Therapeutic Strategies Targeting Pancreatic Islet β-Cell Proliferation, Regeneration, and Replacement

Abstract

Diabetes results from insufficient insulin production by pancreatic islet β-cells or a loss of β-cells themselves. Restoration of regulated insulin production is a predominant goal of translational diabetes research. Here, we provide a brief overview of recent advances in the fields of β-cell proliferation, regeneration, and replacement. The discovery of therapeutic targets and associated small molecules has been enabled by improved understanding of β-cell development and cell cycle regulation, as well as advanced high-throughput screening methodologies. Important findings in β-cell transdifferentiation, neogenesis, and stem cell differentiation have nucleated multiple promising therapeutic strategies. In particular, clinical trials are underway using in vitro–generated β-like cells from human pluripotent stem cells. Significant challenges remain for each of these strategies, but continued support for efforts in these research areas will be critical for the generation of distinct diabetes therapies.

Introduction

Pancreatic Islets of Langerhans and the Burden of Diabetes Mellitus

Pancreatic islets of Langerhans are key regulators of glucose homeostasis containing insulin-producing β-cells, glucagon-producing α-cells, somatostatin-producing δ-cells, and to a lesser extent pancreatic polypeptide (PP)–producing cells and ghrelin-producing ε-cells (1). Islets release insulin and glucagon, among other hormones, in response to a variety of metabolic, endocrine, and neuronal inputs to maintain glucose homeostasis (2). Diabetes mellitus is characterized by an impairment in the production or response to insulin, leading to chronic hyperglycemia and subsequent comorbidities, including cardiovascular disease, retinopathy, nephropathy, and neuropathy (3). In type 1 diabetes mellitus (T1D), the accepted view is that β-cells are destroyed via autoimmunity. This immune attack and loss of β-cell mass can occur in juveniles, adolescents, or adults (4, 5). While the exact causes of T1D are not understood, certain genetic variations can predispose individuals to T1D and environmental factors can trigger autoimmune responses (6, 7). For example, infections with certain viruses have been associated with increased T1D risk (8); however, this idea is still under investigation (9). It is also possible that the β-cell itself instigates its own autoimmune attack (10, 11). In type 2 diabetes mellitus (T2D), genetic and environmental circumstances contribute to β-cell dysfunction and peripheral insulin resistance (12-14). During this process, β-cells are thought to compensate for insulin resistance by increasing their mass and insulin output. Indeed, genome-wide association studies indicate that many of the genes associated with T2D risk are involved in β-cell function (15-18). As the disease develops, β-cell compensation eventually fails and there is loss of functional β-cell mass leading to T2D (19, 20). The resulting hyperglycemia has long-term deleterious effects in T1D and T2D patients, including microvascular and macrovascular complications (21).

In 2021, it was estimated that 537 million adults (≥18 years old) worldwide had diabetes, and projections indicate that by 2045 this number is expected to be near 783 million adults worldwide (22). Furthermore, medical care for those with diabetes presents a significant economic burden. Recent reports indicate that the cost of medical care to those with diabetes is rising in the United States. In 2017, the total cost dedicated to diabetes in the United States was estimated to be $327 billion (23). The rising cost of diabetes is increasing at a greater rate than other diseases, suggesting its treatment is becoming more expensive. From 1987 to 2011, the excess cost of care for a patient with diabetes essentially doubled when compared with a patient without diabetes (24).

Current Diabetes Therapies

While currently there is no cure for diabetes, patients with T1D and late stage T2D are treated using an exogenous subcutaneous insulin regimen to maintain glycemic control (25). Although life-saving and effective, subcutaneous insulin treatments have limitations. When compared with endogenous insulin secreted by pancreatic islet β-cells, recombinant exogenous insulin treatments are inferior in terms of the temporality and fine-tuning involved in the precise regulation of glycaemia (26). For instance, delivery of too much insulin may result in hypoglycemic episodes, which can be life-threatening (27). Prior to development of insulin dependence in T2D, a variety of pharmaceutical treatments exist to help lower insulin resistance and mitigate or delay endogenous β-cell failure. Such therapies include intensive insulin therapy (28), metformin (29), sulfonylureas (30), thiazolidinediones (31), glucagon-like peptide 1 receptor (GLP1R) agonists (32), dipeptidyl peptidase 4 inhibitors (33), and sodium/glucose cotransporter 2 inhibitors (30); many of these therapeutics have been reviewed in detail (30, 34). Artificial pancreas technologies such as closed-loop systems using glucose sensors and insulin pumps are also among strategies currently pursued (35). There are also multiple other pharmaceutical targets and approaches under investigation for T2D including glucokinase activators, GPR40 agonists, and alternative incretin axis stimulators (36). In T1D, major areas of focus for therapeutics include prevention of β-cell loss or replacement of β-cell mass, and, to this end, advances in immunotherapies are promising. Teplizumab is an anti-CD3 monoclonal antibody designed to bind CD3 at the cell surface and consequently activate the T cell receptor/CD3 signaling (37). Activation of the T cell receptor in this way can improve self-tolerance, preservation of regulatory T cells (38). Teplizumab slowed the reduction of β-cell function in new-onset T1D patients for 2 years after clinical diagnosis (39). Additionally, teplizumab delayed time to diagnosis by 24 months in a phase II trial conducted in individuals at high-risk for developing T1D (40). A decision from the FDA is anticipated in late 2022 on the approval of teplizumab for use in T1D patients (41). A second immunotherapy, golimumab, is human antitumor necrosis factor (TNF)α monoclonal antibody designed to abrogate interaction of TNFα with its receptor, thereby preventing inflammatory signaling downstream of the TNFα receptor (42). Golimumab is an FDA-approved for use in multiple autoimmune diseases including rheumatoid arthritis and ulcerative colitis (43). Although golimumab was recently shown to improve endogenous β-cell function in recently diagnosed T1D patients (44), it has not yet been approved by the FDA for use in T1D.

Another approach is the use of anti-interleukin (IL)-21 antibody in combination with the GLP1R agonist, liraglutide. This combination strategy was recently used in a clinical trial in individuals recently diagnosed with T1D (45). GLP1R agonists are known to mitigate β-cell apoptosis under stress (46) and IL-21 promotes recruitment of cytotoxic CD8⁺ T cells to pancreatic islets (47). The anti-IL-21 and liraglutide combination significantly blunted the T1D-related decrease in β-cell function after 1 year of treatment. There is also recent evidence that oral administration of the voltage-dependent Ca²⁺ channel blocker verapamil can delay T1D progression by up to 2 years and may also improve outcomes in combination therapy in T2D (48, 49). The relative contribution of cell types and pathways involved in the effects of verapamil are not yet clear, but may be due to a combination of effects on β-cells and immune cells.

In addition to methods which protect endogenous β-cells, the Edmonton protocol is an islet transplantation method that has been used to successfully treat T1D for extended durations (Fig. 1) (50). This method is useful to a subgroup of T1D patients with refractory hypoglycemia. However, the number of islet transplants performed annually is small due to limited donors, specific selection criteria, and damage during processing (51). Methods to enhance viability and function of islet tissues in vitro prior to transplantation or to promote engraftment and protect against immune destruction in vivo have also been investigated (52, 53). Examples of these approaches include antibodies against the proinflammatory cytokine interferon-γ-induced protein 10 (54), inhibitors of proapoptotic pathways (eg, caspases, JNK), catalytic antioxidants (55, 56), and anti-inflammatory treatments (α-1 antitrypsin (57, 58)). A variety of G protein–coupled receptor (GPCR) pathways also represent potential targets given their protective properties in β-cells exposed to inflammatory cytokines, including agonism of GLP1R (32, 59-61) and cholecystokinin receptors (62), modulation of neuropeptide Y receptors (63, 64), or antagonism of proinflammatory GPCRs like GPR31 (65). In a complementary approach to these methods, new approaches focused on replacing or regenerating endogenous β-cells have been under active development with potential applications in both T1D and T2D patients.

