Reprogramming of Different Cell Lineages into Functional β-Cell Substitutes

Abstract

Since its first use in 1922, insulin therapy has transformed diabetes from a fatal disease to a manageable condition. However, long-term insulin injections lead to significant complications. β-cell replacement, derived from either a limited number of deceased donors or embryonic stem cells, offers an encouraging alternative. While these procedures allow patients to be insulin-independent, they still require systemic immunosuppressants to prevent graft rejection, which poses immunological challenges. Direct reprogramming holds considerable promise as a method for generating β-cells from various sources, enabling autologous therapies that mitigate the risk of immune rejection and eliminate the need to harvest cells from embryos. This review provides an overview of the latest advances in direct reprogramming strategies, with a focus on key transcriptional regulators that drive phenotypic conversion and maintenance of various cell types into β-like cells.

Introduction

According to the International Diabetes Federation, diabetes currently affects more than 537 million people globally. This number is projected to exceed 783 million by 2045 (diabetesatlas.org). Furthermore, the Centers for Disease Control and Prevention identified diabetes as the eighth leading cause of death in the United States as of 2021 (cdc.gov). Diabetes is also recognized as one of the most critical comorbidities in patients infected with the SARS-CoV-2 virus (American Diabetes Association, 2013).

Diabetes comprises a heterogeneous group of metabolic disorders characterized by hyperglycemia (American Diabetes Association, 2013). The most common classifications of diabetes include Type 1 and Type 2. Type 1 diabetes (T1D) is caused by autoimmune destruction of pancreatic β-cells, rendering patients incapable of producing sufficient insulin. In contrast, Type 2 diabetes (T2D) results from the resistance of the body to insulin. This leads to long-term exposure to hyperglycemia, which promotes β-cell dysfunction due to metabolic stress (Cerf, 2013). Consequently, both T1D and advanced T2D patients rely on exogenous administration of insulin.

Since its first injection in January 1922, insulin therapy has transformed diabetes from a terminal illness into a manageable condition (Vecchio et al., 2018). However, insulin injections remain an imperfect solution, as they cannot precisely mimic endogenous β-cell function. In the long run, this leads to complications, including hypoglycemia, weight gain, dizziness, heart disease, and kidney disease, due to glucose levels in the blood that are still not adequately controlled (Efrat, 2002). Therefore, therapies that aim at a constant and perpetual reconstitution of physiological glycemia are extremely desired.

In this context, β-cell replacement offers an encouraging approach. Lantidra, a cell therapy recently approved by the Food and Drug Administration (FDA) (FDA.gov), utilizes pancreatic islet transplantation from deceased donors to provide an endogenous source of insulin. While this procedure allows patients to be insulin-independent (Bellin et al., 2012), it has important limitations that restrict its widespread application to cure diabetes. Besides the risks associated with invasive surgery, patients need to take immunosuppressors to reduce graft rejection. In the long run, this can lead to immunological challenges, including toxic side effects (Ruiz and Kirk, 2015). Importantly, the most critical limiting factor in the application of this therapy is the scarcity of donors (Rickels and Paul Robertson, 2019). Therefore, alternative approaches to human cadaveric islets are urgently sought.

In this regard, the use of human pluripotent stem cells (hPSCs), including induced pluripotent stem cells (iPSCs) and embryonic stem cells (ESCs), as well as reprogramming of different cell lineages to adopt a β-cell phenotype, promises to provide a potentially endless source of patient-specific functional β-cells to be transplanted into patients with diabetes (Shi et al., 2017; Colarusso and Zhou, 2022; Pagliuca et al., 2014). Specifically, ESC-derived islets have already been employed in an FDA trial and patients injected with a single dose of these cells demonstrated glucose-responsive insulin production and insulin independence for up to 1 year (https://news.vrtx.com/news-releases/news-release-details/vertex-announces-positive-results-ongoing-phase-12-study-vx-880). However, this new cell therapy still requires lifelong systemic immunosuppression. In addition, a clinical trial using chemically reprogrammed iPSC from adipose-derived mesenchymal stromal cells to treat T1D is already ongoing in China (Wang et al., 2024).

Direct reprogramming, also known as transdifferentiation, holds significant promise as a method for generating β-cells from different sources, particularly because it could be customized for individual patients in autologous therapies that will mitigate the risk of immune rejection (Fig. 1). Importantly, this approach offers the potential for personalized treatment options, enhancing therapeutic outcomes. The initial breakthrough in this field demonstrated that the exogenous expression of three transcription factors (TFs)—NGN3, PDX1, and MAFA—was sufficient to drive the conversion of mouse pancreatic acinar cells into insulin-producing cells (Zhou et al., 2008). With this review, we aim to provide a comprehensive overview of the most advanced direct reprogramming strategies for β-cell generation, highlighting the key transcriptional regulators and cell sources that are driving progress in this emerging area of regenerative medicine for diabetes treatment.

FIG. 1.

Reprogramming strategies for the cure of diabetes. Reprogramming is a process that involves a change in cell identity that does not include a pluripotent intermediate state. Current reprogramming strategies for the cure of diabetes utilize the exogenous expression of three important pancreatic reprogramming transcription factors by viral delivery: NGN3, PDX1, and MAFA, which turn different types of cells into β-like cells.

