Tim Kieffer

Diabetes is one of the deadliest diseases worldwide. While cadaveric islet cell replacement therapy offers a solution to the disease, the low number of donors severely limits the widespread adoption of this treatment. Here, we developed a scalable 7-stage differentiation protocol, including multilayered static cell culture vessels and vertical-wheel bioreactors, to generate human pluripotent stem cell (hPSC)-derived insulin-producing cells. Furthermore, we investigated stage-specific cell attributes and process parameters of the protocol. During the first four stages of differentiation, the cells utilized a high percent of the medium glucose (50-90%) and produced high lactate concentrations (7-17 mM). These attributes did not depend on the protocol, cell format or cell line. The cells between stages 5-7 had far lower glucose consumption (0-21%) and lactate production (0.3-2 mM). The oxygen consumption rate (OCR) during the first four stages was comparable to hPSCs. The OCR increased by stage 7, but was lower compared to human islets. Thus, during endocrine differentiation from PSC to pancreatic progenitors, there was a gradual energetic metabolic shift in metabolism from predominantly glycolytic to increased oxidative phosphorylation. The cells maintained a viability of >80% throughout the process and expressed the expected stage-specific markers. Although the cell numbers increased during the first four stages, during the period of endocrine differentiation (stages 5-7) there was a gradual decrease in proliferative cell numbers and substantial cell losses. The generated insulin-producing cells responded appropriately to several secretagogues, including high glucose. To de-risk aggregation steps in the bioprocess that could impede scale-up, we evaluated the effects of five bench-scale culture platforms and inoculum concentrations on hPSC aggregate formation and culture. Furthermore, we determined the feasibility of replacing AggreWell™ plates with PBS-Minis during pancreatic progenitor aggregation. Finally, we developed and characterized a fully aggregate cell differentiation suspension culture protocol from hPSC to insulin-producing cells. The initial seeding concentration impacted differentiation towards pancreatic progenitors. Collectively, our work helps establish an operating range of process parameters and cell quality attributes to enable the scalable manufacture of hPSC-derived insulin-producing cells.

View record

Diabetes is a debilitating disorder that affects nearly 500 million people worldwide and is characterized by chronic hyperglycemia caused by pancreatic β-cell dysfunction or death. Those that are affected need to frequently monitor their blood glucose levels and administer exogenous insulin, but this strategy does not match the tight regulation maintained by healthy β-cells. Progress in the field has demonstrated an effective clinical path to treat type 1 diabetes through the success of islet transplants. More recently, groups have demonstrated that human pluripotent stem cells, differentiated into pancreatic progenitors or beyond, can prevent or reverse diabetes in rodents and have shown promising results in human clinical trials. Here I develop a differentiation protocol to further differentiate pancreatic progenitors generated with a commercially available kit into endocrine cells. I identified FGF7 signalling as being important for aggregate survival and Notch-inhibition improved endocrine commitment at the expense of ductal cells. Combining in vitro β-cell differentiation strategies with newly developed precision gene editing techniques allows for disease modelling in a dish and transgene elements can be introduced into cells to provide new functionalities. To this end, I used a novel gene editing technique to correct a single-base mutation in pluripotent stem cells derived from an individual living with a severe form of neonatal diabetes. Using stem cell-derived cell therapies does not come without risk, given the potential for teratoma outgrowth, inappropriate differentiation and function, or an immune response to the graft. Here, I generated cell lines containing an inducible safety-switch, which was shown to eliminate pluripotent and pancreatic progenitor cells in vitro and teratomas in vivo. Finally, I targeted the B2M locus for silencing using an inducible CRISPRi platform, preventing HLA-ABC formation in a reversible manner. This strategy could both allow cells to avoid immune detection as well as act as a safety-switch by deactivating CRISPRi and re-expressing the B2M gene. Overall, the projects described in this dissertation demonstrate the combined potential of pluripotent stem cells, gene editing, and cell differentiation with the goal of advancing the cell therapy field, particularly for diabetes.

