Towards a Functional Cure for Diabetes Using Stem Cell-Derived Beta Cells: Are We There Yet?

Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Associated Data

Abstract

Diabetes mellitus is a pandemic metabolic disorder that results from either the autoimmune destruction or the dysfunction of insulin-producing pancreatic beta cells. A promising cure is beta cell replacement through the transplantation of islets of Langerhans. However, donor shortage hinders the widespread implementation of this therapy. Human pluripotent stem cells, including embryonic stem cells and induced pluripotent stem cells, represent an attractive alternative beta cell source for transplantation. Although major advances over the past two decades have led to the generation of stem cell-derived beta-like cells that share many features with genuine beta cells, producing fully mature beta cells remains challenging. Here, we review the current status of beta cell differentiation protocols and highlight specific challenges that are associated with producing mature beta cells. We address the challenges and opportunities that are offered by monogenic forms of diabetes. Finally, we discuss the remaining hurdles for clinical application of stem cell-derived beta cells and the status of ongoing clinical trials.

1. Introduction

Pancreatic beta cells are critical regulators of blood glucose homeostasis by their unique ability to produce and secrete insulin in response to changing blood glucose levels. Because insulin is the only hormone that is able to decrease blood glucose levels, its release into the bloodstream must be strictly controlled in order to prevent blood glucose levels from surpassing dangerously low or high levels. The loss of beta cell function results in diabetes mellitus, a group of metabolic disorders that are characterized by chronically elevated blood glucose levels. Diabetes mellitus has reached epidemic proportions globally and it currently affects over 463-million people [1]. Although its two predominant forms are type 1 (T1D) and type 2 diabetes (T2D), rare forms of diabetes, including monogenic types that require specific attention, are increasingly diagnosed. T1D is marked by absolute insulin deficiency following autoimmune-mediated beta cell loss [2], while T2D is caused by relative insulin deficiency due to beta cell dysfunction, often in the context of peripheral insulin resistance [3]. Monogenic forms of diabetes result from single gene mutations and they are characterized by beta cell dysfunction, to varying degrees of severity [4]. Because no real cure exists for diabetes, daily insulin injections remain the standard of care for patients with T1D, late-stage T2D, and for a subset of patients with monogenic diabetes. Although this treatment is lifesaving, it conveys a chronic and costly burden of care, a persisting risk for acute and chronic complications, and it still results in an overall decreased life expectancy.

Beta cell replacement holds the potential to truly cure T1D and also possibly T2D and monogenic diabetes. Such cell therapy—through percutaneous infusion of pancreatic islets into the portal vein—is currently applied in some patients with brittle T1D [5], providing prolonged insulin independence. In selected patients, beta cell replacement proves to be superior to insulin administration with regard to overall metabolic control, prevention of severe hypoglycemia, and delaying the progression of micro- and macrovascular complications [6,7,8,9]. Despite this proof-of-principle for beta cell replacement as a genuine cure, donor islet transplantation is unattainable for the vast majority of diabetic patients for several reasons. First, donor islets are in short supply, which contrasts with the global disease prevalence. Donor shortage is further aggravated by a loss of up to half of grafted cells in the first few days after transplantation (reviewed in [10]). Delayed graft revascularization is one of the leading causes for this loss, since, following transplantation, islets enter a poorly vascularized and hypoxic microenvironment [11,12] that compromises islet cell and, in particular, beta cell survival and function. Excessive numbers of beta cells must be grafted to compensate for this early post-transplant cell loss, which further aggravates donor scarcity. Second, the allogenic origin of donor islet grafts necessitates lifelong immunosuppression, thereby increasing the susceptibility to infections and tumorigenesis (reviewed in [13]).