Pancreatic Cell Sources

Acinar cell reprogramming

Acinar cells are specialized exocrine cells in the pancreas that produce and secrete digestive enzymes into the duodenum. While serving distinct functions in the pancreas, acinar cells and β-cells both arise from pancreatic progenitor cells of the foregut endoderm (Mehta et al., 2022). Despite PDX1 being a critical transcription factor in early pancreas development, cells from the exocrine lineage downregulate PDX1 later in development. Instead, PTF1A expression is upregulated and is key for acinar cell identity (Mehta et al., 2022). As mentioned above, the major breakthrough for transdifferentiation was discovering pancreatic acinar cells of the mice could be reprogrammed into insulin-producing cells with three TFs NGN3, PDX1, and MAFA (collectively named NPM factors) (Li et al., 2014a). Mice were injected with adenovirus encoding those three transcription factors directly into the pancreas, and insulin production was verified with immunohistochemistry. In another study, the pancreatic acinar cells have shown plasticity and could be reprogrammed into α, β, and δ cells through modulating the expression of the key transcription factors mentioned prior (Zhou et al., 2008). For instance, NGN3 alone induces the reprogramming of acinar cells to δ-cells, while NGN3 and MAFA together can reprogram acinar cells to α-cells (Zhou et al., 2008). Importantly, the coexpression of NPM factors altogether drives the reprogramming of acinar cells to β-cells (Zhou et al., 2008). This finding was supported in another study using a transgenic mouse model that inducibly expresses the three transcription factors with doxycycline (Clayton et al., 2016). The key finding was that too high levels of NPM factors led to pancreatic inflammation, driving acinar cells to a duct-cell fate rather than a β-cell fate. This issue can be mitigated by attenuating inflammation, either through reduction in the expression of the three transcription factors or by depleting macrophages. More recently, Dahiya et al. described a novel method for the reprogramming of acinar cells to what they termed acinar-derived insulin producing cells (ADIP) (Dahiya et al., 2024). Their process involves the pharmacological inhibition of focal adhesion kinase. The study demonstrated through lineage tracing that ADIP cells migrate into islets and begin expressing insulin, effectively integrating into the islet microenvironment. Furthermore, in a streptozotocin-induced diabetic mouse model, PF-562271 treatment led to improved glucose regulation, highlighting the therapeutic potential of this strategy. However, the treatment did not fully reverse diabetes, indicating that while ADIP cells contribute to glucose homeostasis, additional interventions may be necessary to achieve complete β-cell functionality (Dahiya et al., 2024). These studies in acinar cell reprogramming highlight the plasticity of acinar cells and demonstrate that they are a potential source for cell therapies. However, the challenges are also apparent, such as the ability to directly specify β-cell reprogramming and the issue of maturation to ensure they can reverse diabetes. Future studies should explore more efficient reprogramming of acinar cells to mature β-cells to make acinar cells a viable option for reprogramming as a diabetes treatment.

Transdifferentiating α- to β-cells

α- and β-cells are two highly specialized cell types found in the islets of Langerhans, which derive from the endocrine tissue of the pancreas. α-cells recently became an attractive source for β-cell replacement. In fact, they share the same endocrine gene signature (Herrera, 2000) and post-differentiation share a few TFs with β-cells, including MAFB, PAX6, and NKX2-2 (Saleh et al., 2021). Furthermore, α- and β-cells are located in the same anatomical structure, receiving the same blood supply (Herrera, 2000), as well as sympathetic innervation (Salvioli et al., 2002). Moreover, they share the glucose sensor SLC2A2 and a similar machinery to secrete hormones (Quesada et al., 2008; Zhang et al., 2013). Thus, these similarities might facilitate cell reprogramming. Furthermore, it has long been known that there is high interconversion plasticity between the two cell lineages. For instance, mice lacking PDX1 show a higher number of α-cells in the islets, accompanied by changes in the location of this cell population within the Langerhans islets (Ahlgren et al., 1998). Importantly, the Herrera group showed that massive ablation of β-cells using diphtheria toxin in adult mice leads to spontaneous regeneration of the β-cell population by the natural reprogramming of α-cells (Thorel et al., 2010).

Following these results, much research has gone into improving α-to-β cell conversion for therapeutic purposes. It has been shown in mice that the in vivo exogenous overexpression of Pax4 either into endocrine progenitors or adult α-cells causes the conversion of this lineage into β-cells, expanding the size of this population in islets. Specifically, Ngn3 reactivation was required for the Pax4-mediated α-to-β cell reprogramming (Collombat et al., 2009). Other factors that are crucial for transdifferentiation include HNF4A (Sangan et al., 2015). Repression of Arx and Dnmt1 either genetically or drug induced in the mouse pancreas leads to conversion of α- to β-cells (Chakravarthy et al., 2017). Some studies showed that Arx repression can also be obtained by long-term GABA exposure in wild-type mice, which in turn expands β-cell mass due to α- to β-cell transdifferentiation (Ben-Othman et al., 2017). However, these findings were not reproducible using different strains of mice, putting into question the universal applicability of GABA treatment in diverse human patients (Ackermann et al., 2018). More recently, it was shown in adult mice that the ectopic expression of Pdx1 and Mafa via adeno-associated viral delivery into the pancreatic duct is able to push α-to-β-cell conversion together with normalization of glucose homeostasis in diabetic mice. In this context, it was specifically observed that the α-to-β cell transdifferentiation mediated by Pdx1-Mafa overexpression under the glucagon promoter in adult mice is able to reverse β-cell dysfunction in several autoimmune T1D-induced mice. Furthermore, the newly converted β-like cells were found resistant to immune attack (Guo et al., 2023), suggesting they are not immediately recognized by the immune system. Therefore, this last approach could be improved by combining with immunosuppression drugs to increase the lifespan of the newly generated β-like cells. More studies in this direction showed that cadaveric human non-β-cells (i.e., α and pancreatic polypeptide cells) could also be converted into a β-like phenotype by overexpression of Pdx1 and Mafa with the resulting pseudoislets able to produce insulin and rescue diabetes in mice (Furuyama et al., 2019).

Duct epithelial cell reprogramming

Duct cells of the pancreas are exocrine cells that transport digestive enzymes secreted from acinar cells to the duodenum. Duct cells, such as acinar cells, are part of the exocrine pancreas and arise from the same progenitors (Grapin-Botton, 2005). Interestingly, a population of duct cells that retains NGN3 expression was identified (Gribben et al., 2021). Lineage tracing revealed that they have the capacity to be reprogrammed into β-cells and migrate out of the ductal epithelium (Gribben et al., 2021). This discovery supports that duct cells inherently have plasticity, which could be leveraged for reprogramming. There were successful efforts in reprogramming duct cells into endocrine cells through adenoviral transduction of Ngn3, Mafa, Pdx1, and Pax6, which yielded β-like cells (Lee et al., 2013). The authors reported that these reprogrammed duct cells are as efficient in insulin secretion as β-cells derived from ESCs, which is about 49%–77% of a control β-cell. Another group investigated epigenetic modification of ductal progenitor cells using EZH2 methyltransferase inhibition to drive duct cells from T1D and nondiabetic donors toward a β-cell fate (Al-Hasani et al., 2024). From the three donors, two of which being T1D donors, treatment with either GSK126 or Tazemetostat resulted in an increase in insulin messenger RNA expression and activation of endocrine markers to varying degrees. Additionally, the study reports only a 20% increase in insulin secretion under high glucose in glucose-stimulated insulin secretion, which they attribute to low cell number due to low starting material from the donors. To make duct cells a viable candidate for β-cell regeneration, more work needs to be done in studying the maturation of reprogrammed duct cells so that they can produce enough insulin to reverse diabetes.

