagnosis and that 11% of patients had no notable fall from baseline by 2 years.34 In 2014, a large European multicentre longitudinal study showed an age-dependent decline in β‑cell function (as measured by fasting levels of C‑peptide) in patients with T1DM.35 Taken together, these data suggest that the velocity and severity of β‑cell decline in T1DM is influenced by several factors, including age at diagnosis, BMI, the degree of metabolic control and genetic predisposition.36–39 The nature of these factors and the interactions between them differ for each patient (contributing to the heterogeneity of the disease 40), which indicates that trials should be developed to test personalized approaches to the cure of T1DM that depend on baseline C‑peptide levels and on determinants of β‑cell loss. As the knowledge of the natural history of β‑cell decline is critical to design and interpret trials in T1DM, future research should aim to improve the description of these determinants. Indeed, patient heterogeneity can considerably influence the efficacy of immune intervention trials, as demonstrated by the effect of different HLA genotypes on the clinical outcomes of immunotherapy (Box 1).41
2 | ADVANCE ONLINE PUBLICATION
www.nature.com/nrendo © 2015 Macmillan Publishers Limited. All rights reserved
REVIEWS β-cell damage and β‑cell regeneration A dynamic and intense turnover process replaces damaged or lost β cells. During fetal development, this process is the result of differentiation of stem cells or progenitor cells, whereas after birth, new β cells are produced by replication of pre-existing β cells.42 The dysfunctional autoimmune response in T1DM means that T cells mediate the destruction of β cells. This response disrupts the regeneration mechanism, which means that the β‑cell proliferative capacity is not able to replace the cells lost. As the cellular and humoral immune changes associated with T1DM persist for decades after the diagnosis, the imbalance between β‑cell loss and regeneration is relevant even in patients with long-standing disease.43 Many theories have been developed about T1DM aetiology. These include molecular mimicry that might be due to dietary antigens or viral cross-reaction towards the β cells,44–46 involvement of the gut microbiome47,48 and defects in the expression of antigens on β cells in the pancreatic lymph nodes and thymus. Collectively, these events could determine the loss of immunological tolerance to β cells with consequent inexorable impairment of the mass and function of β cells. T lymphocytes, B lymphoc ytes and macrophages can be found in the islet infiltrate.49 This autoimmune response is made up of many phases and the trigger events have yet to be elucidated. However, CD4+ CD25+ FOXP3+ regulatory T (TREG) cells are known to be key for the maintenance of a noninflamed environment, as they suppress prolifer ation and the cytokine response of immune cells such as CD4+ CD8+ T cells, natural killer cells, natural killer T cells, B cells, dendritic cells and macrophages.50 This action still remains crucial for suppressing inflammation and many studies are focusing on this mechanism in an effort to restore the balance between TREG cells and T effector cells (TEFF) and the reconstitution of immuno logical tolerance to pancreatic autoantigens to avoid β‑cell death.
Combination therapy Immune monotherapies Numerous strategies have been developed to block the autoimmune-mediated β‑cell destruction; these strat egies are generally targeted at preserving C‑peptide levels and increasing the number of β cells with the aim of restoring β‑cell function. Previous studies have focused on different monotherapies, including non-antigen- specific immunomodulators (such as CTLA‑4 Ig,9 IL‑1 receptor antagonist 51 and anti-CD20 monoclonal antibody [mAb]) that aim to expand the TREG cell population8 and antigen-specific immunotherapies that have a high specificity for β‑cell autoimmunity.52 The Immune Tolerance Network initiated the AbATE trial, which used an anti-CD3 mAb (teplizumab) to halt β‑cell destruction and to induce tolerance in patients with T1DM.53 In this trial, two 14‑day courses of teplizumab (one at the beginning of the study and the other after 12 months) resulted in all patients achieving levels of HbA1c
