Protecting islet functional viability using mesenchymal stromal cells

Abstract Islet transplantation is an emerging treatment for type 1 diabetes which offers the prospect of physiological control of blood glucose and reductions in acute hypoglycaemic episodes. However, current protocols are limited by a rapid decline in islet functional viability during the isolation process, culture period, and post‐transplantation. Much of this can be attributed to the deleterious effects of hypoxic and cytokine stressors on β cells. One experimental strategy to improve the functional viability of islets is coculture or cotransplantation with mesenchymal stromal cells (MSCs). Numerous studies have shown that MSCs have the capacity to improve islet survival and insulin secretory function, and the mechanisms of these effects are becoming increasingly well understood. In this review, we will focus on recent studies demonstrating the capacity for MSCs to protect islets from hypoxia‐ and cytokine‐induced stress. Islets exposed to acute hypoxia (1%‐2% O2) or to inflammatory cytokines (including IFN‐γ, TNF‐α, and IL‐B) in vitro undergo apoptosis and a rapid decline in glucose‐stimulated insulin secretion. Coculture of islets with MSCs, or with MSC‐conditioned medium, protects from these deleterious effects, primarily with secreted factors. These protective effects are distinct from the immunomodulatory and structural support MSCs provide when cotransplanted with islets. Recent studies suggest that MSCs may support secretory function by the physical transfer of functional mitochondria, particularly to metabolically compromised β cells. Understanding how MSCs respond to stressed islets will facilitate the development of MSC secretome based, cell‐free approaches to supporting islet graft function during transplantation by protecting or repairing β cells.


| INTRODUCTION
Type 1 diabetes mellitus (T1D) is an autoimmune disorder in which the host immune system mistakenly destroys the insulin-secreting β cells located within the pancreatic islets of Langerhans. T1D was a lethal metabolic disorder before the discovery of insulin in the early 20th century 1 and people with T1D are still dependent on the daily administration of exogenous insulin to survive. The single cell-type pathology of T1D makes it an attractive candidate for cell replacement therapy and the transplantation of isolated pancreatic islets, which are the only source of primary β cells, offers a long-term alternative to conventional insulin therapy. Currently, over 50% of islet graft recipients maintain insulin independence and/or regain hypoglycemic awareness for up to 5 years post-transplantation. 2 However, the number of people with T1D who benefit from this therapy is limited by the scarcity of appropriate donor pancreata and by the excessive loss of β cell functional viability during islet isolation, culture, and transplantation. 2 Islet cells are damaged during the harsh process of collagenase digestion of whole pancreas to isolate the islets. MSCs are multipotent, adult progenitor cells located in the perivascular niche of most post-natal tissues. The minimum criteria to define human MSCs include plastic adherence in standard culture conditions, expression, or lack of expression of specific cell surface markers (CD90 +ve , CD105 +ve , CD73 +ve , CD34 -ve , CD31 -ve , CD45 -ve , and CD14 -ve ) and the capacity to differentiate into osteoblasts, adipocytes, and chondrocytes in vitro. 5 MSCs also exhibit many functional properties which are greatly influenced by the host niche: they lay down extracellular matrix as structural support and a reservoir for bioactive molecules; they regulate and suppress innate and adaptive immune cells through secreted factors and cell-contact dependent mechanisms 6 and they attenuate inflammatory responses through the secretion of anti-inflammatory cytokines. 7 The MSC secretome includes a range of angiogenic growth factors, 8 30% of all known extracellular matrix proteins 9 and anti-apoptotic factors. 10 This wide range of functionalities makes MSCs attractive candidates to support and protect cells in hostile transplantation environments.
There is a convincing body of evidence that coculture of isolated rodent or human islets with MSCs maintains their functional viability in vitro as assessed by glucose stimulated insulin secretion (GSIS), reduced apoptosis, and enhanced β cell mass. [11][12][13][14][15] Beneficial effects of MSCs on islet function in vitro have been reported using MSCs isolated from a range of mouse and human tissues, including adipose, bone marrow, kidney, and pancreas. 16 Similarly, coculture or coadministration of MSCs with islets improves the experimental outcomes of islet transplantation in syngeneic, allogenic, and humanized animal models of diabetes, 17 although experimental studies of islet pretreatment have not yet translated into human clinical trials. However, many experimental studies using rodent tissues do not fully replicate the cellular stresses experienced by human islet grafts. For example, experimental rodent islets are isolated rapidly from young, healthy, lean, and genetically homogenous animals. In contrast, human islets are isolated from pancreata harvested from heart-beating, braindead donors, and factors such as age, body mass index, and duration of brain death impact on the exposure of the islets to insults and on islet function in vitro. 2,3 Technical differences in the islet isolation process also result in human islets being subjected to much greater levels of cellular stress than experimental rodent islets. Human islets experience a prolonged cold ischaemia time resulting in upregulation of hypoxia inducible factor-1α regulated genes. Thus, the gene expression profile of isolated human islets maintained under normoxic conditions (20% O 2 ) resembles that of mouse islets exposed to extreme hypoxia (1% O 2 ) 18 suggesting that isolated human islets are inherently more stressed than experimental rodent islets. This review will therefore focus on the capacity of MSCs to protect β cells from the cellular stressors associated with islet isolation, culture, and transplantation. proportionally high blood supply to support the high rates of mitochondrial oxidative respiration on which the β cell secretory process is dependent. 19 The endocrine islets comprise 2% of the total pancreas but receive around 15% of its blood supply which maintains a higher oxygen partial pressure in islets than in the surrounding exocrine tissue. 20 Consequently, islet cells are subjected to prolonged hypoxia from the time of surgical removal of the donor pancreas to the completion of post-transplantation revascularization ( Figure 1). The duration of cold ischemic time during islet isolation correlates to reduced islet yields and impaired β cell secretory function and can be ameliorated to some extent by maintaining the donor pancreas in a high oxygen environment. 21 During postisolation culture, when islet cells depend on diffusion for their oxygen supply, hypoxia leads to upregulation of genes associated with oxidative stress and apoptosis. 22,23 In animal models, islet graft revascularization begins during the first week and is typically complete within 10 to 14 days. Prior to this, islets infused into the hepatic portal vein rely on portal venous flow so oxygen supply is limited, especially to the cells in the core of the islets which rapidly