Conversion of Other Foregut-Derived Cell Lineages

Hepatocyte reprogramming

Liver appears to be an ideal candidate for developing reprogramming strategies aimed at converting tissues into β-like cells. Indeed, hepatocytes were among the first cells to be converted into a β-like phenotype (Kalo et al., 2022). This is due to three main reasons: the regenerative capability, the accessibility of the adult liver (as opposed to the pancreas), and the location of these two organs on the developing gut tube (Colarusso and Zhou, 2022). In fact, both liver and pancreas derive from the endoderm of the posterior region of the foregut tube. These domains lie in proximity to each other, so they originate from the same precursors (Deutsch et al., 2001; Angelo et al., 2012). Furthermore, in the developed organs, hepatocytes and β-cells employ similar mechanisms to detect glucose, involving the glucose transporter 2 (GLUT2) (Bae et al., 2010). In addition, they share many TFs, including the hematopoietically expressed homeobox (HHEX) (Watt et al., 2007) and the GATA binding protein 4 (GATA4) (Bideyan et al., 2022). Therefore, these biological similarities determine a great intercellular plasticity between the two organs which is persistent throughout adulthood. Indeed, the presence of hepatic foci has been detected in mice pancreas in response to specific conditions (injury, special dietary regimen, carcinogens) (Shen et al., 2003). Most interestingly, it has been recently found that the loss of a liver TF, TBX3, improves pancreatic progenitor differentiation from human pluripotent stem cells, by increasing the expression levels of pancreatic factors such as PDX1, ISL1, and MNX1 (Mukherjee et al., 2021).

It has long been known that the reprogramming of hepatocytes into insulin-producing cells can occur via overexpression of pancreatic TFs (Akinci et al., 2013). As a first attempt, Ferber and colleagues used adenoviral vectors to deliver Pdx1 (Ferber et al., 2000). Based on this work, several other labs tried to transduce Pdx1 in combination with other pancreatic TFs, finding that the overexpression of NPM factors guarantees the highest conversion efficiency (Meivar-Levy and Ferber, 2019). However, NPM overexpression was shown to only partially reproduce β-cell phenotype in vitro and in vivo. For instance, when NPM overexpression was tested in diabetic NOD-SCID mice, the resulting β-like cells were able to secrete insulin but showed mixed properties between liver duct cells and β-cells. These reprogrammed cells originated from SOX9⁺ bile duct cells, which are generally resistant to reprogramming. As a result, the strategy only partially alleviated hyperglycemia (Banga et al., 2012).

A better reprogramming efficiency, although still partial, was obtained via adenoviral delivery of NPM in cultured hepatoblasts. Other experiments involved a constitutionally active form of PDX1, with consequent morphological aberration of the tissue (Imai et al., 2005). Based on these results, new studies are looking in two directions: (i) to determine which specific liver subtype is more prone to reprogramming and (ii) the search of new factors to overexpress in liver cells. In this context, TGIF2 seems to be a good candidate as it is involved in the cell fate decision between liver and pancreas, promoting the expression of pancreatic program, as well as repressing hepatic phenotype (Cerdá-Esteban et al., 2017). Perhaps, the combination of TGIF2 overexpression and repression of TBX3 obtained through protocol modification of culture conditions might provide a starting point for a successful conversion of liver cells into pancreatic β-cells.

Gut epithelial cells reprogramming

The gastrointestinal (GI) tissue represents another great source of cells to reprogram into β-like cells. Indeed, stomach and pancreas share the same endodermal origin from the foregut. Furthermore, native antral endocrine cells share a surprising degree of transcriptional similarity with pancreatic β-cells (May and Kaestner, 2010). In addition, GI epithelium contains endocrine cells that secrete various hormones similar to pancreatic endocrine cells (May and Kaestner, 2010). Most importantly, gut-derived hormones, generally known as incretins, that is, gastric inhibitory peptide (GIP) (Yamane et al., 2016) and glucagon-like peptide 1 (GLP-1) (Wei et al., 2023), can promote β-cell growth and the release of pancreatic hormones in response to glucose. Based on the entero-insular axis concept, previous work showed that intraperitoneal injections of intestinal epithelial cells together with GLP-1 in diabetic rats stimulate them to secrete insulin sufficiently to rescue diabetes in these animals (Suzuki et al., 2003). Furthermore, overexpression of Pdx1 alone or in combination with GLP-1 in intestinal epithelial cells was found to also promote insulin secretion (Koizumi et al., 2006). Finally, it has been shown that intestinal K-cells from mice can be induced to secrete insulin by overexpressing the insulin gene under the 5ʹ regulatory region of the glucose-dependent insulinotropic polypeptide (GIP). K-cells secreting insulin protected mice from developing diabetes once native β-cells were ablated (Cheung et al., 2000).

Conversion of GI cells into β-cells can also be achieved pharmacologically. Previous studies on murine gut enterochromaffin cells and human iPSC-derived gut organoids showed that FOXO1 knockdown, either obtained by dominant negative mutant or lentiviral shRNA, promotes the generation of insulin-secreting cells (Talchai et al., 2012). In fact, FOXO1 is a key TF involved in pancreas versus gut fate decision. It inhibits the Notch pathway, which drives Ngn3 expression. Based on this work, the Accili group developed a fine-tuned combination treatment to increase the efficiency of β-cell conversion in both mice and human adult intestinal organoids, with subsequent conversion of intestinal crypts into β-like cells. Importantly, in vivo, the administration of a triple blockade of FOXO1, Notch, and TGF-β in either insulin-deficient or NOD mice was sufficient to rescue glucose levels (Talchai et al., 2012).