View record

Diabetes affects over 425 million worldwide, costs billions, and causes morbidity and mortality for patients. Though insulin injections are lifesaving, insufficient β-cell mass and function leaves patients facing risks of chronic hyperglycemia and acute risks of hypoglycemia. Replacement of β-cells via transplantation of cadaveric islets is a functional cure but is limited by a paucity of donor tissue. If β-cell replacement or (re)generation therapies were abundantly available, they could be potential cures for diabetes. To this end, investigating β-cell development and function is worthwhile. In the current thesis, we first characterized the role of insulin on β-cell development and maturation by studying insulin knockout mice (Ins1-/-Ins2-/-). Though insulin was necessary for β-cell maturation, insulin replacement by islet transplantation but not insulin injection, supported maturation of endogenous β-cells. Second, we developed and characterized an adeno associated virus (AAV) carrying Cre recombinase regulated by an insulin promoter (AAV Ins1-Cre) for in vivo genetic manipulations. AAV Ins1-Cre produced efficient recombination in β-cells alongside off-target recombination, making it a useful tool when off-target effects are controlled for or deemed unimportant. Third, we assessed the viability of a gene therapy for the Ins1-/-Ins2-/- mouse model of monogenic diabetes. We delivered an insulin gene to β-cells (an Ins1 promoter driving human insulin (INS) or mouse insulin 1 (Ins1)) using AAV Ins1-INS or AAV Ins1-Ins1. Though the AAV delivered the insulin gene to β-cells, Ins1-/-Ins2-/- β-cells retained a processing defect leading to secretion of insulin’s precursor proinsulin. We created adult insulin knockout mice using AAV Ins1-Cre and failed to prevent onset of diabetes with AAV Ins1-Ins1. Finally, in Chapter 5 we assessed the production of mature insulin in human β-cells. Despite consensus on the role of prohormone convertase 2 (PC2) in proinsulin processing, we provide evidence that unlike mouse β-cells, human β-cells produce mature insulin without PC2. This thesis provides insight into the developmental impact of “removing insulin” from β-cells, assesses the viability of a gene therapy “replacing insulin” in β-cells, and revises a longstanding theory on the “processing of proinsulin” in human β-cells. These findings may guide development of gene- and cell- based therapies for diabetes.

View record

Diabetes is characterized by insufficient amounts of insulin, a hormone produced by the beta-cells in the pancreatic islets. While current treatment options such as insulin injections provide a way to manage blood glucose levels, these methods provide suboptimal glucose control. Islet transplantation shows significant therapeutic promise, but the lack of available islets is an obstacle for this therapy. Restoration of beta-cell mass through cell therapy or stimulation of endogenous regeneration to replace beta-cell loss and dysfunction could potentially cure diabetes. Hence, investigating beta-cell development and regeneration is extremely valuable. Because zebrafish and mammalian pancreas development are highly conserved, we utilized zebrafish as a model to study beta-cell development and regeneration in the work described in this thesis. We developed and characterized a lineage tracing model that allowed us to differentiate beta-cells arising from the dorsal and ventral pancreatic buds. We lineage traced the dorsal bud derived beta-cells during development until the adult stage. While we observed that dorsal bud derived beta-cells constitute a small percentage of the total beta-cells in the adult pancreas, we did not identify transcriptional differences between beta-cells that arise from dorsal and ventral pancreatic buds, suggesting that developmental origin does not dictate transcriptional profile. Second, we characterized the role of islet vessels during beta-cell formation. We found that islet vessels are dispensable for alpha-cell and beta-cell development. Finally, we determined the cell sources important in beta-cell maintenance and regeneration. We observed that beta-cell proliferation is the main mechanism of beta-cell maintenance in the adult zebrafish, but that a non-beta-cell source may contribute to beta-cell regeneration. By providing a better understanding of zebrafish beta-cell development and regeneration, the findings in this thesis may help guide cell therapies and regenerative strategies for diabetes.

