Current Status of Islet Cell Replacement and Regeneration Therapy

Abstract

Context: Β cell mass and function are decreased to varying degrees in both type 1 and type 2 diabetes. In the future, islet cell replacement or regeneration therapy may thus offer therapeutic benefit to people with diabetes, but there are major challenges to be overcome.

Evidence Acquisition: A review of published peer-reviewed medical literature on β-cell development and regeneration was performed. Only publications considered most relevant were selected for citation, with particular attention to the period 2000–2009 and the inclusion of earlier landmark studies.

Evidence Synthesis: Islet cell regenerative therapy could be achieved by in situ regeneration or implantation of cells previously derived in vitro. Both approaches are being explored, and their ultimate success will depend on the ability to recapitulate key events in the normal development of the endocrine pancreas to derive fully differentiated islet cells that are functionally normal. There is also debate as to whether β-cells alone will assure adequate metabolic control or whether it will be necessary to regenerate islets with their various cell types and unique integrated function. Any approach must account for the potential dangers of regenerative therapy.

Conclusions: Islet cell regenerative therapy may one day offer an improved treatment of diabetes and potentially a cure. However, the various approaches are at an early stage of preclinical development and should not be offered to patients until shown to be safe as well as more efficacious than existing therapy.

The current progress in deriving pancreatic islet cells for implantation is reviewed, as well as possible approaches towards in situ islet cell regeneration for the treatment of diabetes.

Diabetes is a devastating chronic disease afflicting hundreds of millions of individuals: it is expected to increase inexorably over the coming decades. There is currently no cure for diabetes, and despite great improvements in treatment, the morbidity and increased mortality associated with diabetic complications are still a major concern. Unfortunately, most existing therapies fail to provide adequate control. Thus, there is an urgent need for more effective and durable management of diabetes, and ultimately a cure. This review considers just one such approach: islet cell regeneration therapy, embracing both the replacement/implantation and bona fide regeneration of islet cells in a patient. It is not our purpose to provide a comprehensive review of the literature but rather to describe to the clinician how such therapy may be achieved, where we stand today, and the technical as well as ethical challenges for the future.

Setting the stage: lessons from islet transplantation

The rationale for β-cell replacement therapy is driven by not only the near total deficiency of the cells in type 1 diabetes but also the now accepted finding that β-cell mass is reduced in type 2 diabetes (1,2). Although many patients benefit from pancreas transplants, the drawbacks remain major surgery with a high complication rate and the need for immunosuppression. With simultaneous pancreas kidney transplants, the major benefit is from the kidney, with increased long-term survival (3); the addition of a pancreas improves quality of life and hypoglycemia unawareness but has only modest effects on diabetic complications (4). Serious work on islet replacement began in the early 1970s with the demonstration that isolated islets could reverse diabetes in rodents (5), but the first convincing clinical demonstration was achieved only in 1989 (6). Although the pioneering centers could only provide short-term insulin independence in about 10% of their transplant recipients, even this seemingly modest success was extremely important for individuals with type 1 diabetes, offering the chance of life without diabetes.

A new era opened in 2000 with the report from Edmonton that seven consecutive subjects with type 1 diabetes were rendered insulin independent with islet transplants (7); more current data suggest that 72% of all recipients became insulin independent (8). Because one of the most common indications for transplantation was disabling hypoglycemia, a major success was the striking elimination of this problem. Another success has been the demonstration that increasing numbers of recipients can become insulin independent with infusion of islets from only one donor (9). Unfortunately, sobering realities accompanied this success. Perhaps the biggest disappointment was lack of durability such that more than 50% of those with insulin independence were back on insulin within 2 yr (10). On the positive side, despite this return to insulin injections, the accompanying continued (low grade) insulin secretion from the remaining graft still smoothed glycemic control in many of these patients and reduced the incidence of hypoglycemia. However, progression to complete graft failure occurred within a few years. This rapid loss of function is in contrast to the better durability of islet autotransplants (11), leading to the conclusion that graft loss is largely due to continuing assault by autoimmunity and allorejection as well as toxicity of immunosuppressive drugs. In addition to the problem of graft durability, there are risks from both the transplant procedure and immunosuppression (8).