About a decade ago, the Stanger Lab screened for the best tissue that undergoes conversion upon the NPM cocktail misexpression induced by Dox treatment in vivo. With this strategy, they found that intestinal crypts converted into “neo-islets,” showing ultrastructural and functional properties of β-cells (Chen et al., 2014). Importantly, these neo-islets were shown to be glucose-responsive and able to mitigate hyperglycemia in a murine diabetic model. More studies in this direction involve the overexpression of other important TFs, including PTF1A, which was shown to convert stomach and intestinal tissue into both pancreatic endocrine and acinar cells (Kawaguchi et al., 2002).

Interestingly, the Zhou Lab showed that the NPM cocktail could efficiently convert in vivo antral cells into insulin-secreting cells with a molecular landscape and functionality similar to pancreatic β-cells. Importantly, these GI insulin-producing cells could rescue hyperglycemia in diabetic mice for at least 6 months and show a great regeneration potential upon ablation (Ariyachet et al., 2016).

Gastric stem cell reprogramming

One of the most exciting news on β-cell conversion comes from gastric stem cells. A recent report by Huang et al. showed that these cells are particularly prone to becoming insulin-secreting cells upon sequential activation of NPM factors (Huang et al., 2023). SOX9⁺ gastric stem cells are easily accessible via biopsy of the gastric mucosa. These cells can be isolated, cultured, and expanded on feeder layers to produce billions of cells within ∼60 days. Functional gastric insulin-secreting cells (GINS) can be obtained within 10 days with a straightforward three-step protocol (Fig. 2). Importantly, these cells exhibit key functional features of pancreatic β-cells— they are glucose-responsive, adequately respond to GLP-1, and can reverse hyperglycemia in immunodeficient diabetic mice upon transplantation for over 3 months.

FIG. 2.

Differentiation of gastric insulin-secreting organoids. Gastric stem cells can easily be obtained by stomach biopsy and expanded into billions of cells. The sequential activation of Ngn3 and Pdx1-Mafa can turn them into insulin-secreting organoids within 10 days.

Specifically, two major aspects make these cells particularly attractive for therapeutic purposes. First is the speed and robustness of the differentiation process. In contrast, differentiation of pluripotent stem cells into β-like cells typically takes between 40 and 60 days (Pagliuca et al., 2014), with efficiency greatly varying across different cell lines. The process is even more prolonged in autologous settings, where reprogramming somatic cells into iPSCs can take up to 2 months (Choi et al., 2023). Second—and most critical for β-cell replacement therapy—is the absence of proliferation following transplantation. Unlike pluripotent stem cell-derived islets, which carry a risk of teratoma formation due to residual undifferentiated cells, undifferentiated gastric stem cells undergo apoptosis post-transplantation (Lezmi and Benvenisty, 2022). This feature places GINS at the forefront in terms of safety.

However, single-cell RNA-seq comparison with the available human islet dataset revealed that GINS still show some level of immaturity. For instance, this results in limited insulin secretion capabilities due to low expression levels of SCL30A8 (a gene encoding a zinc transporter responsible for insulin secretion and storage). Furthermore, differently from human islets, gastric-derived islets show a lower amount of α- and δ-cells, which may impair the fine regulation of insulin release and increase the risk of hypoglycemia. The developmental path of GINS during differentiation is unique, involving SOX4⁺ and GALANIN⁺ progenitors. Interestingly, these two factors were never found to be crucial for pancreas differentiation. Finally, other concerns are relative to the expression levels of the transduced factors. Indeed, PDX1^low and MAFA^low β-cells contribute to β-cell homeostasis and synchronization with neighboring β-cells. Continued expression of these two factors might disrupt the functional heterogeneity among β-cells (de Koning and Carlotti, 2023).

In conclusion, this study presents a pioneering strategy for generating functional insulin-producing cells from human stomach tissue. GINS emerged as a compelling alternative to pluripotent stem cell-derived islets in regenerative therapeutics for patients with insulin-dependent diabetes. Future direction should aim at characterizing the immunogenicity profile of GINS, as β-cell loss is caused by autoimmune attack in T1D.

Conversion of Other Cell Lineages

Fibroblasts reprogramming

One of the most critical aspects of any reprogramming strategy is the accessibility, availability, and ease of handling of the cell source. For this reason, fibroblasts have been particularly instrumental for reprogramming strategies, with successful conversion into many cell types (Shahbazi et al., 2016; Vierbuchen et al., 2010; Ieda et al., 2010; Huang et al., 2014). It is well known that fibroblasts can be reprogrammed first to pluripotency to create iPSCs, which can then be differentiated to create islets (Pagliuca et al., 2014). However, direct conversion of fibroblasts into β-like cells has proven more challenging (Zhou et al., 2008; Lee et al., 2013). In 2014 and 2016, the Ding group demonstrated the conversion of fibroblasts into insulin-secreting cells through a small molecule protocol either from mouse or human. Specifically, the β-like cells when transplanted into a surrogate were able to function and rescue diabetes in mice (Li et al., 2014b; Zhu et al., 2016). Recently, Fontcuberta-PiSunyer et al. showed that a cocktail including five endocrine TFs, NGN3, PDX1, MAFA, PAX4, and NKX2-2, can directly reprogram human foreskin fibroblasts into β-like cells while repressing the fibroblast transcriptional network (Fontcuberta-PiSunyer et al., 2023). The converted β-like cells were able to secrete insulin both in vitro and in vivo. Indeed, the sole expression of NGN3, PDX1, and MAFA generated polyhormonal GCG⁺ and SST⁺ cells. However, the sequential addition of PAX4 and NKX2-2—at 3 and 6 days post-NPM transduction, respectively—successfully promoted a β-cell-like identity. These cells are functional and release a similar amount of insulin compared to human islets. While they maintain the capability to secrete insulin even after transplantation in NSG mice, whether these cells could rescue diabetes in a mouse model remains unclear. Interestingly, the authors utilized the anterior chamber of the eye as the transplantation site, which allows for fast engraftment and in vivo imaging (Adeghate and Donath, 1990; Leibiger and Berggren, 2017). In summary, this is the first approach in which fibroblasts were successfully differentiated into β-cells by defined factors. This strategy benefits from a readily available cell source, such as skin fibroblasts, which enhances the translational potential for diabetes treatment.