View record

The hormone leptin reduces food intake and increases energy expenditure; leptin deficient rodents and humans are severely obese with impaired glucose homeostasis, and leptin therapy reverses these metabolic abnormalities. Based on this, leptin gained significant interest as a potential therapeutic agent to combat obesity; however, it was soon discovered that most obese humans are resistant to its anorectic actions and the initial excitement over the therapeutic applicability of leptin dwindled. It is now evident that leptin regulates glucose metabolism independent of its actions on body weight. Underlining this, leptin displays the remarkable ability to reverse hyperglycemia in rodent models of type 1 diabetes. These findings sparked interest in leptin as a therapy for type 1 diabetes. The overarching goal of this thesis was to perform pre-clinical studies to elucidate the mechanism by which leptin lowers blood glucose levels in insulin deficient mice and assess whether resistance to these effects of leptin can occur. To this end, we assessed the plasma metabolomic profile of leptin-treated insulin deficient mice, which revealed global alterations in amino acid metabolism. Thus, we characterized amino acid utilization in leptin-treated mice, tested the mechanistic role of amino acid catabolism genes, and extensively characterized the metabolic changes in the liver using an 'omics' based approach. In addition, we assessed whether recapitulating leptin-mediated changes in the liver using a small molecule can mimic the glucose lowering actions of leptin. Lastly, we assessed whether hyperleptinemia or high-fat intake, which are reported to cause resistance to the weight reducing actions of leptin, also impede glucose lowering effects of leptin. This thesis revealed that leptin globally alters the liver metabolic profile to suppress the utilization of amino acids for glucose production and recapitulating the transcriptomic profile of leptin therapy with a novel small molecule can lower blood glucose levels in insulin deficient mice. Furthermore, dietary fats, but not hyperleptinemia, causes resistance to the glucose lowering actions of leptin in insulin deficient mice. Collectively, these investigations help elucidate the mechanism by which leptin reverses hyperglycemia in insulin deficient mice and shed insight into the suitability of leptin as a therapy for diabetes.

View record

Background: The adipose-derived hormone leptin is well known for its effects on metabolic homeostasis, and in particular its role in suppressing appetite and regulating fat mass. Loss of leptin results in a phenotype characterized by obesity, insulin resistance, hyperinsulinemia and glucose intolerance. Treating leptin deficient mice with leptin is able to lower blood glucose and insulin levels without reducing body weight, indicating that leptin has independent effects on glucose homeostasis. While there have been many studies exploring the role of leptin in metabolism, the glucoregulatory actions of leptin remain to be fully elucidated. Scope of Thesis: The overall goal of this thesis was to understand the role of leptin in glucose regulation. To do so, we utilized a number of in vivo models in which leptin signalling was disrupted. First, we investigated the effects of reducing insulin gene dosage on glucose homeostasis and body mass in leptin deficient ob/ob mice. Next we explored the significance of leptin signalling in pancreatic β cells. To determine whether the glucoregulatory actions of leptin are conserved across rodent species, we explored the effects of leptin deficiency in rats, and used this model to investigate the hierarchy of metabolic defects that arise as a result of impaired leptin action. Finally, we describe a novel mouse model to selectively knock out leptin in a temporal and tissue-specific manner. Conclusions: We show that reduction of circulating insulin levels prevents obesity, but results in hyperglycemia in ob/ob mice. Furthermore, we show that direct leptin signalling in β cells is insufficient to restore euglycemia. In addition, we demonstrate that leptin deficiency produces a similar metabolic phenotype in rats as in mice. Finally, we suggest that loss of insulin sensitivity may be a primary metabolic defect that results from leptin deficiency in rats. Together, these results provide further insight into the glucoregulatory mechanisms of leptin that may be used to develop therapeutic strategies to treat obesity and diabetes.

View record

It has long been thought that the only hormone capable of reversing the catabolic consequences of diabetes is insulin. However, various studies have demonstrated that the adipocyte-derived hormone leptin can potently lower blood glucose levels in rodent models of insulin-deficient diabetes. In addition, the hormone glucagon is elevated in type 1 and type 2 diabetes, and glucagon suppression therapy has shown promise as an agent to treat diabetes in mice. Given the interest in both of these therapies, leptin treatment and glucagon antagonism were propelled into clinical trials for patients with type 1 and type 2 diabetes respectively. The overarching goal of this thesis was to perform preclinical studies to investigate the mechanism of the glucose lowering actions of leptin and the effects of glucagon suppression therapy in mouse models of diabetes. To achieve this, we probed the role of increased leptin action as a result of insulin therapy, determined the function of insulin-like growth factor binding protein-2 (IGFBP2) in the glucose lowering actions of leptin, investigated the necessity of insulin for leptin treatment and glucagon suppression therapy, and explored the potential of glucagon suppression therapy via glucagon receptor (Gcgr) small interfering ribonucleic acid (siRNA) delivered by lipid nanoparticle (LNP) technology. This thesis reveals that elevated leptin levels may contribute to the glucose lowering effect of insulin therapy in insulin-deficient diabetes. In addition, we demonstrate that physiological levels of IGFBP2 are neither sufficient nor required for the action of leptin on glucose homeostasis. Moreover, leptin can normalize many metabolic parameters in the complete absence of insulin, but blood glucose levels are volatile and the length of survival is finite. Furthermore, Gcgr siRNA can improve many diabetic symptoms in mouse models of type 1 and type 2 diabetes. Finally, we report that the metabolic manifestations associated with a complete lack of insulin cannot be overcome by Gcgr gene inactivation. Collectively the findings in this thesis contribute insight into the mechanism of action, and the therapeutic potential of leptin administration and glucagon suppression therapy as a treatment for diabetes.