Figure 1.

Current and future strategies for β-cell replacement therapies. (A) Remaining endogenous β-cells may be coaxed into self-renewal via replication-inducing drugs targeted in a cell type–specific manner. Pharmacological stimulation of β-cell neogenesis or transdifferentiation from progenitor or islet endocrine cells, respectively, is also under investigation. (B) Currently, donor human islet tissue can be transplanted into the liver through the Edmonton protocol. There is potential for expansion of islet endocrine or progenitor cells in vitro using newly discovered small molecules or hormones. Differentiation or transdifferentiation of these expanded cell populations into β-like cells could provide a supplemental source of transplantable tissues. New trials are underway utilizing stem cell–derived pancreatic endoderm or β-like cell clusters either in a ‘naked’ format transplanted via the liver portal vein, or contained within macroencapsulation devices.

Regenerative medicine is aimed at replacing or regenerating human cells with the goal of restoring or establishing normal function (66). A major goal in diabetes research is aimed at applying these methods to replace or replenish lost β-cells by promoting their proliferation or regeneration either in vivo or ex vivo, or by generating de novo β-cells from induced pluripotent stem cells. In the following sections we will begin with the process of β-cell maturation to highlight the functional characteristics and identifying features of an adult β-cell, as well as the heterogeneity within mature β-cell populations. We will also provide an in-depth discussion of β-cell proliferation and regeneration, including recent advancements in this area and how these findings may lead to a successful treatment or cure for diabetes.

β-Cell Development and Maturation

Juvenile, Immature, and Mature β-Cells

Juvenile β-cells (in humans from 1-9 years of age) respond similarly to adult β-cells in glucose-stimulated insulin secretion assays, except that juvenile human islets release less insulin (67). Additionally, juvenile human β-cells respond normally to canonical insulin secretagogues in vitro including glucose and amino acids (68). Juvenile β-cells are therefore considered functionally mature after approximately the first year of development. Mature human β-cells have an increased threshold for glucose-stimulated insulin secretion and express maturation markers, including the gene urocortin 3 (69, 70). Immature β-cells lacking urocortin 3 persist in adult mice and humans within a neogenic niche in the islet (71). These virgin cells could potentially be harnessed for β-cell neogenesis. Transcription factors have been identified that are critical during β-cell maturation, such as NKX6-1 (72), as well as those unique to mature adult β-cells, including MAFA (67) and SIX2/SIX3 (73). β-Cell differentiation is dependent upon lineage-determining transcription factors, but the development of β-cell identity is not sufficient for glucose-stimulated insulin secretion (74). Glucose-stimulated insulin secretion is dependent upon signal-dependent transcription factors, which are regulated via environmental signals (74). Guided by lineage-determining transcription factors, signal-dependent transcription factors mediate the required fine-tuning to confer functional and metabolic properties of β-cells.

β-Cell heterogeneity

Adult β-cells are heterogeneous in their functions and abilities. Some adult β-cells may have greater capacity to secrete insulin than others, and this may change over time (75). Such functional heterogeneity has been described in the case of leader/first-responder β-cells (76, 77) and hub β-cells (78) in both mouse and human islets. In addition, there is heterogeneity in the proliferative capacity of adult β-cells. Based upon this notion, the defined “maturity” of a β-cell cannot be determined only by the amount of insulin expression or the presence of specific transcription factors. For instance, 1 subpopulation of β-cells is known as “hub”’ or “leader” cells. These cells contain a lesser amount of insulin, PDX1, and NKX6.1 when compared with other β-cells (78). Hub β-cells are under active investigation, and transcriptomic distinctions between hub/leader and follower cells implicated differences in primary cilia number and length may be involved. Furthermore, 2 other subpopulations of β-cells differ in functionality and were identified in mice marked by the Wnt/planar cell polarity effector Flattop. Flattop-negative β-cells have a greater proliferative ability, while Flattop-positive cells have better functional responses (79). During times of need, such as pregnancy, Flattop-negative β-cells can proliferate and change to Flattop-positive cells to handle the altered metabolic state (79). In times of physiological stress or disease, this variability within β-cell populations is imperative in order to maintain normoglycemia (75). However, the exact mechanisms underlying the origin and durability of different β-cell subpopulations is not fully understood (75, 78, 80-85), and further investigation is warranted.

Avenues to Formation of New β-Cells

The major paths to β-cell replacement include proliferation, transdifferentiation, neogenesis, replication, and stem cell–based approaches (86-88) (Fig. 1). In replication (also referred to as proliferation), a β-cell divides into 2 new β-cells (88). During transdifferentiation, other pancreatic cells, such as ductal cells (89) or islet endocrine cells (90), can convert into β-cells. In the process of neogenesis, new β-cells are created from undifferentiated progenitors (69, 71, 85). Although transdifferentiation and neogenesis describe different mechanisms, they encompass the same outcome, which is generation of new insulin-expressing β-cells from noninsulin-expressing cells (91). In taking advantage of β-cell neogenesis and replication, the therapeutic goal is to essentially re-enact the signaling mechanisms that occur during the development of the pancreas. β-Like cells can also be generated in vitro from either human embryonic stem cells (hESCs) or human induced pluripotent stem cells (hiPSCs). hESCs are uniquely able to indefinitely self-renew and to differentiate into tissues from all 3 germ layers (mesoderm, endoderm, ectoderm) (92). hiPSCs are patient-derived somatic cells reprogrammed into pluripotent cells. Stem cell–based strategies are performed in vitro with the goal of transplanting functional β-cells into humans (93). The following 3 sections will discuss how recent findings in these areas are being developed toward diabetes therapies.

β-Cell Proliferation/Replication

Overview of Cell Cycle Regulation and β-Cell Replication

The cell cycle is a tightly controlled process. By analogy, the cyclin-dependent kinases (CDKs) can be thought of as engines of the cell cycle and the cyclins as the gas pedal, which is either engaged, to power the engine through the cell cycle, or disengaged, allowing the engine to remain in an idle state. First discovered in yeast as Cdc2 (94), CDK1 was later identified in humans in 1987 (95). CDKs drive the cell cycle through their kinase activity and are regulated by the binding of distinct cyclin proteins (96). In β-cells, the cell cycle and proliferation are regulated by many factors including, but not exclusively, nutrients (eg, glucose, amino acids), growth factors (eg, IGF-I, lactogens), and GPCRs; subsequent signaling including Akt, mTOR, MAP kinases, cAMP-PKA; and downstream effects on transcriptional regulators such as FoxO1, STAT proteins, CREB, and β-catenin (for comprehensive review see (97-99)). In most cases of diabetes, including T1D, a small population of β-cells can persist (100). Considering this, it may be possible to stimulate the replication of these cells to treat diabetes. A major roadblock to this approach is that β-cells are postmitotic and have a low replication rate throughout adult life, estimated to be ∼0.1% in mice (101). Recent work using multi-isotope imaging mass spectrometry–electron microscopy indicated that many islet β-cells are as old as the animal itself (102). These findings are in agreement with earlier estimates that adult human β-cells are long-lived and the cell population is largely stable after 20 years of age in healthy individuals (103). In the fetal and neonatal stages of development, β-cells replicate rapidly, and as adulthood is reached β-cells become quiescent (104). However, adult β-cells do replicate under certain conditions, such as pregnancy and the initial compensatory response to obesity (105).