Pluripotent Stem Cell Reprogramming

The most revolutionary discovery in regenerative medicine comes from the Yamanaka factors, a cocktail of four TFs: OCT4, SOX2, KLF4, and MYC (commonly known as OSKM), which allow terminally differentiated cells to be reprogrammed into iPSC. While this strategy opened the doors to the possibility of autologous treatment for many pathological conditions, including cancer and diabetes, the process itself can be quite inefficient. This is due to multiple reasons, including tumor suppressors and various microRNAs present in any cell. Cells that do not undergo dedifferentiation can either die or become senescent (Chondronasiou et al., 2022). These “senescent cells” cannot always be distinguished from PSC and prolonged senescence can result in tumor development, chronic inflammation, and stem cell exhaustion (Huang et al., 2022). Furthermore, iPSCs maintain epigenomic memory of the original terminally differentiated cells, and this can represent a barrier for differentiation and or to maintain stemness over time. Finally, while pluripotent stem cells can undergo differentiation through the addition of small molecules in the culture (avoiding viral delivery) (Pagliuca et al., 2014), they lack an efficient system for the purification of differentiated β-cells (Pellegrini et al., 2022). Therefore, the heterogeneous population of pancreatic endocrine cells combined with the length of the differentiation poses a limitation for the broad application of this approach.

Recently, infection of these cells with a lentivirus carrying the NPM cocktail was attempted in human iPSCs (Jeyagaran et al., 2024). Upon infection and subsequent selection, cells were treated with doxycycline at the stage of pancreatic progenitors, and cells were shown to be glucose-responsive in a monolayer culture. However, these cells are dependent on doxycycline, as endogenous insulin production was not observed even after 10 days of NPM overexpression. Nevertheless, the enhancement of the forward programming of iPSCs into mature functional β-cells could represent the solution to the problem of scaling and allow the commercialization of insulin-secreting cells for transplantation into patients suffering from diabetes.

Future Directions

The reprogramming strategy shows great potential to provide a cure for diabetes (Table 1). In this context, the generation of GINS from gastric stem cells might take the lead among the systems presented in this review, since they demonstrated a great efficiency of conversion, feasibility for large-scale differentiation thanks to the simple culture conditions, therapeutic benefits in mice, as well as sustainability of the new differentiation state in vivo. Indeed, GINS transplantation in mice did not result in teratoma formation up to 6 months upon transplantation. In fact, consistent with previous literature on the reliance of Wnt signaling to survive, GSCs die upon transplantation, as also demonstrated by Huang et al. in their experiments. This shows little risk of uncontrolled proliferation that can lead to tumor formation in the body. Importantly, reprogramming technologies, such as GINS, bypass harvesting cells from embryos.

Table 1.

Summary of Reprogramming Strategies

Cell type	Factors	Resulting direct reprogramming	Reported efficiency	Citation
Acinar	Ngn3, Pdx1, Mafa	Beta	47 ± 8%	Li et al., 2014a
	Ngn3	Delta	40 ± 3%
	Ngn3, Mafa	Alpha	9 ± 5%
Alpha	Pax4	Beta	35%	Collombat et al., 2009
	HNF4A		Not reported	Sangan et al., 2015
	Repression of Arx and Dmnt1		50%–80%	Chakravarthy et al., 2017
	Pdx1, Mafa		65% recovery of β-cell mass	Guo et al., 2023
Duct	Ngn3, Pdx1, Mafa, Pax6	Beta	7%–11%	Lee et al., 2013
Duct	Repression of EZH2 methyltransferase	Beta	Not reported	Al-Hasani et al., 2024
Hepatocyte	Ngn3, Pdx1, Mafa	Beta	12%	Akinci et al., 2013
Hepatocyte	PDX1	Beta	0.1%–1%	Ferber et al., 2000
Gut epithelial	Repression of FOXO1	Beta	4.5%–10.5%	Talchai et al., 2012
Intestinal crypt	Ngn3, Pdx1, Mafa	Beta	31.6%	Chen et al., 2014
Gastric stem cell	Ngn3 (2 days), Pdx1, Mafa	Beta	65.4 ± 5.2%	Huang et al., 2023
Fibroblasts	Ngn3, Pdx1, Mafa, Pax4, Nkx2-2	Beta	67.9 ± 6.2%	Fontcuberta-PiSunyer et al., 2023

With that said, the differentiation of GINS requires lentiviral transduction of NPM factors. The random integration of the transgene raises safety concerns for clinical translation. In this context, the employment of targeted genome editing techniques might solve the problem. One such example includes CRISPR/CAS9-mediated integration into the AAVS1 locus, which is considered to be a safe harbor (Hayashi et al., 2020) for genome editing. On the contrary, many recent approaches that assess genomic instability at high resolution have been developed and could address safety concerns due to random integration. For instance, copy number variation (CNV) can detect deletion or duplication with 1 KB to 5 MB resolution (Zhou et al., 2025; Pös et al., 2021). Indeed, this technique is already employed to address concerns due to CRISPR off-targets (Leavens et al., 2024). However, a small molecule protocol would be a preferred method, although the differentiation efficiency may be compromised (Baker et al., 2023).

In addition, the ongoing cell therapy trials still employ products that require patients to take systemic immunosuppressants. In this sense, reprogramming strategies offer a potential solution with autologous transplantation (Wang et al., 2024). However, the converted β-cells may still need protection from autoimmune attack.

Immune shielding strategies

As previously discussed, the field of regenerative medicine is developing a range of platforms for β-cell replacement therapies, including stem cell-derived β-cells and direct reprogramming strategies. In autologous transplants, immune rejection by the host is naturally avoided; however, autoimmune destruction—particularly relevant in T1D—remains a major challenge. For instance, GINS has been shown to survive for more than 6 months in immunocompromised mice (Huang et al., 2023). Nevertheless, whether GINS can survive after engraftment in patients with T1D remains unknown. To overcome this challenge, many research groups are now integrating immune evasion strategies with regenerative therapies to generate β-cells resistant to autoimmune attack.

One such strategy involves engineering grafted cells to express CD47, a surface protein that binds to SIRPα on macrophages and NK cells, delivering a “don’t-eat-me” signal that prevents phagocytosis and cytotoxicity (Xu et al., 2024). Another approach focuses on incorporating immune-inhibitory molecules to counteract both cellular and humoral immune responses. In this context, the Deuse Lab showed that synthetic engagers with agonistic functionality to their inhibitory receptors, such as TIM3 and SIRPα, can effectively protect engineered HLA-deficient stem cell-derived endothelial cells from NK cell and macrophage-mediated killing (Gravina et al., 2023). Furthermore, inclusion of a truncated version of CD64 helps shield the grafted cells from antibody-dependent cellular cytotoxicity (Gravina et al., 2023).