View record

Diabetes mellitus results in elevated blood glucose levels due to an insufficiency of the glucose lowering hormone insulin. In type 1 diabetes, insulin loss is due to an autoimmune destruction of the insulin producing pancreatic β-cells. One treatment for type 1 diabetes is the transplantation of cadaveric islets, although this process is limited by a lack of donor tissue. We and others are examining factors that influence the generation of new β-cells from stem cells. Specifically, the aims of this thesis were to examine the role of transcription factors in the formation of endocrine cells from stem cells. To do this, we developed a high content screening approach in human amniotic fluid stem cells to assess the effect of six pancreatic transcription factors on insulin expression. From this screen and subsequent studies, we observed that while transcription factor overexpression was capable of driving insulin expression, the resulting cells were unable to reverse diabetic hyperglycemia upon transplantation. Given the well-established developmental capacity of human embryonic stem cells (hESCs), we next characterized a novel hESC line (CA1S), which is amenable to high throughput screening and pancreatic differentiation. Using these cells, we found that the number of cells seeded into a culture system had a significant effect on the formation of endodermal, pancreatic progenitor and pancreatic endocrine cells. This effect correlated with hESC cell cycle status and resulted in the formation cells co-expressing insulin, glucagon and somatostatin. We next examined the effects of the transcription factors PAX4 and ARX on pancreatic endocrine specification. We revealed that increased PAX4 expression reduced ARX and glucagon expression leaving insulin positive cells. Reduced ARX expression by genomic editing resulted in fewer glucagon positive cells and increased PAX4 levels. Loss of ARX was also associated with an abundance of somatostatin positive cells and a partial reduction in insulin, which was rescued with re-expression of ARX adenoviral gene delivery methods. Collectively, the data presented in this thesis emphasise the role of transcription factor expression as a primary control point for the possible generation of β-cells from stem cells, which represents a potential cellular therapy for type 1 diabetes.

View record

Type 1 diabetes is a life-long disease, with devastating consequences and no cure. While the hormone insulin has been the only effective treatment to date for type 1 diabetes, emerging evidence has revealed that the fat-derived hormone leptin can also exert profound glucose lowering effects, and reduce mortality in type 1 diabetes. This has generated interest in the therapeutic potential of leptin as an anti-diabetic treatment, and propelled leptin into clinical trials for type 1 diabetes. The fact that leptin monotherapy (without insulin administration) can so potently lower blood glucose in insulin-deficient diabetes is surprising, given that for almost a century, insulin has been assumed to be the only hormone that can lower blood glucose in type 1 diabetes. The overarching goal of this thesis was to perform pre-clinical studies to elucidate the mechanism of the anti-diabetic effects of leptin in type 1 diabetes. To this end, we thoroughly assessed the changes in metabolic and energy homeostasis that occur in a mouse model of type 1 diabetes receiving leptin therapy. The roles of hepatic and neuronal leptin receptor signalling in the anti-diabetic action of leptin were also investigated, through tissue specific disruption of leptin signalling using the Cre-lox method. In addition, we assessed whether leptin therapy can serve as an adjuvant to islet transplantation therapy in type 1 diabetes. This thesis revealed that leptin therapy lowers blood glucose in a mouse model of type 1 diabetes, correlating with decreased hepatic glucose production and enhanced insulin sensitivity. The anti-diabetic action of leptin is not blunted in mice with disrupted neuronal leptin signalling outside of the arcuate and ventromedial hypothalamus, or in mice with disrupted hepatic leptin signalling, suggesting that leptin acts through alternate or redundant pathways to lower blood glucose. Finally, low-dose leptin administration dramatically enhanced the ability of islet transplants to restore euglycemia, suggesting that leptin and islet co-therapy could be a successful therapeutic strategy for type 1 diabetes. Collectively the findings in this thesis contribute insight into the mechanism of action, and the therapeutic potential of leptin as a treatment for type 1 diabetes.

