Improving Clinical Islet Transplantation Outcomes

Transplantation of islets purified from donor pancreata is one of only a few clinical strategies that can reverse type 1 diabetes (T1D) and restore glycemic control (1–3). Significant progress has been made by improving transplantation strategies in areas including immune suppressive regimes, islet quality, transplanted β-cell mass and purity, and donor selection (4–6). However, while improved metabolic control and reduced glycemic side effects such as hypoglycemia persist long term in the majority of T1D patients, insulin independence beyond 1 year is only achieved in a minority of cases (2). Early loss of islet mass upon portal vein infusion is believed to have a major negative impact (7). Recurrence of insulitis, the hallmark of T1D immunopathogenesis leading to autoimmune mediated β-cell destruction, and islet allograft rejection pose additional therapeutic hurdles (8,9).

Recently, neutrophils were suggested as part of the innate immune system contributing to development of T1D (10). Chemokines produced by distressed islet tissue attract leukocytes to the site of implantation (Fig. 1). The chemokine CXCL8, produced by recruited and resident macrophages, attracts neutrophils carrying its receptors CXCR1/2 to the lesion. In this issue of Diabetes Care, Piemonti and colleagues (11) now report the results of a phase 3 clinical trial assessing whether interference in homing of leukocytes to the implanted islet graft using reparixin, a small molecule that interferes with CXCR1/2 signaling, may improve transplantation outcome. The rationale for this strategy is outstanding. Early studies in mice and humans were very encouraging. Treatment with reparixin prevented and reversed autoimmune diabetes in NOD mice and improved graft survival in experimental models of islet transplantation (12,13). Such positive effects were confirmed in a pilot clinical trial in T1D patients receiving allogeneic islet transplantation. The results led to a pivotal phase 3 trial to compare this new therapeutic strategy with standard clinical practice.

The investigators are to be applauded for their diligent design of this clinical study, which is unique in clinical islet transplantation. They elected to test chemokine interference using a preferred design in clinical trials, i.e., a randomized, double-blind, parallel-assignment study, in a multicenter setting. Completion of this head-to-head comparison of different islet transplant protocols in a clinical trial was a tour de force. While other islet transplantation trials have suffered from slow accrual rates and completion, the participating clinical centers in Europe and North America managed to complete the study in a remarkably fast fashion. Nonetheless, the outcome of this trial was disappointing; the rate of stimulated C-peptide production was identical in patients treated with or without reparixin (2.75 ± 2.74 vs. 3.64 ± 1.94 ng/mL peak C-peptide, respectively, following a mixed-meal tolerance test), as was the rate of insulin independence early or 1 year after islet transplantation (32% vs. 31% of patients, respectively) (11). Also, no significant differences were seen between treatment groups at any time point for any secondary measures of glycemic control. However, there was one glimmer of hope: although perhaps not surprising (14), conditioning therapy with thymoglobulin rather than basiliximab in islet recipients with TID proved to associate with superior clinical outcome, which tended to be better when used in combination with reparixin. This was in spite of higher cytokine and chemokine secretion following thymoglobulin therapy.

The incentive for the reparixin trial had everything going for it. Why did it fail to deliver? Earlier clinical studies assessing interference of cytokines of the innate immune system (tumor necrosis factor, interleukin-1), to blunt cytokine release syndrome associated with thymoglobulin induction therapy and to reduce the immediate loss of islet allograft mass, successfully improved clinical outcome (15). Also, the preclinical studies in mice were very encouraging. However, this would not be the first time that studies in mice have been misleading (16). As with every immune suppressive strategy, targeting the CXCR1/2-CXCL8 axis may be a double-edged sword: in cancer, this axis contributes to the migration of a heterogeneous group of immune cells referred to as myeloid-derived suppressor cells (MDSCs) to the tumor lesion (17). MDSCs contribute to immune suppression in the hypoinflammatory tumor microenvironment that hinders the immune system from dealing with cancers. However, in the case of T1D, immune cells such as MDSCs might be favorable, reducing inflammation at the site of islet implantation or during insulitis. Another issue that may have contributed to the unfortunate lack of efficacy of CXCR1/2 inhibition may be related to an additional critical chemokine pathway that remained untouched. CXCL10, produced by distressed β-cells, attracts the infamous islet autoreactive T cells to the islet lesion, contributing to insulitis and β-cell destruction (Fig. 1) (18). It is conceivable that the same mechanism is active in the case of islet allograft transplantation. Injecting islets in the portal vein, thereby exposing endocrine tissue directly to the bloodstream, may add to the complexity and challenges of preventing migration of autoreactive or alloreactive T cells and other inflammatory immune cells to the islet graft (19).

Clinical islet transplantation faces numerous problems beyond donor shortage and preimmunized recipients (7). Differences in clinical practice of immune suppression hinder multicenter studies; there is no consensus on preferred immune-suppressive drugs. This may have an advantage, as it allows for serendipitous findings in subsets of patients or clinical centers using different immune-suppressive regimes. This is illustrated by the present phase 3 trial with regard to differences between thymoglobulin and basiliximab, which would have remained unnoticed if one standard induction therapy had been elected (11). As with any advanced medicinal cell therapy, standardizing the graft quality and composition within and between clinical centers and continents is a complex and daunting task (20). It may be hard to avoid unintentional “cherry picking” of superior islet grafts to achieve better outcome of novel islet transplantation strategies, and it is here where a randomized and double-blind trial design is so welcome. The reparixin trial is the first of its kind in islet transplantation to do so. Phase 3 trials of similar large scale have been performed in islet transplantation before to reverse hypoglycemic unawareness and restore glycemic control; smaller scale phase 1 and 2 trials have been done to validate different islet isolation and infusion techniques, implantations sites, and immune intervention strategies (2,21). However, none have been designed as head-to-head comparisons between islet transplantion strategies in a blind fashion and multicenter international setting, a gold standard for clinical trials.

While the outcome of this trial is a setback, the diabetes research and clinical community should be grateful to the investigators for their exquisite study design, diligent efforts to assess efficacy of reparixin in the context of islet transplantation, and successful execution of this challenging trial. They point the way forward demonstrating that randomized, double-blinded studies are feasible and valuable to guide new strategies to improve clinical outcome of intervention therapy in T1D, even in islet transplantation. Quod erat demonstrandum.