This donor islet shortage has fueled the search for alternative beta cell sources. Recent advances in differentiation protocols have positioned human embryonic stem cells (ESCs) and human induced pluripotent stem cells (iPSCs) as a promising and theoretically unlimited beta cell source. Currently, encapsulated pancreatic endoderm cells, differentiated from human ESCs, have already entered the first clinical trials (NCT03162926 (completed), NCT03163511 (recruiting), NCT02239354 (active, not recruiting), and NCT02939118 (enrolling by invitation)) and some preliminary results have been disseminated [14]. In this review, we define some key characteristics of adult beta cells and call attention to heterogeneity within the beta cell population. Next, we discuss the use of stem cells as an alternative beta cell source for transplantation. We elaborate on the current status of stem cell differentiation protocols in order to generate beta cells, the cues that are needed for functional beta cell maturation, and a number of hurdles that still need to be overcome. We address the following hurdles in detail: (i) the limited functional maturation of in vitro generated beta cells, (ii) the difficulties of graft survival upon transplantation due to the immune response and delay in graft revascularization, and (iii) safety concerns, especially regarding residual immature cells that may give rise to teratoma formation. Finally, we highlight the potential of gene-editing to generate patient- or disease-tailored beta(-like) cells for personalized medicine, discuss ongoing clinical trials, and offer some perspectives to the field.

2. Beta Cell Maturation and Heterogeneity

2.1. The Mature Beta Cell

Because the ultimate goal of stem cell differentiation protocols in diabetes research is to generate mature, fully functional beta cells, it is important to define what constitutes a functionally mature beta cell. The foremost fundamental characteristic of a beta cell lies in its ability to secrete appropriate amounts of insulin in response to glucose. Beta cells have a specialized machinery for fulfilling this role. In brief, glucose uptake by beta cells is mediated by specialized low affinity glucose transporters (mainly GLUT1 in human). Upon entry, glucose is phosphorylated by glucokinase (GCK) and is used to generate ATP through oxidative phosphorylation. This rise in ATP increases the ATP/ADP ratio, which, in turn, leads to the closure of ATP-sensitive potassium (K_ATP) channels and cell membrane depolarization. The latter electric signal then triggers the opening of voltage-dependent calcium channels (VDCCs), leading to Ca 2+ influx, resulting in the release of insulin granules through exocytosis [15,16]. Subsequently, beta cells generate a second, even larger, wave of insulin release (known as biphasic glucose-stimulated insulin secretion) that is K_ATP-channel independent, but rather beta cell metabolism-dependent [17,18,19]. Mature beta cells can generate such responses—in a uniquely sensitive way—to minor physiological variations in plasma glucose levels.

Testing the maturity of beta cells with functional assays that interrogate the specialized cellular machinery that is implicated in glucose sensing and insulin secretion while using physiologically relevant variations in glucose concentrations is cumbersome, time-consuming, and often difficult. Therefore, cell markers have been used as surrogate measures for beta cell maturation. Typical beta cell maturation markers include the transcription factors GLIS3 [20], MAFA [21], NEUROD1 [22], NKX6.1 [23], PAX6 [24], PDX1 [25], SIX2 [26], and UCN3 [27,28]. Notably, besides positive regulators, beta cell maturity is equally determined by the absence of ‘disallowed’ or ‘forbidden’ genes that interfere with beta cell function including Ldha, Mct1, SLC16A1, Hk1, Hk2, and Rest (reviewed in [29]). The disallowed genes are upregulated under mild hyperglycemia as deduced from scRNA-seq data of beta cells isolated from diabetic rats, pointing to beta cell dysfunction occurring already in the early stages of diabetes [30]. In addition, one of the most important aspects of adult beta cells is their post-mitotic nature. Beta cell mass expansion mainly occurs shortly after birth by a process that rapidly declines with age [31,32], to reach an extremely low proliferation rate of 0.1–0.4% in the adult human [33,34]. Mathematical modeling of the accumulation of lipofuscin bodies in human beta cells suggests that, after the age of 20 years, long-lived beta cells age with the body [35,36].