Additional studies have already shown that induction of PD-L1 overexpression alone or in combination with HLA class I knock out protects stem cell-derived β-cells from diabetogenic CD8⁺ T cells attack (Santini-González et al., 2022). A particularly promising direction involves leveraging a specific subset of CD4⁺ T cells known as regulatory T cells (Tregs). Under physiological conditions, Tregs play a critical role in maintaining immune tolerance by suppressing autoreactive immune responses through multiple, well-established mechanisms (Rajendeeran and Tenbrock, 2021). Notably, the Ferreira Lab has demonstrated that stem cell-derived β-cells expressing a truncated form of the epidermal growth factor receptor, which is biologically and immunologically inert, can activate Tregs to suppress cytotoxic T cells (Barra et al., 2024). Finally, Tregs can also be activated by specific cytokines (Zong et al., 2024); therefore, other approaches involve overexpression of IL-10, TGF-β, and IL-2, which can in turn activate Tregs and induce immune tolerance (Yu et al., 2015). In conclusion, combining strategies that protect against cytotoxic T cell attack with approaches that promote Treg-mediated immunosuppression may provide multilayered immune protection for β-cell replacement therapies.

Concluding Remarks

Further research in the direction of unraveling the developmental transcriptional mechanisms and gene regulatory networks, as well as chromatin landscapes and epigenomic rearrangement, promises to help build a more robust reprogramming outcome. This will offer an exceptional opportunity for cell replacement therapies for diabetes while allowing patients to choose among multiple systems for customized treatment of their unique condition.

Ethical Statement

This study was approved by the Weill Cornell Medicine Review Board.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Funding Information

No funding was received for this article.

References

Ackermann

, Moss

, Kaestner

. GABA and artesunate do not induce pancreatic α-to-β cell transdifferentiation in vivo. Cell Metab, 2018; 28(5):787–792.e3.

Adeghate

, Donath

. Morphological findings in long-term pancreatic tissue transplants in the anterior eye chamber of rats. Pancreas, 1990; 5(3):298–305.

Ahlgren

, Jonsson

, et al. β-cell-specific inactivation of the mouse Ipf1/Pdx1 gene results in loss of the β-cell phenotype and maturity onset diabetes. Genes Dev., 1998; 12(12):1763–1768.

Akinci

, Banga

, Tungatt

, et al. Reprogramming of various cell types to a beta-like state by Pdx1, Ngn3 and MafA. PLoS One, 2013; 8(11):e82424.

Al-Hasani

, et al. EZH2 inhibitors promote β-like cell regeneration in young and adult type 1 diabetes donors. Signal Transduct. Target. Ther, 2024; 9:30–32.

American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care, 2013; 36.

Angelo

, Guerrero-Zayas

, Tremblay

. A fate map of the murine pancreas buds reveals a multipotent ventral foregut organ progenitor. PLoS One, 2012; 7(7):e40707–e11.

Ariyachet

, et al. Reprogrammed stomach tissue as a renewable source of functional β cells for blood glucose regulation. Stem Cell, 2016; 18(3):410–421; doi: 10.1016/j.stem.2016.01.003

Bae

, Kim

, et al. Transcriptional regulation of glucose sensors in pancreatic β cells and liver: An update. Sensors (Basel), 2010; 10(5):5031–5053.

10.

Baker

, Arany

, Baur

, et al. CAR T therapy beyond cancer: The evolution of a living drug. Nature, 2023; 619(7971):707–715.

11.

Banga

, Akinci

, Greder

, et al. In vivo reprogramming of Sox9+ cells in the liver to insulin-secreting ducts. Proc Natl Acad Sci USA, 2012; 109(38):15336–15341.

12.

Barra

, Robino

, Castro-Gutierrez

, et al. Combinatorial genetic engineering strategy for immune protection of stem cell-derived beta cells by chimeric antigen receptor regulatory T cells. Cell Rep, 2024; 43(11):114994.

13.

Bellin

, Barton

, Heitman

, et al. Potent induction immunotherapy promotes long-term insulin independence after islet transplantation in type 1 diabetes. Am. J. Transplant., 2012; 12(6):1576–1583.

14.

Ben-Othman

, Vieira

, Courtney

, et al. Long-term GABA administration induces alpha cell-mediated beta-like cell neogenesis. Cell, 2017; 168(1–2):73–85.e11.

15.

Bideyan

, López Rodríguez

, Priest

, et al. Hepatic GATA4 regulates cholesterol and triglyceride homeostasis in collaboration with LXRs. Genes Dev, 2022; 36(21–24):1129–1144.

16.

Cerdá-Esteban

, Naumann

, Ruzittu

, et al. Stepwise reprogramming of liver cells to a pancreas progenitor state by the transcriptional regulator Tgif2. Nat Commun, 2017; 8:14127.

17.

Cerf

. Beta cell dysfunction and insulin resistance. Front Endocrinol (Lausanne), 2013; 4:37–12.

18.

Chakravarthy

, Gu

, Enge

, et al. Converting adult pancreatic islet α cells into β cells by targeting both Dnmt1 and Arx. Cell Metab, 2017; 25(3):622–634.

19.

Chen

Y-J

, Finkbeiner

, Weinblatt

, et al. De novo formation of insulin-producing ‘neo-β cell islets’ from intestinal crypts. Cell Rep., 2014; 6(6):1046–1058.

20.

Cheung

, Dayanandan

, Lewis

, et al. Glucose-dependent insulin release from genetically engineered K cells. Science, 2000; 290(5498):1959–1962.

21.

Choi

, Cayabyab

, Yoshihara

. A guide from the stomach to β cells. Nat Cell Biol, 2023; 25(5):637–638.

22.

Chondronasiou

, Martínez de Villarreal

, Melendez

, et al. Deciphering the roadmap of in vivo reprogramming toward pluripotency. Stem Cell Reports, 2022; 17(11):2501–2517.

23.

Clayton

, Osipovich

, Stancill

, et al. Pancreatic inflammation redirects acinar to β cell reprogramming. Cell Rep, 2016; 17(8):2028–2041.