β-Cell Replication and Quiescence

Challenges in β-cell replication

There are multiple barriers to exploiting endogenous β-cell proliferation for diabetes therapy. One challenge is that in order for this process to be possible there must be pre-existing β-cells to proliferate or β-cell precursors to differentiate. The smaller the population of remaining β-cells in a patient with diabetes, the more difficult this treatment process will be. Another challenge is that there have been no clinical trials on humans to determine an optimal therapeutic rate of β-cell replication. For example, a patient with T1D with a small population of β-cells would most likely need to take a putative β-cell proliferative drug for much longer than a T2D patient who may still have half of their β-cells remaining (20). These therapeutic rates will need to be determined via clinical trials.

As mentioned previously, there is also the issue of β-cell heterogeneity, the existence of β-cell subpopulations, and their relevance to diabetes disease states (83, 106, 107). For instance, molecular markers like Flattop mark nonproliferative islet cells (79), urocortin 3 marks mature β-cells (70), and RBP4⁺ β-cells correlate with reduced function (108). Determining the specific subtypes of β-cells that have the ability to proliferate may be critical when populations of these cells are small, as in diabetes. Such knowledge would inform studies with proliferative drugs which could be most effective if they are designed to target those specific subtypes. Further, inflammation in T1D is not uniform throughout the islet and some regions may be free of inflammation (109). These findings support the idea that there may be a subset of cells which are resistant to immune killing. A 2017 study identified a subpopulation of β-cells during the progression of T1D in NOD (nonobese diabetic) mice that survived immune attack (110). These novel β-cells had stem-like features, were less differentiated, had lower expression of mature β-cell markers, and contained lower granularity (110). In line with this idea, the autoimmune destruction of human β-cells in T1D is not always complete, and some individuals with longstanding disease retain detectable insulin concentrations (111, 112). As researchers continue to identify and characterize β-cells in these systems, proliferative treatments may be more likely to succeed. A final issue which requires consideration is that any drug which induces proliferation also has the potential to cause cancer. Hence, it is imperative that these drugs be extremely specific and well tested.

Mechanisms underlying β-cell replication and quiescence

After maturation, β-cells eventually become quiescent (104). Knowledge of the mechanistic underpinnings of β-cell quiescence is important in the quest to determine methods to stimulate β-cell proliferation. Over the course of evolution, adaptations arose to prevent the proliferation of certain cell types (including β-cells) in adults. In the absence of these mechanisms, unregulated β-cell mass expansion can cause hyperinsulinemia and result in dangerous hypoglycemia. Examples of this can be found in endocrine cancers, such as insulinomas (113, 114), or in nesidioblastosis (115). Different pairs of CDKs and cyclins have been shown to induce β-cell proliferation (Cdk4-cyclin D1) (116), or to traffic to the β-cell cytoplasm (Cdk6-cyclin D3) (117, 118). When cyclin D3 (Ccnd3) and Cdk6 are overexpressed, they can traffic to the nucleus to cause phosphorylation of cell cycle inhibitors (eg, retinoblastoma protein Rb) and drive entry of β-cells into the cell cycle (118). Furthermore, mitogenic stimuli have been shown to “awaken” quiescent β-cells by means of controlling the expression and activation of specific cyclins and CDKs, allowing β-cells to re-enter the cell cycle. For example, GLP1R ligands can act as mitogens which promote β-cell proliferation and this has been shown to require epidermal growth factor (EGF) receptors (119-121). The effects of GLP1R agonism on proliferation were primarily observed in juvenile β-cells and required calcineurin/NFAT (nuclear factor of activated T cells) signaling to induce cyclin A (CCNA1) and CDK1 expression (122).

Development of Pharmacological β-Cell Protectants and Proliferation Inducers

Early efforts to stimulate β-cell proliferation via pharmacological methods were quite challenging, but the advent of high throughput chemical screening identified agents that stimulate human β-cell proliferation in vitro (123, 124). Small molecules modulating a variety of targets have been discovered to alter β-cell proliferation including the kinases DYRK1A, adenosine kinase, SIK, and glucokinase; receptors for transforming growth factor (TGF)β, EGF, insulin, glucagon, GLP1, and prolactin; and oligonucleotide-mediated depletion of cell cycle regulators (Table 1).

Table 1.