Whereas we can look forward to better outcomes from continuing improvements in immunosuppression and the quality of islets, because of the limited islet supply, we can expect only hundreds rather than thousands of transplants annually for the next several years (the maximum number of available pancreases in the United States is only about 4000 per year and only a minority of these will provide islets of sufficient quality).

Moving to the postcadaveric islet transplantation era

The continuing work with transplants using cadaver islets prepares us for a new era of better sources of insulin producing cells and protection of these cells from immunological destruction. For the cell source problem, an extraordinary amount of progress is being made with stem cell biology, transdifferentiation, and xenotransplantation, and there are indications that β-cell regeneration within the pancreas may be possible. Although not the major focus of this review, it will obviously be necessary to protect new β-cells, regardless of their origin, from renewed autoimmune attack in individuals with type 1 diabetes and rejection in the example of xeno- or allotransplantation. Suffice it to say that progress is being made with our understanding of immunological tolerance and inflammation, and protective immunological barriers may prove useful.

The possibility of transplanting porcine islet cells continues to receive considerable attention because of their availability and similarities between porcine and human glucose metabolism. The hyperacute rejection of xenotransplants caused by preformed antibodies interacting with the Gal-α-Gal epitope is a major problem with organ transplantation but less so with cellular transplants. Strikingly improved results have been obtained in nonhuman primates using adult porcine islets (12), neonatal pancreatic cell clusters (13), or embryonic pancreatic tissue (14). Although these results are encouraging, there are concerns as to whether the immunosuppression used for these experiments in nonhuman primates will be efficacious in humans and not too toxic. Active research to control xenorejection includes the production of transgenic pigs to make islets more resistant to rejection by various maneuvers including removal of the Gal-α-Gal epitope (13) and the use of immunobarriers. For more than 30 yr, there has indeed been exploration of the potential of semipermeable membranes to protect islets from immune destruction (15). The principle is that glucose, nutrients, and oxygen can reach the encapsulated cells and insulin can be released, but immune cells and perhaps antibodies are excluded. Various configurations have been developed. Those receiving the most attention include microencapsulation in which islets are contained in small (400–1000 μm) beads of an alginate gel and macroencapsulation in which islets are enclosed within an implantable device consisting of planar sheets.

Progress has been frustrating because impressive success in rodents (16) has not yet been reliably reproduced in large animals or humans. The question of how much permselectivity is needed is very complex. It is possible that exclusion of cells will be the only requirement and that cytokine penetration will not be an issue. One can wonder about the need to restrict the release of antigens. The bigger issue is probably the biocompatibility of the materials. An inflammatory reaction with neutrophils and macrophages on the outside will result in dead islet cells inside, probably due to suffocation by competition for oxygen rather than by cytokines. If the biomaterials are inert, there should be no attached cells, which raises the question of where cytokines would come from and whether they would ever threaten the islet cells inside. Despite the difficulties, the work continues because safe control of autoimmunity and transplant rejection may take many years to achieve.

Regeneration and transdifferentiation

In human pancreas, it now seems clear that there is a slow rate of β-cell turnover whereby β-cells replicate and new islets are formed, probably from exocrine duct cells through the process of neogenesis (1,17,18). The relative rates of these two processes are not known, but the rate of β-cell replication seems to slow with age and neogenesis can be stimulated by injury. After decades of type 1 diabetes, small numbers of β-cells can invariably be found in the pancreas, supporting the concept that β-cells continue to be formed throughout life but are then killed by autoimmunity. Limits to the capacity for regeneration in type 2 diabetes are exemplified by pathological studies that do not find evidence for active regeneration (1).

Approaches to in situ regeneration are to stimulate either β-cell replication or neogenesis. In vitro expansion is also possible, in a process likely to involve dedifferentiation of β-cells in tissue culture, proliferation, and then redifferentiation (19,20). Success with such an in vitro intervention on islets derived from cadavers could lead to an expanded population of β-cells suitable for transplantation. Finding a molecular intervention that can be safely used in vivo seems more challenging but not impossible.