Significance statement
Understanding how mesenchymal stromal cells respond to and protect islets damaged by transplantation-related stressors will facilitate the development of mesenchymal stromal cell secretome-based, cell-free approaches to supporting islet graft function during transplantation therapy for type 1 diabetes by protecting or repairing β cells. undergo necrosis. 24 Even when revascularization is complete, the oxygen tension in transplanted islets remains low regardless of transplantation site 25 and this is thought to be a major contributor to posttransplantation loss of β cell function and graft failure. 22

| Innate immune response and inflammation
As with any transplanted tissue, the clinical success of islet transplantation is constrained by the recipient immune response (Figure 1).
Immune system-mediated damage to islets begins in the donor pancreas in situ in response to systemic increases in inflammatory cytokines in brain-dead organ donors ("cytokine storm"). This is accentuated by more directed islet cell destruction by the host innate immune system immediately after islet transplantation. Infusion of islets into the hepatic portal vein triggers the nonspecific instant blood-mediated inflammatory reaction, characterized by the activation of the complement cascade, localized thrombus formation, and the infiltration of innate immune cells. 26 A variety of activated innate immune cells-including neutrophils, macrophages, and Kupffer cellsinduce β cell damage by the localized release of free radicals and proinflammatory cytokines, 4,27 which is estimated to adversely affect the functional viability of up to 50% engrafted β cells. 26 In vitro exposure of islet cells to cocktails of the pro-inflammatory cytokines IFN-γ, IL-1β, and TNF-α is often used as a model of acute, high-grade inflammation in type 1 diabetes or islet transplantation ( Figure 1). Isolated human islets are more consistently affected than rodent islets by the cytotoxic effects of these cytokine cocktails, most F I G U R E 1 Islets are subjected to hypoxic and inflammatory stressors throughout the isolation and transplantation process. Islets are removed from their dense vasculature during the isolation process and maintained at a lower oxygen environment in the subsequent culture period. Following transplantation, islet graft revascularization begins within a week but islets must rely on a limited oxygen supply from portal venous blood flow prior to this. Infusion of islets into the hepatic portal vein triggers the nonspecific inflammatory and thrombotic reaction called the immediate blood-mediated inflammatory reaction (IBMIR). Activated innate immune cells continue to interact with islets post-transplantation, releasing free radicals and pro-inflammatory cytokines. The sum of these prolonged stressors leads to an increase in islet cell apoptosis and a decrease in glucose stimulated insulin secretion, which reduces the long-term efficacy of the graft. Peri-and post-transplantation conditions can be modeled in vitro by treatment with pro-inflammatory cytokines (IFN-γ, IL-1β, and TNF-α) and incubation in 1% to 2% O 2 , with similar effects on cell death and function likely because of their inherently higher levels of cellular stress, as discussed above. Exposure of both human and rodent islets to proinflammatory cytokines impairs insulin secretion, increases the expression ER stress markers and apoptosis, and results in increased β cell death, 28,29 consistent with the observed loss of functional islets during the immediate post-transplantation period in vivo.