β-Cell protection, proliferation and regeneration modulating agents

Agent/chemical/factor	Agent family	Target(s) or Pathway(s)	Target class	Effect of agent at target	Target tissue(s)	Effects on tissue	References
Aminopyrazine (GNF7156, GNF4877)	Small molecule	DYRK1A, GSK3β	Kinase	Inhibition	β-Cell	↑ β-cell proliferation	(130)
6-Azaindole derivative (GNF2133)	Small molecule	DYRK1A, GSK3β	Kinase	Inhibition	β-Cell	↑ β-cell proliferation	(130)
CC-401	Small molecule	DYRK1A/B	Kinase	Inhibition	β-Cell	↑ β-cell proliferation	(130)
1,5-Naphthridine	Small molecule	DYRK1A	Kinase	Inhibition	β-Cell	↑ β-cell proliferation	(130)
Denosumab	Monoclonal antibody	RANKL	Cytokine	Inhibition	circulation	↑ β-cell proliferation	(190)
Osteoprotegerin	Endogenous humoral	RANKL, GSK3	Kinase	Inhibition	β-Cell	↑ β-cell proliferation	(190)
Harmine	Small molecule	DYRK1A	Kinase	Inhibition	β-Cell	↑ β-cell proliferation	(130)
5-Iodotubercidin (5-IT)	Small molecule	DYRK1A/adenosine kinase	Kinase	Inhibition	β-Cell	↑ β-cell proliferation	(130, 146)
Serpin B1	Endogenous humoral	proteases	Protease	Inhibition	β-Cell	↑ β-cell proliferation	(85, 167)
GLP1R agonist/DYRK1A inhibitor combination	Pharmacological co-treatment	GLP1R/DYRK1A	GPCR/kinase	Inhibition	β-Cell	↑ β-cell proliferation	(137)
HB-EGF	Endogenous humoral	EGFR	Kinase	Inhibition	β-Cell	↑ β-cell proliferation	(134)
WS6	Small molecule	EGFR/NFkB pathways	Kinase	Activation	β-Cell	↑ β-cell proliferation	(288, 289)
Glucose	Endogenous metabolite	glycolysis, IRS2, mTOR, cyclin D2	Metabolism	Activation	β-Cell	↑ β-cell proliferation	(290-293)
Glucokinase activators (eg, TTP399; Ro28)	Small molecule	glucokinase	Glycolysis	Activation	β-Cell, liver	↑ β-cell proliferation	(149-151, 294)
Placental lactogen; prolactin	Endogenous humoral	Prolactin receptor and growth hormone receptor (JAK-STAT signaling)	Kinase	Inhibition	β-Cell	↑ β-cell proliferation	(175, 295)
S961; OSI-906	Synthetic peptide	Insulin/IGF1 receptors	Kinase	Inhibition	Liver, adipose	↑ β-cell proliferation	(162-164)
ABT-702	Small molecule	Adenosine kinase	Kinase	Activation		↑ β-cell proliferation	(146)
A-134974	Small molecule	Adenosine kinase	Kinase	Inhibition	β-Cell	↑ β-cell proliferation
5′-N-ethylcarboxamide adenosine (NECA)	Small molecule	ADORA2A	Gpcr	Inhibition	β-Cell	↑ β-cell proliferation	(147)
HG-9-91-01	Small molecule	SIK1/2; RIPK3	Kinase	Inhibition	β-Cell	↑ β-cell proliferation (concentration-dependent)	(140-142)
GLP1R agonist (eg, exendin-4; liraglutide)	Synthetic peptide	GLP1R	Gpcr	Synergistic activation	β-Cell	↑ β-cell proliferation; β-cell protection	(46, 119, 121, 122, 296)
AS1842856	Small molecule	FoxO1	Transcription factor	Inhibition	Multiple	↑ β-cell transdifferentiation	(225, 226)
IGFBP1	Endogenous humoral	IGFR signaling	Kinase		α-Cell, β-cell	↑ β-cell transdifferentiation	(206)
Glucagon receptor antibodies	Antibody	GCGR signaling	Gpcr	Activation	Liver, α-cell (indirect)	↑ α-cell proliferation, β-cell transdifferentiation	(211, 215, 216)
GLP1-estrogen	Small molecule-linked peptide	GLP1R/ER	GPCR/nuclear receptor	Inhibition	β-Cell, brain	β-Cell protection	(195-197)
GLP1-ASOs	Oligonucleotide-linked peptide	CHOP	Transcription factor	Activation	β-Cell	β-Cell protection	(188)
Anti-IL-21/liraglutide	Peptide/antibody co-treatment	IL-21	Cytokine; GPCR	Inhibition; activation	Circulation; β-cell	β-Cell protection	(45)
Golimumab	Antibody	TNFα	Cytokine	Inhibition	Circulation	β-Cell protection	(42-44)
Verapamil	Small molecule	L-type Ca2+ channel	Ca2+ channels	Inhibition	Multiple	β-Cell protection	(48, 49)
sCCK-8	Synthetic peptide	CCKRB	Gpcr	Activation	β-Cell	β-Cell protection	(62)
Neuropeptide Y	Synthetic peptide	NPYRs	Gpcr	Activation/inhibition	β-Cell	β-Cell protection	(63, 64)
Anti-IP-10	Antibody	IP-10/CXCL10	Cytokine	Inhibition	Circulation	β-Cell protection	(54)
α-1 Antitrypsin	Purified protein	Proteases	Protease	Inhibition	Circulation	β-Cell protection	(57, 58)
BMS-986165	Small molecule	TYK2	Kinase	Inhibition	Multiple	β-Cell protection	(281)
Teplizumab	Antibody	CD3	T cell receptor	Activation	Regulatory T cells	↑ immune tolerance; β-cell protection	(37, 39-41)
Fabkin	Endogenous humoral	P2Y1	Gpcr		β-Cell	↓ β-cell function ↑ β-cell death	(148)
Insulin	Endogenous humoral	Insulin receptor signaling	Kinase	Inhibition	Liver, β-cell	Regulate β-cell mass in some models	(160, 291, 295, 297)
CID661578	Small molecule	MNK2	Kinase	Block binding to eif4g	β-Cell	↑ β-cell neogenesis from duct	(251)
(R)-DRF053; roscovitine	Small molecule	Cdk5	Kinase		Multiple	↑ β-cell neogenesis from duct	(232)
GABA	Endogenous small molecule	GABAAR; GABABR; metabolism	Ligand-gated Cl– channel, GPCR	Synergistic activation/inhibition	β-Cell	Variable observations in β-cell neogenesis	(178, 219-222, 224)

Different small molecules, peptides, proteins, antibodies, oligonucleotides, or combination therapeutics have been demonstrated to affect β-cell proliferation, regeneration/neogenesis, or to protect β-cell function or viability from stressors. The targets of these agents fall into a variety of classes including GPCRs, kinases, proteases, or other circulating factors. While some combinations of agents have been tested, most have been tested alone, leaving the possibility of synergy between multiple different agents targeted against independent pathways for selective β-cell effects.

Kinase targets in β-cell proliferation

DYRK1A

Inhibitors of DYRK1A (dual-specificity tyrosine phosphorylation regulated kinase 1A) have been widely studied (80, 125-128) and have been shown to stimulate the highest β-cell proliferation rates through inhibition of calcineurin/NFAT/DYRK1A signaling (125-127, 129). DYRK1A inhibitors include aminopyrazines, harmine, 5-iodotubercidin, GNF2133, 1,5-naphridine, and CC-401 (85, 130). In 2015, Wang et al used a high-throughput chemical screen to identify analogs of harmine that could be useful therapeutic drugs in the treatment of diabetes (125). This study identified the likely target of harmine to be DYRK1A and downstream signaling to NFAT as the mechanism by which it induced proliferation and differentiation of mouse and human β-cells. Harmine treatment caused an expansion of islet mass and improved glucose control in mouse models of diabetes (125). Harmine was further confirmed to induce human β-cell proliferation (127, 131). Harmine is also known to target multiple proteins, warranting the development of analogs that may improve specificity and potency (132, 133). A recent study further investigated the effects of harmine, glucose, and heparin-binding epidermal growth factor-like growth factor (HB-EGF) (134). β-Cells were stimulated by harmine and HB-EGF, but these molecules also both increased the proliferation of insulin-negative and glucagon-negative cells. Harmine also greatly stimulated the proliferation of α-cells. Proliferation due to the presence of glucose was dependent upon both the donor and the performed assay. This study also highlights the need for in-depth identification of cell types using multiple markers, as non-β-cells may be misidentified as β-cells (134). DYRK1A is not β-cell specific, but rather expressed widely and is known to be associated with other diseases (135). This is a safety issue that must be overcome with future research. One strategy is to exploit synergistic effects of multiple β-cell proliferation-inducing pathways. For example, GLP1R agonists have been attributed to the promotion of cell survival as well as reducing β-cell ER stress (46, 136). There is a potentiating effect when using DYRK1A inhibitors in combination with GLP1R agonists (137), as well as when combined with TGFβ pathway inhibitors (130, 138). This combination method also allowed for lower dosage of each agent, limiting unintended effects because these are not β-cell–specific pathways. This is a significant finding, considering that a deficiency, excess, or dysregulation of DYRK1A has been associated with central nervous system effects (139). Although effective at stimulating β-cell proliferation, it is not known whether DYRK1A inhibitors will be safe in humans. The utility of DYRK1A inhibition will be unclear until there are data on human safety and preclinical studies on animals (130).