There are ways in which neogenesis might be stimulated to expand β-cell mass. Based on animal experiments, agents that may contribute include exendin-4, an analog of the incretin hormone glucagon-like-peptide-1; gastrin; epidermal growth factor; and islet neogenesis-associated protein (21). Much work is now underway to determine the potential of these agents used alone or in combination.

Transdifferentiation is a process in which a differentiated cell can be reprogrammed to change its identity. A considerable amount of work has been devoted to trying to convert cells in the liver into β-cells (22,23), and this effort continues (24,25). An exciting new possibility has emerged with the demonstration that pancreatic acinar cells may possibly be reprogrammed in mice with injections into the pancreas of adenoviruses expressing just three transcription factors, pancreatic duodenal homeobox-1, musculoaponeurotic fibrosarcoma oncogene homolog A, and neurogenin-3 (26). However, until this study is developed further, it is not possible to distinguish clearly between such reprogramming of acinar cells and β-cell neogenesis from precursor cells.

Recapitulating normal development of the endocrine pancreas

Early attempts to generate β-cells directly from embryonic stem (ES) cells or other cell sources, although apparently successful in producing cells that contained insulin (27,28,29), failed to produce true β-cells (30,31). Usually these cells expressed very little insulin, failed to respond adequately to normal secretagogues like glucose, and had non-β-cell characteristics, often looking suspiciously like neurons (32). In fact, during normal development in mammals, certain non-β-cells produce low amounts of insulin, including cells in the extraembryonic membranes and neurons in the central nervous system (33); these early attempts likely generated some of these interesting, but ultimately clinically impractical, cells. One lesson from these efforts was that recapitulation of normal development was the surest route to normal β-cells.

During normal development, the pancreas, and subsequently β-cells, arise from the endoderm germ layer (for basic reviews see Refs. 34,35,36). Early protocols for generating β-cells from ES cells did not attempt to go through endoderm, and this shortcut may have ultimately doomed those efforts. In the past few years, however, several groups have succeeded in generating endoderm (37,38). This effort has been rewarded by more efficient derivation from human ES cells of insulin-producing cells containing more insulin and with more β-cell-like characteristics (39). However, these in vitro-derived cells still fail to function like normal β-cells [even if maturation in vivo after implantation of the β-like cells into mice has been shown to promote their further differentiation toward a more convincing β-cell phenotype (40)], and once again normal development may hold the answers.

From mouse studies we know that there are four phases of β-cell generation. Before embryonic day (e) 13 in the mouse pancreas, the few endocrine cells generated (predominantly glucagon with a few insulin producing cells) differ from normal α- and β-cells (41,42): the insulin-expressing cells contain low levels of insulin, often coexpress glucagon, and lack the mature β-cell markers Nkx6.1, MafA, and pancreatic duodenal homeobox-1. Then at e13, β-cell neogenesis accelerates dramatically, and these new cells look like mature β-cells (although they may still lack robust glucose stimulated insulin secretion). At the same time, the exocrine acinar cells start to differentiate, and the progenitor cells become restricted to the ducts.

These synchronous differentiation events have been termed the secondary transition (43), and the atypical endocrine cells that differentiate before the secondary transition can be called primary endocrine cells to distinguish them from the mature islet cells that appear later. Although originally discussed 35 yr ago (43), the issue of the difference between primary endocrine cells and mature islet cells remains very relevant today because the insulin-producing cells generated to date from ES cells in vitro appear to have the characteristics of primary endocrine cells (39), rather than the normal mature β-cells of the secondary transition.