2.2. Heterogeneity in the Adult Beta Cell Population

A next level of complexity with regard to the adult beta cell population is its heterogenous nature. The concept of beta cell heterogeneity was developed in the 1990s following metabolism-centered studies of dissociated cell populations, wherein some beta cells showed differential glucose-responsiveness [37,38] based on differences in glucokinase activity [39]. Recent studies unveiled particular underlying molecular hallmarks to provide compelling examples of beta cell heterogeneity. For example, Johnston et al. identified two distinct beta cell populations, ‘hub’ or ‘leader’ cells (90% of beta cells) upon insulin secretion challenges. Leader cells display high-potential mitochondria, high glucokinase expression, reduced insulin content, and lower expression of the beta cell markers PDX1 and NKX6.1 [40]. Bader et al. identified two distinct beta cell subpopulations that are based on the expression of Flattop (Fltp), a Wnt/planar cell polarity (PCP) effector, and reporter gene. Fltp - beta cells are more proliferative and able to expand under conditions of increased metabolic demand, such as pregnancy, whereas Fltp + beta cells are more mature, and they display higher insulin secretory capacity and increased mitochondrial function [41]. van der Meulen et al. identified immature beta cells in a neogenic niche at the adult islet periphery that lack UCN3 and are derived from transdifferentiated alpha cells [42]. Because these UCN3-negative beta cells only constitute~1.5% of the total beta cell population, they are likely distinct from the more abundant leader or follower cells and from the Fltp + and Fltp − beta cells, since the latter express UCN3 [41].

3. General Properties of Stem Cells

Stem cells are clonal cells with the potential to both self-renew and differentiate into a variety of functional somatic cells [43]. Stem cells are classified into three main types, according to their origin, potential, and developmental stage, as: (i) ESCs—which are isolated from the early mammalian embryo, (ii) adult stem cells—which can be found in various adult tissues (e.g., neural, hematopoietic, mesenchymal, and epidermal stem cells), and (iii) iPSCs—which are derived from adult cells that are reprogrammed back to an embryonic-like pluripotent state.

ESCs are highly undifferentiated cells that are isolated from the inner cell mass of mammalian blastocysts, i.e., early embryonic cells. ESCs can, on the one hand, proliferate indefinitely while maintaining their pluripotency and on the other differentiate into cells of all three germ layers [44]. In 1998, Thomson et al. established ESC lines from human blastocysts [45]. Because human ESCs can be maintained in culture for extended periods of time and can be differentiated into any desired target cell type, they represent a promising cell source for regenerative medicine to treat a host of diseases, including Parkinson’s disease, spinal cord injury, and diabetes [45]. However, even if functional cells and organs can be generated from ESCs, their transplantation remains subject to allograft rejection in the same manner as conventional donor organ transplants. Because human ESCs are derived from surplus human embryos, their application in regenerative medicine also raises ethical concerns that can be addressed by the use of iPSCs that are generated by reprogramming somatic cells, such as peripheral blood mononuclear cells or dermal fibroblasts, into the pluripotent state through the overexpression of a defined set of transcription factors (Oct4, Sox2, Klf4, and c-Myc) [46]. Such a straightforward method to produce human iPSCs directly from patients’ own cells opens the possibility of studying disease and screening drugs in vitro in a patient-specific manner. In addition, the inherent autologous nature of iPSCs provides them with unique immunological advantages over other cell sources in the context of cell therapy.

4. Current Status of Human ESC and iPSC Differentiation Protocols

The discovery of ESCs and iPSCs has opened the possibility to generate cells or tissues in vitro that can be used for the study of disease mechanisms, drug screening, and cell replacement. Regarding the latter, these cells promise an unlimited source of virtually any cell type that can be transplanted into patients, including pancreatic beta cells for people with diabetes.

4.1. Generation of Stem Cell-Derived Beta Cells

The main approach for differentiating stem cells into beta cells is by adherent cell culture with progressive, stepwise lineage commitment while using combinations of cues added to the culture medium. Using such an in vitro lineage differentiation approach, D’Amour et al. were the first to succeed in robust induction of definitive endoderm differentiation [47], with subsequent generation of pancreatic endocrine hormone-producing cells [48]. However, the endocrine cells that were generated were mainly polyhormonal (e.g., insulin and glucagon co-expressing cells) that are more akin to immature islet cells [49]. The resulting insulin-expressing cells also lacked the essential beta cell transcription factors NKX6.1 and PDX1 [50]. Over the ensuing years, strategies have been devised and optimized in order to generate monohormonal insulin-expressing cells that co-express NKX6.1 and PDX1 by modifying the composition and timing of growth factor and small molecule addition [51,52,53]. In 2014, Rezania et al. [53] and Pagliuca et al. [52] reported the successful generation of functional stem cell-derived beta-like cells that possessed many beta cell-specific traits, including glucose-responsive insulin secretion. Importantly, the transplantation of these cells was able to reverse diabetes in mice. However, the beta-like cells that were generated by these protocols [52,53] and follow-up studies [51,54,55,56] still displayed poor glucose-induced insulin secretion when compared to human islets.