| Hypoxia
There is a growing body of evidence that MSCs can protect β cells from some of the deleterious effects of the hypoxia associated with islet transplantation. In particular, cotransplantation of MSCs is known to enhance islet graft revascularisation. 13,30,31 However, most islet function is lost within the first few days after transplantation, well prior to revascularization, so the beneficial effects of MSCs on angiogenesis are unlikely to be important in the immediate posttransplantation period. The initial protection from hypoxia by MSCs may therefore be a direct effect on the islet β cells.
The acute hypoxia of the post-transplantation environment can be modeled in vitro by incubation under an atmosphere of 1% to 2% O 2 , which is reported to increase apoptosis and decrease GSIS in rat islets within 8 hours. 28 The deleterious effects of hypoxia on insulin secretion were significantly reduced when the islets were cocultured on a monolayer of bone marrow-derived MSCs, consistent with a direct effect on the β cells. 28 The upregulation of islet cell apoptosis in both early and late stage hypoxia was also significantly reduced by coculture on MSC monolayers, as demonstrated by reduced expression of ER stress markers, BIP and CHOP, at both mRNA and protein levels. 32 29 Similarly, two recent studies have shown that conditioned medium from human umbilical cord MSCs can protect porcine islets from hypoxia-induced (24-48 hours) apoptosis, an effect which was associated with inhibition of oxidative stress and downregulation of total and mitochondrial reactive oxygen species (ROS) production in cocultured islets in one study, 34 and increased autophagy in cocultured islets in the other. 35 The conditioned medium in the latter study contained high concentra-  In studies using rodent or human islets exposed to cocktails of proinflammatory cytokines, MSC coculture was reported to preserve GSIS, 40,41 protect against cytokine-induced apoptosis, and maintain islet morphology. 41 Human islets cultured in MSC-conditioned medium were protected from apoptosis following cytokine exposure, and maintained normal islet morphology. 42 These effects have been attributed to a range of MSC secreted factors which can, to some extent, mimic the effects of MSC coculture. Concentrations of cytoprotective factors VEGF-A, fibroblast growth factor-2 (FGF-2), and nerve growth factor were elevated in MSC-conditioned medium following MSC exposure to hypoxia. 42 Islets incubated in these media show lower rates of cytokine-induced apoptosis compared to those cultured in conditioned medium from MSCs incubated under normoxia. During coculture with islets, HGF and MMP-2 production by MSCs was increased following exposure to cytokines. Exogenous HGF alone improved β cell glucose responsiveness, albeit not to the same extent as MSC coculture, and glucose release remained elevated at basal and stimulatory concentrations of glucose. 41 We have identified a number of MSC-derived G-protein coupled receptor agonists which protect mouse 43  Intercellular transfer of intact mitochondria has been implicated in stem cell-mediated repair of damaged cells in a variety of experimental models and pathologies. 47 In many studies using MSCs, recipient cell stress is required for mitochondrial transfer with MSCs donating more mitochondria to damaged endothelial cells and cardiomyocytes than healthy cells. 47 These studies are consistent with our recent observations of increased mitochondrial transfer to hypoxia-exposed mouse islets, and human islets 18 which exhibit a hypoxic gene signature in culture not observed in rodent islets during culture, 23