It is not clear whether the events that occur around the secondary transition are simply time dependent, result from some extracellular signal (44), or a combination of both. Whatever their basis, the pancreatic progenitor cells, from which the endocrine and exocrine cells differentiate, themselves undergo distinct morphological and gene expression changes during the secondary transition (41,45,46) and successively acquire the capacity to generate each of the mature islet cell types (45). Therefore, the stage of the progenitor cells may determine the type of endocrine cells they can generate (primary endocrine cells vs. mature islet cells). The process of differentiation of progenitor cells down the endocrine lineage is initiated by the transient expression of the transcription factor neurogenin-3, which then activates a cascade of transcription factors that drive differentiation and determine the final characteristics of the resultant endocrine cell (34,35,36).

The genesis of β-cells from the secondary progenitor cells peaks around e14–15 in the mouse and has largely ceased by e18 (47). The third phase of β-cell formation starts shortly before this termination: existing β-cells start to proliferate, resulting in a marked expansion in β-cells that lasts through the first few weeks of postnatal life (48,49). A similar expansion occurs in human infants (50).

Unfortunately, we know the least about the final phase of β-cell generation: replacement or expansion of β-cells in the adult. We know it occurs and underlies the ability of β-cell mass to fluctuate in response to metabolic needs (1,51), but we do not know the signals that drive the process, the source of the new cells, or the pathway by which they are generated. In mice, adult β-cell regeneration depends predominantly on the proliferation of preexisting β-cells rather than neogenesis (52), although in certain forms of pancreatic damage, neogenesis of β-cells through a neurogenin-3-expressing stage does occur (51). Humans, however, may use the two pathways, neogenesis and proliferation, quite differently (1). In any case, the cell of origin for this neogenesis pathway, whether a differentiated adult cell or a professional stem cell, is not known, although in mice adult duct cells, which share certain characteristics of secondary pancreatic progenitor cells, can contribute to the pathway (18).

Islets vs. β-cells for regeneration therapy?

β-Cells exist in the highly specialized microenvironment of the islet, with other β- or non-β-cells as their immediate neighbors, in contact with extracellular matrix components deposited by themselves and/or endothelial cells, richly irrigated by the islet microvasculature and innervated. Contacts between islet cells (53) and with the extracellular matrix (54,55) impact directly on β-cell function, survival, and replication. There is close coordination and cross talk between the different islet endocrine cell types, with profound consequences for normal glycemic control. In particular, postprandial glycemia is contained within physiological limits by the combination of increased insulin and decreased glucagon secretion. Ironically it has been recognized since 1970 that type 2 diabetes is the consequence of dysfunction of the islet as a microorgan, with insufficient insulin and excessive glucagon acting together to drive hyperglycemia (56). Despite new evidence for a direct effect of insulin on glucagon secretion (57), the jury is still out regarding exactly how β-cells speak to α-cells to regulate glucagon secretion (i.e. via secretion of insulin, Zn²⁺ or γ-aminobutyric acid); however, there is no doubt that such communication occurs and is important for coordinated islet hormone secretion (58). Finally, neural regulation of both insulin and glucagon secretion was demonstrated more than 30 yr ago (59). Even disregarding the additional complexity of the known cross talk among other islet cell types, it thus seems reasonable to anticipate notable differences in the function of β-cells on their own and those within islets. However, equivalent glycemic control has been demonstrated in diabetic mice transplanted with aggregates of pure β-cells or islets (60), even if no such equivalence has ever been demonstrated clinically.

Quality control and safety

β-Cells derived in vitro will have to be characterized completely before any clinical trials. The first step would involve phenotypic characterization, with comprehensive gene profiling and a complete battery of in vitro functional tests to monitor the dynamics of insulin secretion in response to glucose and other secretagogues. This would be followed by preclinical testing in vivo, first in small animals and next in nonhuman primates. This sounds perfectly straightforward, but in reality there are serious issues preventing clean interpretation of the data. Although we know much about the β-cell, it may hold more secrets. It will be hard to decide exactly which functional aspects of this highly specialized cell are the most essential (61). What is the gold standard that surrogate β-cells must match to be given the go for clinical testing? Given the known differences between rodent and human β-cells on the one hand and between β-cells within an islet vs. pure cells on the other, a preparation of pure primary human β-cells will be the only valid comparator. Whereas it is feasible to obtain a highly enriched population of human β-cells sorted from other islet cell types (62), there are many uncontrollable variables that may impact the quality and performance of these cells, thereby compromising their validity as the comparator vs. surrogate β-cells. These would include the age, health, and cause of death of the donor; warm and cold ischemia time of the donor pancreas; and the isolation procedure itself (i.e. collagenase digestion and harvesting of islets).