More recently, Nair et al. [57] and Velazco-Cruz et al. [58] succeeded in generating functional human stem cell-derived beta cells showing dynamic glucose-stimulated insulin secretion that is similar to human islets. Nair et al. reaggregated immature human stem cell-derived beta-like cells into enriched-beta cell clusters after fluorescence-activated cell sorting (FACS) and demonstrated that the reaggregation/clustering is paramount to the generation of functionally superior beta cells with robust dynamic insulin secretion in vitro [57]. Mechanistically, the clustering of beta-like cells induced metabolic maturation by driving mitochondrial oxidative respiration, which is central to stimulus–secretion coupling in mature beta cells. This strategy not only increased the structural resemblance with native islets, but also the functional resemblance, both at the cellular (transcriptomic) and cluster (functional) level [57]. Velazco-Cruz et al. mainly focused on permitting TGF-β signaling during the final stage of differentiation (from endocrine progenitor to beta-like cell) and, additionally, on controlling cellular cluster size [58]. Although the TGF-β pathway is claimed to play a critical role during beta cell differentiation from stem cells [59], its role in beta cell function and insulin secretion remains controversial, as TGF-β signaling suppresses insulin transcription and reduces insulin protein levels and secretion [60], while others posited that TGF-β signaling is required for maintaining beta cell mass and regulating insulin secretion [61,62]. Further studies on TGF-β signaling in beta cells are needed, but, with respect to the differentiation of highly functional beta cells from stem cells, TGF-β fulfills a dual role, as its inhibition is required at early stages while its signaling is beneficial at late stages. Recently, Yoshihara et al. demonstrated that non-canonical WNT4 signaling drives the metabolic maturation of beta-like cells—essential for robust insulin secretion—in large part through the induction of an ERRgamma gene network [63], while Li et al. developed a beta cell differentiation protocol that is based on three previously published protocols [52,53,64] to obtain beta cells with high glucose-responsiveness and insulin production [65].

In addition to maturity, the purity of beta cells that are differentiated from human pluripotent stem cells must be taken into account for in vitro differentiation protocols, since a higher percentage of differentiated cells implies a lower percentage of contaminating immature and therefore possibly teratogenic cells. Sorting strategies while using a GFP reporter under the control of the insulin gene promoter or an antibody to the beta cell surface marker CD49a enabled obtaining 80–90% pure beta cells [57,66]. Alternatively, the cell-surface marker CD9 can be exploited for negative selection to enrich for glucose-responsive human beta-like cells [65]. These strategies may also prove to be useful for sorting other hormone-producing cells, such as alpha cells, which contribute to disease etiology by elevating blood glucose levels [67,68,69]. See Table 1 for a summary of the approaches for the generation and purification of stem cell-derived beta cells.

Table 1

Overview of protocols for the generation of stem cell-derived beta cells.

ROCK: Rho-associated protein kinase, TGF-β: transforming growth factor β, FACS: fluorescence-activated cell sorting, WNT4: wingless-type murine-mammary-tumour virus integration site family member 4, CD9: cluster of differentiation 9.

4.2. Generation of Non-Beta Pancreatic Endocrine Cell Types

Co-transplantation with other pancreatic endocrine cell types will likely benefit transplantation outcomes, as the release of insulin is regulated via complex paracrine interactions (discussed in more detail in the next chapter). To this end, efforts are ongoing for generating non-beta pancreatic endocrine cell types. In 2011, the directed differentiation of human ESCs into glucagon-positive cells, expressing the key alpha cell transcription factor ARX was reported [73]. Although these alpha-like cells secreted glucagon in vitro to some extent, proper regulation was lacking, which is indicative of their immature state [73]. More recently, Peterson et al. developed a differentiation protocol for the generation of alpha cells that express and secrete glucagon in response to low glucose and some glucagon secretagogues, and that elevate blood glucose levels upon transplantation in mice [74]. Efforts to develop differentiation protocols for somatostatin-producing delta cells, ghrelin-producing epsilon cells, and pancreatic polypeptide cells, and cell surface antibody-based sorting strategies will further contribute to the generation of islet-like clusters that will provide a unique resource for studying cell biology, disease modelling, drug screening, and cell replacement therapy.