Salt-inducible kinases

Among the 3 salt-inducible kinases (SIKs) (SIK1, SIK2, SIK3), SIK2 has been shown to be important for β-cell function in mice, as β-cell–specific Sik2 knockout animals become glucose intolerant due to impaired insulin secretion (140). Nevertheless, recent work from Charbord et al found pharmacological inhibition of SIKs to induce human β-cell proliferation (141). In a separate study by Iorio et al, human β-cell proliferation was shown to be induced via targeting the GPCR GPR3-SIK2 pathway (142). In their study, GPR3 silencing induced human β-cell proliferation by increasing SIK2 activity. Silencing of SIK2 only, but not SIK1 or SIK3, prevented increased proliferation rates in GPR3-silenced β-cells (142). Overexpression of SIK2 had the opposite effect. Interestingly, in the study by Iorio et al, the pan-SIK inhibitor HG-9-91-01 at 2 µM reduced proliferation caused by GPR3 silencing, while in the study by Charbord et al, the same inhibitor (but at 0.1-1 µM) was found to induce proliferation in zebrafish, mouse, and human β-cells; at 3 µM HG-9-91-01 had little impact on proliferation. It is possible these discrepancies are due to differential effects of SIK family members SIK1 (143) and SIK2 (140), acute sensitivity of β-cells to the concentration of HG-9-91-01, the potential polypharmacology of the drug, or different proliferation assay protocols (SV40 T-antigen expression and EdU⁺/Ins⁺ incorporation in Iorio vs Ki67⁺/Ins⁺ staining in Charbord). Indeed, HG-9-91-01 was also recently shown to inhibit the kinase RIPK3 at similar dose ranges (0.5-5 µM) (144), and RIPK3 is a critical regulator of TNFα-induced β-cell death in T1D mouse models (145). Future studies of the SIK family kinases and their related pathways will be important for progressing potential human β-cell therapeutics.

Adenosine kinase

Inhibitors of adenosine kinase were discovered as β-cell regeneration/proliferation inducers using a primary rat islet screening platform (146) and through a high-throughput in vivo zebrafish screen (147). In the rat islet screen, 2 adenosine kinase inhibitors were identified, 5-iodotubercidin and ABT-702, and were shown to act on β-cell proliferation via a nuclear isoform of ADK and in an mTOR-dependent manner (146). In the zebrafish screen, among the top hits were adenosine kinase inhibitor A-134974 and adenosine receptor agonist 5′-N-ethylcarboxamide adenosine (NECA). The authors focused on NECA which they determined to promote mouse β-cell proliferation/regeneration in an ADORA2A-dependent mechanism (147). Related to findings with adenosine kinase, a recent investigation discovered a circulating complex, named Fabkin, comprised of fatty acid–binding protein 4, adenosine kinase, and nucleoside diphosphate kinase (148). This complex was suggested to inhibit β-cell function and promote cell death, in part through alteration of ADP/ATP ratio near β-cells and modulation of signaling through purinergic receptors. Overall, targeting adenosine signaling pathway components appears to be a promising area of research which contributes to our understanding of β-cell proliferation and regeneration.

Glucokinase

Glucokinase has also been shown to be critical in β-cell proliferation using rodent models, where deletion prevented β-cell replication and provision of a small molecule glucokinase activator (Ro28-1675) enhanced replication (149). While most glucokinase activator trials have been aimed at human T2D (36), there have been clinical trials in T1D using a hepatoselective activator, TTP399, which resulted in improved glycemic control (150, 151) (ClinicalTrials.gov identifier: NCT03335371). Other glucokinase activators (Table 1) have been shown to enhance human β-cell proliferation in a PKCζ-dependent manner (152), but have yet to be tested in human trials.

Humoral factors in β-cell proliferation

Insulin

While insulin is an endocrine growth factor released from β-cells, it can also signal in an autocrine/paracrine manner. It follows that components of the insulin receptor signaling pathway have been found to be involved in regulation of β-cell mass and expansion. As examples, β-cell mass has been shown to be decreased in a variety of models including β-cell–specific insulin receptor knockout (βIRKO) mice (153, 154), IRS-2 whole-body knockout mice (155), β-cell–specific PDK1 knockout mice (156), and S6K1 knockout mice (157). Alternatively, heterozygous deletion of insulin receptors, IRS-1 and IRS-2, cause increases in β-cell mass (158), while dominant-negative Akt transgenic animals had no β-cell mass changes (159). In the case of insulin receptor, careful investigations in newer β-cell–specific genetic mouse models have found a more nuanced role for β-cell insulin receptor. Skovso et al deleted insulin receptor from β-cells using Ins1-Cre and observed no alterations in β-cell mass, but instead observed increased glucose-stimulated insulin secretion, in support of β-cell insulin resistance causing hyperinsulinemia (160).

Insulin resistance also plays a role in the proliferation of β-cells. This is shown, for example, by studies in mice with liver-specific insulin receptor knockout (161), or antagonist peptides S961 against the insulin receptor (162) and OSI-906 against both insulin receptor and insulin-like growth factor 1 receptor (163, 164). Under these types of conditions, there is evidence that the liver may generate a signal that leads to β-cell expansion (165, 166), but the identities of putative secreted factors are still under investigation. One candidate, Serpin B1, a hepatocyte-secreted protease inhibitor, is a molecule which has been linked to β-cell proliferation in humans, mice, pigs, and zebrafish (167, 168) (Table 1). Additionally, serum Serpin B1 was also found to be decreased in T2D subjects (169), and, under conditions of insulin resistance, a deficiency of Serpin B1 has been shown to be related to reduced β-cell proliferation (167). Because Serpin B1 is a protease inhibitor, the implication is there is a specific protease involved in the inhibition of β-cell proliferation. Indeed, Serpin B1 was found to inhibit the protease elastase to promote human β-cell proliferation (167). Recent studies using of S961 and OSI-906 in β-cell–specific insulin receptor knockout mice have identified the existence of an adipocyte-derived factor which acts through β-cell E2F transcription factor 1 (E2F1) to induce proliferation (170). While the identity of the putative circulating factor is not known, the β-cell proliferative effects required CENP-A and E2F1, potentially upstream of the master regulator transcription factor FoxM1 (171, 172).

Other humoral contributors

Early studies of nerve growth factor suggested that it enhanced β-cell proliferation via promoting production of activin A and betacellulin, an EGF receptor ligand (173). Additionally, EGF-mediated expansion of other cell types, such as pancreatic duct cells (174), has been explored for its transdifferentiation potential. More recently, heparin-binding EGF was shown to induce both EdU incorporation and Ki67⁺ staining of human β-cells, but also caused proliferation of other non-β-cell types (134). Indeed, as noted earlier, EGF receptor can also be transactivated by GLP1R agonists to induce β-cell proliferation (119, 120).

Multiple other circulating factors have been explored for their potential to induce β-cell proliferation/regeneration, including lactogens (175, 176) as well as neurotransmitters like serotonin (177) and γ-amino butyric acid (GABA) (178). However, later testing of many of these factors did not induce proliferation in human β-cells using Ki67⁺ as a readout, while the DYRK1A inhibitor harmine did (128). A recent study using single-cell RNA sequencing of human stem cell–derived β-cells found a subpopulation of proliferative cells which responded to leukemia inhibitor factor (179). This work revealed additional targetable pathways including STAT3 and CEBPD downstream of leukemia inhibitor factor which could be exploited for future β-cell proliferative strategies.