Next, the cells will have to be tested in vivo. It is obvious that the first tests will have to be performed in rats or mice, and since 1972 we have certainly gained a lot of experience from studies using transplanted islets (5). However, glucose homeostasis differs from one species to another, and considerable care will have to be taken in extrapolating the data of glucose tolerance tests from rodents to man as discussed previously in greater detail (61).

Any attempt to treat individuals by regeneration therapy must of course first meet the strictest standards of safety. Even though this cannot be the major focus here, it must be recognized that all approaches raise major concerns with regard to safety. Any therapy based on increasing β-cell mass and function must have robust built-in measures to avoid an overshoot, with possible uncontrolled oversecretion of insulin and life-threatening hypoglycemia. Ensuring that newly regenerated or implanted β-cells are close to perfect in terms of differentiated function will be part of this safety measure. If the approach chosen is to make new cells, it is vital that we ensure that self proliferation is in check to avoid excessive insulin production with resultant hypoglycemia. If the chosen approach is to implant β-cells derived from ES or induced pluripotent stem (iPS) cells, there is the inherent risk of teratoma: no remaining pluripotent cells should be present in the implanted cell population. Indeed, the successful in vivo differentiation in mice of β-cells derived from human ES cells was accompanied by a high incidence of teratomas (40).

Complex requirements for insulin release by the replacement of β-cells in vivo

The development and use of β-cell replacement therapy, no matter what form it may take, will require the ultimate of the surrogate cells: they will have to secrete insulin in a regulated manner while coping with ever-changing circumstances. These shifting situations will require the cells to not only increase insulin output in response to nutrient ingestion but also to decrease output appropriately as glucose levels decline.

The β-cell’s response is larger when nutrients are presented orally than when administered iv (63). This difference, known as the incretin effect, results in part from the concomitant release of glucagon-like-peptide-1 and glucose-dependent insulinotropic peptide (64). Thus, the replacement cells will need to incorporate more than a single nutrient response system, and in turn these will need to act in concert to ensure appropriate timing of insulin release.

Whereas many may feel the critical aspect of β-cell function that will need to be reproduced is simply the release of insulin, one cannot overemphasize the vital importance of the timing of insulin exocytosis. It is now very well documented that the insulin response has to occur early if glucose is to be metabolized normally (65,66). Thus, when this early response is deficient, impaired glucose tolerance or diabetes result (65,66). Interestingly, under circumstances when glucose intolerance exists, the β-cell will frequently release more insulin late after nutrient presentation, but the timing is inappropriate and release is occurring only because the glucose levels have become excessively elevated. Thus, the replacement cells will have to respond rapidly if they are not simply going to lower glucose but do so efficiently and maintain glycemia well within the normal range. Last but by no means least, this need for a rapid and precise on switch for insulin secretion must be complemented by an equally responsive and dynamic off switch in response to falling glycemia (which would require intact cellular mechanisms responding to inputs such as catecholamines) that will ensure a prompt decrease in β-cell secretion if hypoglycemia is to be avoided.

Adapting to changing β-cell secretory demand: the need for differing β-cell function and mass