24.

Colarusso

, Zhou

. Direct reprogramming of different cell lineages into pancreatic β-like cells. Cell Reprogram, 2022; 24(5):252–258.

25.

Collombat

, Xu

, Ravassard

, et al. The ectopic expression of Pax4 in the mouse pancreas converts progenitor cells into α and subsequently β cells. Cell, 2009; 138(3):449–462.

26.

Dahiya

, Saleh

, Rodriguez

, et al. Acinar to β-like cell conversion through inhibition of focal adhesion kinase. Nat Commun, 2024; 15(1):3740.

27.

de Koning

EJP

, Carlotti

. Human stomach tissue as alternative source of insulin-producing cells. Nat Rev Endocrinol, 2023; 19(9):503–504.

28.

Deutsch

, Jung

, Zheng

, et al. A bipotential precursor population for pancreas and liver within the embryonic endoderm. Development, 2001; 128(6):871–881.

29.

Efrat

. Cell replacement therapy for type 1 diabetes. Trends Mol Med, 2002; 8(7):334–339.

30.

Ferber

, Halkin

, Cohen

, et al. Pancreatic and duodenal homeobox gene 1 induces expression of insulin genes in liver and ameliorates streptozotocin-induced hyperglycemia. Nat. Med., 2000; 6(5):568–572.

31.

Fontcuberta-PiSunyer

, García-Alamán

, Prades

, et al. Direct reprogramming of human fibroblasts into insulin-producing cells using transcription factors. Commun Biol, 2023; 6(1):256; doi: 10.1038/s42003-023-04627-2

32.

Furuyama

, Chera

, van Gurp

, et al. Diabetes relief in mice by glucose-sensing insulin-secreting human α-cells. Nature, 2019; 567(7746):43–48; doi: 10.1038/s41586-019-0942-8

33.

Grapin-Botton

. Ductal cells of the pancreas. Int J Biochem Cell Biol, 2005; 37(3):504–510.

34.

Gravina

, Tediashvili

, Zheng

, et al. Synthetic immune checkpoint engagers protect HLA-deficient iPSCs and derivatives from innate immune cell cytotoxicity. Cell Stem Cell, 2023; 30(11):1538–1548.e4.

35.

Gribben

, Lambert

, Messal

, et al. Ductal Ngn3-expressing progenitors contribute to adult β cell neogenesis in the pancreas. Cell Stem Cell, 2021; 28(11):2000–2008.e4.

36.

Guo

, Zhang

, Lu

, et al. Specific reprogramming of alpha cells to insulin-producing cells by short glucagon promoter-driven Pdx1 and MafA. Mol Ther Methods Clin Dev, 2023; 28:355–365.

37.

Hayashi

, Kubo

, Izumida

, et al. Efficient viral delivery of Cas9 into human safe harbor. Sci Rep, 2020; 10(1):21474.

38.

Herrera

. Adult insulin- and glucagon-producing cells differentiate from two independent cell lineages. Development, 2000; 127(11):2317–2322.

39.

Huang

, Zhang

, Gao

, et al. Direct reprogramming of human fibroblasts to functional and expandable hepatocytes. Cell Stem Cell, 2014; 14(3):370–384.

40.

Huang

, Hickson

LTJ

, Eirin

, et al. Cellular senescence: The good, the bad and the unknown. Nat Rev Nephrol, 2022; 18(10):611–627.

41.

Huang

, Gu

, Zhang

, et al. Stomach-derived human insulin-secreting organoids restore glucose homeostasis. Nat Cell Biol, 2023; 25(5):778–786.

42.

Ieda

, Fu

J-D

, Delgado-Olguin

, et al. Direct reprogramming of fibroblasts into functional cardiomyocytes by defined factors. Cell, 2010; 142(3):375–386.

43.

Imai

, Katagiri

, Yamada

, et al. Constitutively active PDX1 induced efficient insulin production in adult murine liver. Biochem Biophys Res Commun, 2005; 326(2):402–409.

44.

Jeyagaran

, Urbanczyk

, Layland

, et al. Forward programming of hiPSCs towards beta-like cells using Ngn3, Pdx1, and MafA. Sci Rep, 2024; 14(1):13608.

45.

Kalo

, Read

, Ahlenstiel

. Reprogramming—Evolving path to functional surrogate β-cells. Cells, 2022; 11(18).

46.

Kawaguchi

, Cooper

, Gannon

, et al. The role of the transcriptional regulator Ptf1a in converting intestinal to pancreatic progenitors. Nat Genet, 2002; 32(1):128–134.

47.

Koizumi

, Nagai

, Kida

, et al. Forced expression of PDX-1 induces insulin production in intestinal epithelia. Surgery, 2006; 140(2):273–280.

48.

Leavens

, Osorio-Quintero

, Yeuteuh

, et al. Generation of a fluorescent mNeonGreen insulin reporter line in the H1 (WA01) hESC background. Stem Cell Res, 2024; 81:103559.

49.

Lee

, Sugiyama

, Liu

, et al. Expansion and conversion of human pancreatic ductal cells into insulin-secreting endocrine cells. Elife, 2013; 2:e00940; doi: 10.7554/eLife.00940

50.

Leibiger

, Berggren

. Intraocular in vivo imaging of pancreatic islet cell physiology/pathology. Mol Metab, 2017; 6(9):1002–1009.

51.

Lezmi

, Benvenisty

. The tumorigenic potential of human pluripotent stem cells. Stem Cells Transl Med, 2022; 11(8):791–796.

52.

, Nakanishi

, Zumsteg

, et al. In vivo reprogramming of pancreatic acinar cells to three islet endocrine subtypes. Elife, 2014a;3:e01846–e20. 2014.

53.

, Zhu

, Russ

, et al. Small molecules facilitate the reprogramming of mouse fibroblasts into pancreatic lineages. Cell Stem Cell, 2014b;14(2):228–236.

54.

May

, Kaestner

. Gut endocrine cell development. Mol. Cell. Endocrinol, 2010; 323(1):70–75.

55.

Mehta

, Hopson

, Smadi

, et al. Development of the human pancreas and its exocrine function. Front Pediatr, 2022; 10:909648.

56.