UPR^ER-related factors in β-cell proliferation

Insulin hypersecretion is known to elicit the unfolded protein response of the endoplasmic reticulum (UPR^ER) in β-cells (180), and the UPR^ER is involved in disease progression in T1D β-cells (181). MANF, previously known as ARMET, is a secreted protein and a prosurvival UPR^ER factor in β-cells (182-184). Whole-body deletion or β-cell–specific deletion of MANF in mice leads to reduced β-cell mass and diabetes (184, 185). Stimulation with MANF in vitro stimulated mouse β-cell proliferation (185), and also human β-cell proliferation under conditions of TGFβ inhibition (183). Unmitigated β-cell ER stress will lead to a terminal UPR^ER, involving the transcription factor C/EBP homologous protein (CHOP), encoded by Ddit3/DDIT3 (186). Prolonged CHOP expression will eventually induce apoptosis in β-cells. In mice with β-cell–specific depletion of CHOP, or knockdown of CHOP by GLP1-linked antisense oligonucleotides (GLP1-ASOs), expansion of β-cell mass and improved function is observed, respectively (187, 188). Alternatively, suppressing β-cell activity using inhibitory G_i/o–GPCR signaling, for instance through the α₂-adrenergic receptor, restricts β-cell expansion perinatally and inhibition of this pathway specifically in β-cells can promote their expansion (189).

Other β-cell protectants and proliferation inducers

Inhibitors of the NFκB pathway, specifically osteoprotegerin and denosumab (Table 1), have also effectively induced proliferation of β-cells (190). Similarly, suppression of the cell cycle regulators CDKN2C/p18 and CDKN1A/p21 successfully induced replication of human β-cells (191). Under diabetic conditions, the high demand for secretion of insulin itself may also cause ER stress and this is relieved by the genetic deletion of both isoforms of insulin in mice, thereby promoting the ability of β-cells to proliferate via the Akt-Cyclin D1 axis (192).

Given the separately known β-cell protective effects of both GLP1 (46) and estrogen (193, 194), these 2 molecules have been covalently linked and tested by multiple groups in diabetes models (195-197). The combined results have shown significant body weight reduction in rodents (197) combined with protection of β-cells in streptozotocin-treated mouse models (195) and cytokine-treated human islet β-cells (196). Recent findings also show that suppressing expression of DDIT3, which encodes the terminal UPR^ER transcription factor CHOP, in vivo in β-cells using GLP1-ASOs prevented loss of β-cell function in mouse models of diabetes (188). GLP1-ASOs activate GLP1R and become internalized to deliver the conjugated oligonucleotide (198). In addition, GLP1 has been covalently linked to PPARα/γ dual agonists as a diabetes therapy (199). Each of these efforts support the overall concept that linking GLP1 to protective or proliferation-inducing therapies can be used to more specifically target the β-cell in vivo.

Advances in β-Cell Transdifferentiation and Neogenesis

Stimulating proliferation of endogenous insulin-producing β-cells may be a promising approach to treat diabetes. However, in millions of diabetes cases, most notably T1D, β-cells are nearly completely absent. Without sufficient existing β-cells to proliferate, this approach becomes hampered. Therefore, generation of new β-cells via transdifferentiation or neogenesis may be a viable approach. Sources of cells that have been demonstrated to have β-cell transdifferentiation potential include α-cells, δ-cells, PP cells, and acinar cells; while cells with neogenic potential include pancreatic ductal progenitors and niches of immature β-cell populations (Fig. 1).

Transdifferentiation of β-Cells

The posterior foregut region of the developing embryo is the region which gives rise to the pancreas. The posterior foregut is also the origin of the liver, the posterior part of the stomach, and the proximal region of the gut (200). Considering that these regions are similarly related in their cellular makeup, focusing on cell types from these regions is a rational approach to identifying potential candidates to transdifferentiate into functional β-cells.

α-Cells

A major cell type in β-cell transdifferentiation research is the pancreatic islet α-cell, which can convert to insulin-producing β-cells under certain conditions. Additionally, cells in an intermediate phase of α to β transdifferentiation have been shown to comprise a neogenic niche of immature β-cells at the islet periphery, representing a potentially targetable pool of β-cell progenitors in humans and mice (69-71). To understand mechanisms of α- to β-cell transdifferentiation, studies of transcription factors involved in islet endocrine cell differentiation found opposing roles for Aristaless-related homeobox (Arx) and Pax4 (201). Arx expression positively correlated with α-cell number, while Pax4 repressed α-cell number and increased the numbers of β- and δ-cells. These findings indicate that the relative amount of α- and β-cells is related to the expression of Arx or Pax4, respectively. Indeed, a gain-of-function mutation in mouse Arx resulted in the loss of β- and δ-cells and a gain in the number of α-cells and PP cells (202). Alternatively, in a mouse model of β-cell ablation, lineage-tracing showed many β-cells regenerated from α-cells (90). Deleting Arx specifically in α-cells early in development promoted their transdifferentiation into β-cells (203). In conditional Arx α-cell knockout mice, α-cells readily converted to β-cells even upon Arx deletion in adult animals. Consequently, inhibition of Arx was suggested to be a promising area of future research and in the treatment of diabetes. Further work in adult mice indicated that α-cell–specific deletion of both Arx and DNA methyltransferase 1 (Dnmt1) promoted α- to β-cell conversion (204). α-Cell–specific deletion of Arx alone caused a loss of α-cell identity, while loss of Dnmt1 alone did not impact α-cell identity. Glucose-stimulated insulin release was also observed after this conversion, suggesting the generation of functional β-cells (204). In T1D, some glucagon⁺ cells show a loss of ARX and DNMT1, and produce β-cell products such as insulin. These results suggest that, in humans, ARX and DNMT1 are associated with the maintenance of α-cell identity (204). Notably, another DNA methyltransferase, Dnmt3a, is important for methylating the Arx promoter in rodent β-cells to maintain β-cell identity (205). α-Cells may also be transdifferentiated to β-cells by endogenous factors. The secreted protein IGFBP1 was found in zebrafish to promote α- to β-cell transdifferentiation, and circulating IGFBP1 levels in humans were inversely correlated with T2D risk (206). IGFBP1 was presumed to act through suppressing IGF receptor signaling, as β-cell proliferation effects were also seen with 2 other IGF receptor inhibitors (206).

An in vivo mouse study in 2018 showed adenoviral expression of Pdx1 and MafA delivered to the pancreas via the bile duct caused α to β conversion and reversed diabetes (207). In vitro, overexpression of Pdx1 and MafA in human islets induced insulin expression in α-cells (207). In a separate study, Pdx1 and MafA overexpression in human α-cells resulted in the production and glucose-sensitive secretion of insulin, but the cells maintained α-cell characteristics including expression of Arx (208). Transplantation of these reprogrammed α-cells into streptozotocin-treated mice mitigated diabetes and the grafts continued to express insulin for at least 6 months (208).

Another approach under investigation to regenerate β-cells is via modulating α-cell function. Beneficial effects have been found for injected antiglucagon receptor antibodies in humans with T1D (209) as well as nonhuman primates (210, 211) and rodents (212). Disruption of glucagon receptor activation either genetically or through use of injected antibodies results in α-cell hyperplasia (213, 214), offering a potential source of new β-cells if they can be transdifferentiated. Multiple recent studies in mouse models of T1D and T2D have suggested that a combination of α- to β-cell transdifferentiation and β-cell replication can occur in response to glucagon receptor antagonist antibodies (211, 215-217). These promising findings await further confirmation in humans (218).