The ability of the β-cell to respond to stimulation is also critically modulated by differences in insulin sensitivity (67). Insulin resistance requires increased insulin output both in the basal state and in response to stimulation to maintain normal glucose tolerance, whereas improvements in insulin sensitivity place the β-cell in the position of having to reduce insulin release to avoid hypoglycemia. These changes in insulin sensitivity that require adjustment of insulin output can occur quite rapidly or over longer periods of time. The mechanisms responsible for these changes clearly vary and involve changes in both β-cell function and β-cell mass, although in most instances it appears that functional changes predominate. As discussed in greater detail previously (61), this flexibility will be an absolute and critical requirement of the replacement β-cells when glucose uptake increases over the short term, as occurs with exercise (68), or decreases rapidly, as an acute illness develops (69). In addition to functional adaptation to such rapid changes in insulin sensitivity, the β-cell must also alter its activity when this critical modulator changes for more prolonged periods. Under such conditions one envisages both β-cell secretory function and β-cell mass playing complementary roles. In pregnancy, during which there is an absolute need for increases in both basal and stimulated insulin release, the rather limited assessment of β-cell mass in humans suggests this to be increased (70), a finding well demonstrated in animals (71). However, function is also increased and typically beyond what the predicted increase in mass could be responsible for (72). Furthermore, the rapid decline in insulin demand that occurs immediately after parturition (73) underscores an additional critical flexibility the replacement β-cells will need to manifest in this instance.

In healthy subjects obesity alters the character of β-cells by requiring them to increase both their function and mass. These functional and morphological changes have been recognized for many years (74,75), with current estimates being that in obese individuals β-cell function is increased on average about 3-fold (74,76), whereas mass may be enhanced by only 20–40% (2,75). Here again the changes are complementary, with the functional aspect being more rapidly regulated as demonstrated by the prompt decline in insulin release after bariatric surgery, well before marked weight loss has occurred (77). What mediates these adaptive changes is uncertain; however, a number of possibilities exist including alterations in plasma levels of glucose, fatty acids, or incretin hormones and changes in neural tone. Whereas the exact mechanism remains unclear, the replacement cells will have to be responsive to these substrates and peptides. If central nervous system control is a critical aspect, then the site of placement of these cells will have to ensure that neural signals reach them and that they are responsive.

Keeping the replacement cell and the recipient out of dangerous territory

The source of replacement cells will also in many ways dictate processes that could handicap cell function and possibly hamper their survival. Cells obtained from sources other than the recipient can be expected to require immunosuppression therapy. Individuals with type 1 diabetes will be confronted with both autoimmunity and allo- or xenorejection, whereas those with type 2 diabetes will be spared from autoimmunity. Immunosuppression is problematic because it can be associated with insulin resistance (78) and can have an inhibitory effect on insulin release (79,80), thereby making it even more difficult for the replacement cells to maintain optimal glucose control. Of additional importance is the observation that use of certain of these medications has been associated with an increased risk of malignancy [although proven only after kidney transplantation (81)], raising another specter altogether.

With type 1 diabetes, there may be value in monitoring T cell reactivity and some of the antibodies typically associated with this process such as insulin autoantibodies and glutamic acid decarboxylase, islet cell and IA-2 antibodies (82,83). Recurrence and/or increasing titers of these antibodies may signal activity of the β-cell destruction process. With the current focus on immunotherapy to prevent the onset of type 1 diabetes in at-risk individuals (84), it is quite possible that advances in this area could also be applied to prevent disease progression in individuals who manifest antibodies after cell replacement therapy. Whether use of encapsulation will provide sufficient protection from immune assault and allow adequate β-cell function remains an important and unanswered question. It has recently been shown that it is possible to derive β-cells using iPS cells from an individual with type 1 diabetes (85). Whereas such cells will surely be useful for studies on the etiology and pathophysiology of type 1 diabetes, until methods have been developed for their protection, it seems likely that they would be targeted for rapid autoimmune destruction if implanted into that same individual.

Should cell replacement therapy prove a viable option for type 2 diabetes and especially if cells can be manufactured through manipulation of stem cells, the approach will surely be evaluated for potential widespread application. The concept of β-cell replacement therapy for type 2 diabetes has been more controversial because a substantial proportion of these patients are not insulin dependent. We know that β-cell replacement in the form of pancreas transplants can normalize glucose levels (86). Thus, transplantation of sufficient numbers of β-cells will also predictably reverse diabetes in these patients. However, if the replacement cells were to be derived from the host, there are two confounding issues. First, the identified genes for type 2 diabetes are predominantly associated with β-cell dysfunction (87,88). Second, patients with the disease have reduced numbers and function of β-cells (89), the pathogenic basis of which has not yet been clearly delineated but that may be related to alterations in these so-called β-cell genes. Thus, implantation of replacement cells derived from an individual with type 2 diabetes may well not lead to sustained glucose control over time because of a decrease in either or both of these parameters. This is akin to the situation in type 1 diabetes discussed above, although the autoimmune-based destruction of host-derived ß-cells would be predicted to occur more rapidly.