View record

Incidences of obesity and type 2 diabetes have risen worldwide at alarming rates. While there is an undeniable correlation between obesity and type 2 diabetes, a clear mechanistic link between these two conditions has not been fully elucidated. The adipocyte-derived hormone leptin may play a role linking obesity and type 2 diabetes. Leptin can regulate body weight through its effects on the brain to decrease food intake and increase energy expenditure, but these effects are disrupted in obesity. Interestingly, leptin also has effects on glucose and lipid metabolism independent of its effects on body weight, so it is possible that disrupted leptin signalling during obesity can also perturb glucose and lipid metabolism, leading to symptoms associated with type 2 diabetes. Since the liver plays a critical role in integrating and controlling glucose and lipid metabolism, it was hypothesized that leptin resistance in the liver could play a role in the development of diabetic symptoms. To investigate this hypothesis, three complementary mouse models were used to help identify the role of leptin signalling specifically in the liver. It was found that in lean mice, hepatic leptin resistance results in increased insulin sensitivity in the liver, leading to reduced hepatic glucose output but also increased lipid accumulation and secretion of larger, more triglyceride-rich very low density lipoprotein (VLDL) particles without an increase in total plasma triglycerides. In obese, hyperinsulinemic mice lacking hepatic leptin signalling, the effects of lost leptin signalling on triglyceride metabolism were exacerbated, resulting in decreased triglyceride clearance and elevated plasma triglycerides compared to controls. These effects on plasma triglycerides were reversed when hepatic leptin signalling was restored in a mouse model of total leptin resistance. Collectively, these data reveal a possible role for hepatic leptin resistance in the development of diabetic symptoms during obesity.

View record

Adenosine is precursor and a metabolic intermediate of adenosine triphosphate in energy transfer, and cyclic adenosine monophosphate in signal transduction. Recent studies have demonstrated that the role of adenosine in the body is much more than just structural as it can also behave as an important regulator of homeostatic functions. Adenosine signalling relies on the activation of the A₁, A₂A, A₂B and A₃ adenosine receptors. Through the development of pharmacological tools and genetic knockout mouse models of specific receptor subtypes, the involvement of these receptors in various physiological systems is quickly being established. This thesis investigates the function of adenosine in the digestive system and specifically how adenosine regulates the release of gastric and pancreatic peptides. With the use of a novel vascularly perfused isolated mouse stomach model and specific A₁ and A₂A receptor knockout animals, the role of adenosine on the release of somatostatin and ghrelin was determined. Lower concentrations of adenosine can inhibit the release of somatostatin and ghrelin via the activation of A1 receptors, while higher concentrations can stimulate their release via activation of A₂A receptors. Given the importance of somatostatin in regulating gastric acid secretion and motility, and ghrelin in regulating systemic energy balance, better understanding of how the release of these two peptides is regulated may reveal potential therapies for eating disorders, gastrointestinal dysfunctions and metabolic diseases.In the pancreas, adenosine was shown to regulate both insulin and glucagon secretion from the pancreatic islets. Studies presented in this thesis demonstrate that adenosine signalling interacts with the effects of the incretin hormone GLP-1 in the pancreas such that concomitant administration of adenosine and GLP-1 in the perfused pancreas induced greater insulin release than GLP-1 administration alone. Furthermore, A₁ receptor knockout mice exhibited more frequent pulses of insulin secretion, which may have contributed to their superior glucose tolerance compared to wild type control mice. These findings on the role of adenosine signalling in the pancreas may have implications in the etiology of diabetes mellitus. The involvement of adenosine signalling in the digestive tract further illustrates the importance of adenosine as a metabolic regulator in homeostasis.

View record