5. Current Hurdles for Stem Cell Therapy and Possible Ways to Tackle Them

5.1. Functional Immaturity of Stem Cell-Derived Beta Cells

Beta cells reside within the pancreatic islets of Langerhans, being clustered together along with alpha, delta, epsilon, and pancreatic polypeptide cells. Beta cells intensively interact with a variety of cell types in their microenvironment, including the other endocrine and non-endocrine cells within the islet, but also with exocrine cells outside the islet. The numerous signals that beta cells receive are crucial for their functionality, survival, differentiation, and proliferation. First, beta-to-beta cell interactions are of major importance for a properly coordinated and synchronized insulin secretory response and for insulin gene expression, storage, biosynthesis, and release [75,76,77]. Paracrine cell interactions between different endocrine islet cell types are also paramount in fine-tuning the islet secretory response. [78,79]. Indeed, glucagon is a well-known regulator of insulin secretion, whereas somatostatin inhibits both insulin and glucagon secretion [27]. The importance of inter-endocrine cell-to-cell contacts for beta cell maturation is demonstrated by the enhanced maturation of human stem-cell-derived beta cells upon in vitro mimicking of cell clustering [57]. Therefore, the transplantation of islet-like clusters, rather than just beta cells alone, is likely to benefit glycemic control. Besides interactions with other endocrine cells, beta cells also receive signals from the extracellular matrix [80], endothelial cells [81], pericytes [82], neurons [83,84], and immune cells [85].

All of these cues from the islet microenvironment are essential for beta cell maturation. Notably, differentiation does not equal maturation, as Potten and Loeffler stated in 1990: “Differentiation can be defined as a qualitative change in the cellular phenotype that is the consequence of the onset of synthesis of new gene products, that lead ultimately to functional competence. Maturation in contrast can be regarded as a quantitative change in the cellular phenotype or the cellular constituent proteins leading to functional competence” [86]. This notion implies that, even when stem cell-derived beta cells express all known adult beta cell markers and produce high levels of insulin, they are not per se functionally mature, since the proof of functionality should be sought in functional tests that dynamically interrogate the beta cell glucose sensing and insulin secretory machinery.

It is very difficult, if not impossible, to provide all of the signals that beta cells receive from their microenvironment through simplified in vitro differentiation protocols, which is why current protocols rely on in vivo maturation for beta cells to become functional [52,53]. An alternative approach seeks not to generate mature cells, but uses early definitive endoderm or pancreatic progenitor cells and relies strongly on self-directed differentiation and maturation in order to obtain a functional stem cell-derived beta cell mass in vivo [70,87,88]. This is the strategy that was also chosen in the ongoing clinical trials, as mentioned in the introduction. However, while using this approach, it will take several months for the beta cells to become functional and this black box of in vivo maturation does not allow for an easy dissection of its mechanisms. Augsornworawat et al. performed single-cell RNA sequencing on human iPSC- and ESC- islet grafts that were transplanted for six months in diabetic mice and compared their gene expression profiles to stage 6 [89] ungrafted hPSC- islets and cadaveric human islets. The analyses confirmed that transplanted stem cell-derived beta cells possessed insulin secretory capacity and acquired expression of the beta cell maturation markers INS, G6PC2, MAFA, MNX1, SIX2, and UCN3 [90]. Such studies provide a comprehensive resource for understanding human beta cell maturation and to improve differentiation strategies.