Lastly, GABA has been studied for its potential impacts on α to β transdifferentiation. GABA induced β-cell proliferation of human islets transplanted into streptozotocin-treated NOD-SCID mice and in vitro (219). Ben-Othman et al found that long-term GABA administration was also suggested to induce β-cell neogenesis from α-cells both in mice and in transplanted human islets (178). A different study published in the same issue suggested the artemisinin class of antimalaria drugs could promote α to β transdifferentiation through activation of GABA receptor signaling (220). However, follow-up studies indicated that artemisinins promote islet endocrine cell dedifferentiation and not β-cell transdifferentiation (221) and that long-term treatment with GABA or artemisinins had no impact on α to β transdifferentiation in mice (222). Nevertheless, recent clinical trials (ClinicalTrials.gov identifier: NCT03635437 and ClinicalTrialsregister.eu identifier: EudraCT2018-001115-73) found GABA administration to be effective for improving the counterregulatory response to hypoglycemia in humans with T1D (223), but direct effects on α- or β-cells were not studied. Despite disparate findings with regard to transdifferentiation, GABA clearly has a role in islet function (see detailed review (224)).

δ-Cells and PP cells

Besides α-cells, islets contain other endocrine cell types that can transdifferentiate into β-cells, including δ-cells and γ-cells (also known as PP cells).

While α to β transdifferentiation occurs in adult animals after β-cell ablation (90), in juvenile mice β-cell transdifferentiation was found to occur through δ-cells and not α-cells (225). This process involved suppression of amitogenic FoxO1 signaling; pharmacological FoxO1 inhibition (AS1842856 (226)) potentiated the transdifferentiation process. A similar process has been described in zebrafish, where β-cell ablation led to the appearance of somatostatin⁺/insulin⁺ cells which derived from a distinct δ-cell subtype with pro-β identity (227, 228). As these animals recovered normoglycemia, it was suggested that bihormonal cells are functional and important. Single-cell transcriptomic comparison analysis of zebrafish islets to human islet cells suggested this subtype of δ-cell has similarity to human γ-cells (227). Interestingly, lineage-tracing studies of mouse γ-cells showed these cells can also transdifferentiate to different bihormonal cell types including PPY⁺/insulin⁺ cells, and these bihormonal γ-cells were also found in human islets (229).

Acinar and antral stomach cells

Viral gene therapy is an approach that has been successful in mouse models of T1D. In vivo transduction of pancreatic acinar cells with the combination of Ngn3, Pdx1, and MafA led to cellular reprogramming and expression of a β-cell–like phenotype (230). Another source of tissue for β-cell production is the antral stomach. Transgenic overexpression of this same trio of factors, Pdx1, MafA, and Ngn3, in antral stomach tissue resulted in the induction of insulin-positive cells with the ability to rescue mice from streptozotocin-induced diabetes (231). The therapeutic value of this approach was demonstrated by isolation of the transgenic stomach tissue, growth as organoids, and transplantation into diabetic mice prior to induction of Pdx1, MafA, and Ngn3. Mice were rescued from diabetes in those experiments (231).

Differentiation of β-Cells From Progenitors

Cdk5 inhibitors have recently been shown to induce β-cell differentiation from ductal progenitor cells in zebrafish, mice and hiPSCs (232). Additionally, when an adult mouse pancreas is injured, Ngn3⁺ endocrine progenitors have the ability to differentiate into all of the islet cell types through the Notch signaling pathway (233-235). Although these findings arose from studies on rodents, they provide valuable insight that may be translated to human diabetes research. There is an evident role of multipotent pancreatic progenitors during embryonic development. While the existence of a pancreatic islet stem cell that could be targeted in adult humans remains controversial (91), some studies suggest this possibility. A 2020 study used single-cell RNA sequencing to identify a population of mouse islet cells termed protein C receptor-positive (Procr⁺) cells (236). These Procr⁺ cells were suggested to act as progenitors with the ability to be passaged as organoids in vitro and differentiate into each of the islet endocrine cell types. However, other analyses of mouse islet single-cell RNA-sequencing did not detect this Procr⁺ population (237).

The potential of pancreatic ductal cells as a source of β-cell progenitors has been investigated and debated (238-242). Overall, observing substantial β-cell neogenesis from ductal cells seems to require special conditions, such as surgical damage to the pancreas or genetic modifications. Pancreatic ductal cells have the ability to regenerate both endocrine and exocrine tissue in rodents after duct ligation or pancreatectomy (234, 243), or by diphtheria toxin–based ablation of exocrine and ductal tissue (244). Sancho et al found that adult ductal cells can be reprogrammed to multiple islet endocrine cell types (α, β, δ) by deleting Fbw7 in the pancreas (245). They also found that as a ubiquitin ligase component, Fbw7 regulates the stability of Ngn3. Ngn3 is a critical transcription factor required for the development of the pancreatic endocrine compartment (246). Additionally, lineage tracing showed some ductal cells express Ngn3, and over time these Ngn3⁺ ductal cells expressed somatostatin and a subset eventually expressed insulin, suggesting a role in β-cell neogenesis (247). Most of this understanding comes from work in rodents; however, human pancreatic ductal cells can gain β-cell identity when transplanted into mouse models of insulin resistance (248), and ductal cells can express immature β-cell markers in pregnant humans (249). There is now mounting evidence in the field for these ductal progenitors in human pancreas (250), and it may be possible to target such progenitors in diabetic patients to generate new β-cells (Fig. 1). Indeed, pharmacological modulation of MAPK interacting Ser/Thr kinases 2 (MNK2) was discovered to be responsible for β-cell neogenesis from pancreatic ductal cells in vivo (zebrafish) and in vitro (pigs and humans) (251). A role for MNK2 was discovered in this study by deconvoluting the target of the small molecule CID661578. CID661578 turned out to be a MNK2-interacting compound which increased protein synthesis by preventing MNK2 from binding eIF4G of the translation initiation complex. This alteration of MNK2 signaling and enhanced cap-dependent mRNA translation was hypothesized to contribute to the β-cell neogenic effect.

Pluripotent Stem Cell–Derived β-Cells to Replace Endogenous β-Cell Mass

hESCs and hiPSCs are similar in their abilities to differentiate into many different cell types (93). Therefore, these stem cells are excellent candidates for generating insulin producing β-cells. In 1998, Thomson and colleagues derived hESCs, a pluripotent stem cell derived from the inner cell mass of the blastocyst with the ability to give rise to any cell type of the 3 human germ layers (252). Thomson used donated human embryos, which became available to him from unused fertility treatments (252). Despite ethical implications, multiple developments have been made using hESCs as precursors in the formation of insulin producing β-cells (253). Alternatively, hiPSCs are reprogrammed from somatic dermal fibroblasts or peripheral blood mononuclear cells (254). In 2006-2007, the Yamanaka laboratory pioneered a somatic cell reprogramming process by expressing 4 specific genes in mouse and human fibroblasts (255, 256). The overexpression of these specific genes (OCT4, SOX2, KLF4, and CMYC) was necessary and sufficient to induce pluripotency (256). These studies were a milestone in stem cell research, as it enabled the creation of pluripotent stem cells without the use of human embryos. Like hESCs, hiPSCs have the ability to differentiate into multiple cell types. hESCs and hiPSCs can be coaxed through a complex multistage process to result in insulin-producing β-like cells. Typically stem cells must be subjected to a weeks-long multi-step process first requiring generation of definitive endoderm (257), followed by primitive gut tube cells and pancreatic progenitors (258), and finally immature pancreatic endocrine cells before differentiation of β-like cells (259-261). There have been limitations with stem cell–derived β-like cells, notably reduced insulin production and secretion as compared to primary β-cells. However, the process has been further refined to permit more steps to occur in monolayer culture (262, 263), and recent work has shown stem cell–derived islet function quite close to that of primary human islets (264). Additional work has demonstrated generation of enhanced β-cell clusters (eBCs) by reaggregating FACS-enriched hESC-derived β-like cells (265). These eBCs had improved mitochondrial respiration and increased similarity to endogenous β-cells in terms of function, such as dynamic insulin secretion. Further work with eBCs determined that self-aggregation yielded development of more mature self-enriched β-cell clusters (seBCs) (266). It was discovered that the most mature of these β-like cells are marked by a cell surface protein, ENTPD3. This study, and others like it (267), which provide large transcriptomic datasets will aid in discovering relevant pathways and markers to improve mature β-cell differentiation in vitro.