Advances in genetics have clearly identified a number of genes to be associated with reduced β-cell function and/or mass in type 2 diabetes (90), which raises questions about whether such cells might secrete less insulin per cell and be lost to apoptosis at a faster rate. It will be interesting to learn whether the differential effects of medications believed to slow the progression of β-cell disease (91) will also prove beneficial for replacement cells. Another concern is islet amyloid, which occurs in the majority of patients with type 2 diabetes (92) and has recently been shown to occur in transplanted human islets (93) and in those derived from a transgenic mouse model of islet amyloid (94).

Conclusion and Other Considerations

There have been impressive advances in the derivation of β-like cells from human ES cells that provide great hope that such an approach may one day offer a plentiful source of β-cells for transplantation (95,96). There continues to be remarkable progress in this field, including the identification of small molecules that promote the differentiation of human ES cells into pancreatic progenitor cells (97). There is also great excitement surrounding iPS cells (98) that may provide an alternative to ES cells for in vitro derivation of β-cells (99). The field has exploded since the original publication in 2006 (100), and most recently human iPS cells were produced by direct delivery of proteins into somatic cells without the need to introduce the corresponding genes (101). Importantly, iPS cells offer the tantalizing prospect of generating patient-specific β-cells, and they would also circumvent political, religious, or moral objections to use of ES cells that we shall not dwell on further here. In situ regeneration of islets may be a distant dream, but progress in the understanding of the development of the endocrine pancreas and the mechanism of adaptive β-cell regeneration (102) has provided a paradigm shift. Transdifferentation may offer an interesting alternative approach, even if formidable obstacles still lie on the road to safe and controlled islet cell regeneration using either approach.

The clinical requirements being demanded of the replacement β-cells are challenging. They will need to adapt to ever-changing needs from their host, some of these being more rapid than others. Under some circumstances a change in β-cell mass may be required, and this clearly needs to occur in a controlled manner whether it be increasing or decreasing. Furthermore, the host and replacement cells have to work in concert to reduce the likelihood that β-cells will be lost due to a recurrence of the destructive process that required their replacement in the first place. One could argue that if this were to occur, it may be possible to simply top up the β-cell reservoir. However, whereas this may seem and in reality be easy, we cannot begin to understand the possible long-term implications of doing so.

Of course, in addition to being as safe if not safer than existing therapies, islet cell regeneration therapy should provide a significant improvement in the management of the disease and its complications. Ideally, it should be readily available to all in need, regardless of where they live or their socioeconomic status. Quite understandably, individuals with diabetes and their families are increasingly impatient for better treatment and ultimately a cure. Improved communication of science has increased public awareness of major advances in biomedical research and demand for rapid translation of new discoveries from bench to bedside. Here too diabetes is at the forefront and individuals with this disease are almost always mentioned as likely beneficiaries when a breakthrough in stem cell research is communicated to the general public. This all provides fertile ground for unscrupulous medical practice: stem-cell based therapy is today’s snake oil and diabetes is taking center stage in this modern age drama. Direct-to-consumer advertising drives stem cell tourism (103), feeding off patients’ suffering and frustration in the absence of controlled investigation. The treatment that is offered has to be set against a background of peer review and clinical trials in accordance with guidelines for transition of stem cell therapy to clinical applications (104,105), rather than relying on hearsay and unverifiable reports from patients.

In conclusion, there are many reasons to believe that islet cell replacement or regeneration therapy will become a clinical reality, in full respect of good clinical practice and providing true benefit to the patient, so reducing the burden of diabetes on the individual and society.