Ensuring quick islet revascularization following transplantation will likely benefit beta cell survival and functionality. Endothelial cell-derived signals are crucial for embryonic beta cell development [81,91] and for adult beta cell proliferation, survival, differentiation, and function [81]. Hypovascularization is not only one of the factors contributing to islet cell loss following transplantation, but it may also lead to beta cell dedifferentiation, since intraportal implants of human primary beta cells lose their mature phenotype and expression of key beta cell markers [92]. Strategies for improving graft revascularization have focused on the delivery of pro-angiogenic factors, primarily vascular endothelial growth factor A (VEGF-A), to islet cells. VEGF-A is a key angiogenic factor that is secreted by beta cells to promote endothelial cell migration, proliferation, and survival, and to regulate vascular permeability [81]. VEGF-A protein can be delivered in many ways, but one particularly interesting approach is by the transfection of Vegfa mRNA. Not only is this approach much safer when compared to viral-vector based gene delivery methods, its inherent short-term expression is beneficial as compared to long-term expression [93], which is detrimental for islet function and survival [94]. Recent work from our group showed that liposome-mediated transfection of human and mouse islet cells with synthetic modified Vegfa mRNA improves graft revascularization and increases beta cell mass [95].

Taken together, signals from the islet microenvironment are crucial in ensuring proper beta cell maturation and functionality. The translation of these insights to stem cell differentiation protocols is expected to improve transplantation outcomes.

5.2. Risk of Tumorigenesis

The therapeutic potential of pluripotent stem cells is vast, and it promises to transform medicine. The capacity of stem cells to self-renew and differentiate into any desired cell type underlies this promise, but these desirable features also bring along dangers, such as a risk of developing tumors. Therefore, several methods and tools have been developed for guaranteeing safety in therapeutic applications of human ESCs and iPSCs. First, tumorigenesis can be prevented by transplanting only differentiated cells. By using techniques, such as endocrine cell clustering and targeting specific signaling, recent differentiation protocols aim to generate higher percentages of mature functional beta cells and minimize the number of undifferentiated progenitor cells, as discussed above. Thus, protocol optimization in terms of improved differentiation minimizes the risk for tumorigenesis. Second, approaches for eliminating or sorting out undifferentiated human pluripotent stem cells in vitro have been developed, such as the use of chemical inhibitors [96], immunological targeting of undesired cell types [97,98,99], or the introduction of suicide-genes in the stem cell genome [100,101].

Chemical eradication of undifferentiated stem cells is possible by adding small molecules to the culture medium to selectively kill pluripotent cells. A high-throughput screen identified inhibitors of the key oleate biosynthesis enzyme stearoyl-CoA desaturase (SCD1) as agents that specifically compromise stem cell viability [96]. Knowing that SCD1 is abundantly expressed in iPSC-beta cells and required for beta cell identity, the effect of these SCD1 inhibitors on beta cell function warrants further investigation [102]. The immunological targeting of undifferentiated stem cells relies on the use of (cytotoxic) antibodies that are directed against specific stem cell markers to either kill them or separate them from differentiated cells. SSEA-5 glycan, for example, is a human ESC surface marker that has been used in order to selectively remove undifferentiated teratoma-forming cells prior to transplantation [97]. Gene editing can also aid in the removal of undifferentiated human pluripotent stem cells. In an innovative approach to selectively remove tumorigenic cells, two suicide gene cassettes were introduced in ESC-derived beta cells. The first safety cassette was a herpes simplex virus thymidine kinase (HSV-TK) that was driven by the telomerase gene promoter that is selectively active in undifferentiated cells, the second a nitroreductase (NTR) flanked by two loxP-sites, which is removed upon Cre-expression that is driven by the human insulin gene promoter. HSV-TK and NTR are sensitive to, respectively, ganciclovir and CB1954, which enables the elimination of tumorigenic undifferentiated cells at any desired moment, both in vitro and in vivo [100]. Finally, engineered human iPSCs with an inducible Caspase-9 suicide gene have been developed, in which the small molecule chemical inducer of dimerization can effectively induce apoptosis in >99% of the cells [101]. While these methods are innovative and promising, they are not a validated approach for guaranteeing safety after clinical transplantation. See Table 2 for an overview of the possible approaches to prevent teratoma formation.

Table 2

Approaches to prevent teratoma formation of stem cell-derived grafts.