The multistep differentiation protocols which coax hESCs and hiPSCs into β-like cells require the modulation of many signaling pathways at each step. Furthering understanding of the signaling pathways governing β-cell differentiation and maturation provides the necessary data to improve stem cell–derived β-cell generation. Recent investigations of TGFβ, Wnt, and the F-actin cytoskeleton are prime examples. TGFβ signaling has been shown to have an effect on developing and adult β-cells, as removing the actions of TGFβ inhibitors can promote the β-cell maturation process (268). Indeed, the TGFB2 gene was found to be upregulated in ENTPD3⁺ eBCs (266). Wnt pathway signaling is also involved in pancreatic differentiation and β-cell maturation (79, 269, 270). The Wnt family member Wnt4 is highly expressed in human islets and inclusion of Wnt4 in the terminal stages of hiPSC-derived β-cell maturation improved production of insulin and urocortin-3 (269). Wnt4 was also recently shown to be involved in β-cell heterogeneity, with Wnt4⁺ cells being more mature and Wnt4^– cells being more proliferative (271). Finally, a very useful discovery was made by the Millman laboratory, which found that adding the F-actin depolymerizing agent latrunculin A at a precise stage of the differentiation protocol improved expression of NEUROG3 and subsequent efficiency of stem cell–derived β-cells from both hESCs and hiPSCs (262, 263). Identifying new markers of β-cell maturation and innovative technical advancements like above will advance the creation of replacement cell therapies that are increasingly similar to endogenous human islets.

Completed and ongoing clinical trials are testing the potential to use stem cell–derived pancreatic endoderm and β-like cells by transplanting either encapsulated or naked tissues. For example, ViaCyte has developed pancreatic endoderm cells (PEC-01) from an hESC line CyT49 (272); the PEC-01 cells mature to β-cells after implantation. Initial tested devices for implanting PEC-01 cells (VC-01) failed due to insufficient function after implantation due to hypoxia and necrosis (ClinicalTrials.gov identifier: NCT02239354). Therefore, a new PEC-Encap (VC01-103) semipermeable macroencapsulation device was designed to prevent access to the immune system, but retain improved permeability to prevent hypoxia within the implant. The improved PEC-Encap device is in phase 2 clinical trials with results anticipated in 2023 (ClinicalTrials.gov identifier: NCT04678557). In addition, ViaCyte is testing PEC-01 cells contained an alternate device with pores for capillary access and improved β-cell maturation (VC-02) (273, 274), which is also in phase 2 trials (ClinicalTrials.gov identifier: NCT03163511). VC-02 requires immunosuppression because of the circulatory access of the device. Finally, a new experimental treatment, VCTX210, developed in a collaboration between ViaCyte and CRISPR Therapeutics, uses the CyT49 hESCs (272) engineered with CRISPR gene editing to be immune evasive and differentiated to pancreatic endoderm (PEC-QT), which are contained within a macroencapsulation device for implantation. VCTX201 is currently being tested in an ongoing clinical trial (ClinicalTrials.gov identifier: NCT05210530).

An alternative strategy from Vertex Pharmaceuticals, VX-880, uses naked stem cell–derived β-like cells transplanted into the liver via the portal vein, similar to how primary human islet transplants are performed (ClinicalTrials.gov identifier: NCT04786262). Therefore, this approach still requires immunosuppressive drugs to prevent tissue destruction. Current results are limited to only a few patients, and while the study has not been peer-reviewed, the grafts are reported to be tolerated and yield reduced reliance on exogenous insulin. Interestingly, Vertex recently acquired ViaCyte (275), forecasting multiple potential combination approaches between macroencapsulation technology and CRISPR-edited stem cell–derived functional β-like cells.

Ultimately, the final products of any cell replacement approach must be adapted to mass production for clinical usage, protected from the immune system either with physical barriers or immunosuppression and also ensured not to spread to the recipient any potential teratogenic or tumorigenic cells (93, 276). In addition to the kinases and signaling pathways discussed targeting β-cell proliferation or transdifferentiation, there have been advancements in targeting immune response pathways in the β-cell to protect them from immune destruction. For example, the Janus kinase family member tyrosine kinase 2 (TYK2), which mediates signaling downstream of interferon receptors, has been associated with multiple autoimmune diseases, including T1D (277-279). Knockout or pharmacological inhibition of TYK2 protects β-cells from interferon-α–mediated effects (280, 281). Other strategies to protect β-cells include genetic modification of stem cells prior to differentiation to β-like cells. Deletion of multiple human leukocyte antigen (HLA) genes, importantly while retaining one (HLA-A2), reduced immune rejection of transplanted hiPSC-derived islet cells (282).

Additional consideration should be given to islet innovation and vascularization. Endogenous islets pancreatic islets are highly innervated, influencing islet development and function (283). Islets are also vascularized continuously with the pancreatic exocrine tissue where islet-resident pericytes help regulate capillary constriction (284, 285). Recapitulating this endogenous vascularization and innervation represents a potential hurdle for transplanted tissue aimed at replacing endogenous islet function. These are challenges that must be overcome in order for such methods to be effective clinically. Advancements in producing large quantities of functional β-cells that can safely avoid destruction after engraftment are imperative for future successful treatments.

Conclusion

The potential for regenerative medicine to treat diabetes has progressed significantly in the last 40 years. Endogenous β-cell regeneration can occur via numerous methods; however, a majority of these methods have succeeded in animals but have not yet led to accessible human therapies (85). A major issue in humans is the immune attack of transplanted β-cells or the autoimmune destruction of β-cells derived from neogenesis in T1D. Strategies are under continued development to defend β-cells against these insults, including encapsulation technologies (286) and immunotherapies replenishing anti-inflammatory T regulatory cells to confer tolerance to new or transplanted β-cells (287). Despite the challenges, the future is bright in the field of regenerative medicine to treat diabetes.

Abbreviations

CDK

cyclin-dependent kinase

CHOP

C/EBP homologous protein

eBC

enhanced β-cell cluster

EGF

epidermal growth factor

GABA

γ-amino butyric acid

GLP1R

glucagon-like peptide 1 receptor

GPCR

G protein–coupled receptor

HB-EGF

heparin-binding epidermal growth factor-like growth factor

hESC

human embryonic stem cells

hiPSC

human induced pluripotent stem cell

IL

interleukin

NECA

5′-N-ethylcarboxamide adenosine

NFAT

nuclear factor of activated T cells

PP

pancreatic polypeptide

SIK

salt-inducible kinase

T1D

type 1 diabetes mellitus

T2D

type 2 diabetes mellitus

TGF

transforming growth factor

TNF

tumor necrosis factor

UPR^ER

unfolded protein response of the endoplasmic reticulum

Disclosures

The authors have nothing to disclose.

Data Availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.