Meivar-Levy

, Ferber

. Liver to pancreas transdifferentiation. Curr. Diab. Rep., 2019; 19(9):76.

57.

Mukherjee

, French

, Gadue

. Loss of TBX3 enhances pancreatic progenitor generation from human pluripotent stem cells. Stem Cell Reports, 2021; 16(11):2617–2627.

58.

Pagliuca

, Millman

, Gürtler

, et al. Generation of functional human pancreatic β cells in vitro. Cell, 2014; 159(2):428–439.

59.

Pellegrini

, Zamarian

, Sordi

. Strategies to improve the safety of iPSC-derived β cells for β cell replacement in diabetes. Transpl Int, 2022; 35:10575–10515.

60.

Pös

, Radvanszky

, Buglyó

, et al. DNA copy number variation: Main characteristics, evolutionary significance, and pathological aspects. Biomed J, 2021; 44(5):548–559.

61.

Quesada

, Tudurí

, Ripoll

, et al. Physiology of the pancreatic α-cell and glucagon secretion: Role in glucose homeostasis and diabetes. J Endocrinol, 2008; 199(1):5–19.

62.

Rajendeeran

, Tenbrock

. Regulatory T cell function in autoimmune disease. J Transl Autoimmun, 2021; 4:100130.

63.

Rickels

, Paul Robertson

. Pancreatic islet transplantation in humans: Recent progress and future directions. Endocr Rev, 2019; 40(2):631–668.

64.

Ruiz

, Kirk

. Long-term toxicity of immunosuppressive therapy. In Transplantation of the Liver (3rd ed.). W.B. Saunders: Philadelphia, PA, 2015; pp. 1354–1363.

65.

Saleh

, Gittes

, Prasadan

. Alpha-to-beta cell trans-differentiation for treatment of diabetes. Biochem Soc Trans, 2021; 49(6):2539–2548.

66.

Salvioli

, Bovara

, Barbara

, et al. Neurology and neuropathology of the pancreatic innervation. Jop, 2002; 3(2):26–33.

67.

Sangan

, Jover

, Heimberg

, et al. In vitro reprogramming of pancreatic alpha cells towards a beta cell phenotype following ectopic HNF4α expression. Mol Cell Endocrinol, 2015; 399:50–59.

68.

Santini-González

, Castro-Gutierrez

, Becker

, et al. Human stem cell derived beta-like cells engineered to present PD-L1 improve transplant survival in NOD mice carrying human HLA class I. Front Endocrinol (Lausanne), 2022; 13:989815–989814.

69.

Shahbazi

, Moradi

, Nemati

, et al. Conversion of human fibroblasts to stably self-renewing neural stem cells with a single zinc-finger transcription factor. Stem Cell Reports, 2016; 6(4):539–551.

70.

Shen

, Horb

, Slack

JMW

, et al. Transdifferentiation of pancreas to liver. Mech Dev, 2003; 120(1):107–116.

71.

Shi

, Inoue

, Wu

, et al. Induced pluripotent stem cell technology: A decade of progress. Nat Rev Drug Discov, 2017; 16(2):115–130.

72.

Suzuki

, Nakauchi

, Taniguchi

. Glucagon-like peptide 1 (1–37) converts intestinal epithelial cells into insulin-producing cells. Proc Natl Acad Sci USA, 2003; 100(9):5034–5039.

73.

Talchai

, Xuan

, Kitamura

, et al. Generation of functional insulin-producing cells in the gut by Foxo1 ablation. Nat Genet, 2012; 44(4):406–412, S1.

74.

Thorel

, Népote

, Avril

, et al. Conversion of adult pancreatic α-cells to Β-cells after extreme Β-cell loss. Nature, 2010; 464(7292):1149–1154.

75.

Vecchio

, Tornali

, Bragazzi

, et al. The discovery of insulin: An important milestone in the history of medicine. Front Endocrinol (Lausanne), 2018; 9:613.

76.

Vierbuchen

, Ostermeier

, Pang

, et al. Direct conversion of fibroblasts to functional neurons by defined factors. Nature, 2010; 463(7284):1035–1041.

77.

Wang

, Du

, Zhang

, et al. Transplantation of chemically induced pluripotent stem-cell-derived islets under abdominal anterior rectus sheath in a type 1 diabetes patient. Cell, 2024; 187(22):6152–6164.e18.

78.

Watt

, Zhao

, Li

, et al. Development of the mammalian liver and ventral pancreas is dependent on GATA4. BMC Dev. Biol., 2007; 7:1–11.

79.

Wei

, Cui

, Jiang

, et al. Glucagon acting at the GLP-1 receptor contributes to β-cell regeneration induced by glucagon receptor antagonism in diabetic mice. Diabetes, 2023; 72(5):599–610.

80.

, et al. Expression of a mutant CD47 protects against phagocytosis without inducing cell death or inhibiting angiogenesis. Cell Reports Med, 2024; 5.

81.

Yamane

, Harada

, Inagaki

. Mechanisms of fat-induced gastric inhibitory polypeptide/glucose-dependent insulinotropic polypeptide secretion from K cells. J Diabetes Investig, 2016; 7(Suppl 1):20–26.

82.

, Snowhite

, Vendrame

, et al. Selective IL-2 responsiveness of regulatory t cells through multiple intrinsic mechanisms supports the use of low-dose IL-2 therapy in type 1 diabetes. Diabetes, 2015; 64(6):2172–2183.

83.

Zhang

, Ramracheya

, Lahmann

, et al. Role of KATP channels in glucose-regulated glucagon secretion and impaired counterregulation in type 2 diabetes. Cell Metab, 2013; 18(6):871–882.

84.

Zhou

, Dong

, Jiang

, et al. A copy number variation detection method based on OCSVM algorithm using multi strategies integration. Sci Rep, 2025; 15(1):3526.

85.

Zhou

, Brown

, Kanarek

, et al. In vivo reprogramming of adult pancreatic exocrine cells to β-cells. Nature, 2008; 455(7213):627–632.

86.

Zhu

, Russ

, Wang

, et al. Human pancreatic beta-like cells converted from fibroblasts. Nat Commun, 2016; 7:10080.

87.

Zong

, Deng

, Chong

. Regulation of Treg cells by cytokine signaling and co-stimulatory molecules. Front Immunol, 2024; 15:1387975–1387915.