The Potential of Pancreatic Organoids for Diabetes Research and Therapy

ABSTRACT

The success of clinical transplantation of pancreas or isolated pancreatic islets supports the concept of cell-based cure for diabetes. One limitation is the shortage of cadaver human pancreata. The demand–supply gap could potentially be bridged by harnessing the self-renewal capacity of stem cells. Pluripotent stem cells and adult pancreatic stem cells have been explored as possible cell sources. Recently, a system for long-term culture of proposed adult pancreatic stem cells in a form of organoids was developed. Generated organoids partially mimic the architecture and cell-type composition of pancreatic tissue. Here, we review the attempts over the past decade, to utilize the organoid cell culture principles in order to identify, expand, and differentiate the adult pancreatic stem cells from different compartments of mouse and human pancreata. The development of the culture conditions, effects of specific growth factors and small molecules is discussed. The potential utility of the adult pancreatic stem cells is considered in the context of other cell sources.

1. Introduction

One great breakthrough in the field of stem cell (SC) research is the recent development of the organoid cell culture system. It can be loosely described as an “organ in a dish,”¹ as it simulates organ-like growth in vitro.² Thus, the organoids gradually make solid organs accessible for stem cell research, which until recently was possible for hematopoietic tissue only.

The cell source in the organoid culture can either be pluripotent stem cells or primary epithelial stem/progenitor cells with the potential to differentiate into organ-specific cell types.^1,3,4 According to a generally accepted definition, organoids have the intrinsic capacity to expand cells and spontaneously grow into self-organized three-dimensional (3D) structures, which at least partially mirror tissue architecture, cell-type composition, and functionality of a given organ.^2,5–8 As elegantly demonstrated in the intestinal organoids, the stem cells differentiate into organ-specific lineages under specific culture conditions, which activate or inhibit specific signaling pathways.^9,10 The extracellular factors directing the fate of stem/progenitor cells were originally conceptualized for hematopoiesis as the “stem cell niche.”¹¹

The term organoid was first used in 1946 to describe the tissue of dermoid cystic “teratoma.”¹² Sometimes the term organoid is applied incorrectly to spheroids² or islet-like cell clusters, which also form 3D cellular structures, but the cells are not attached to the extracellular matrix and the culture media is different from the media used for the organoid culture.¹³

Organoids provide a useful tool for disease modeling,^4,14 drug testing,^15,16 or cancer research.¹⁷ They also represent an opportunity to study tissue–pathogen interaction in vitro.^18,19 Organoids can help in studying stem cell niches²⁰ and organ development.³ In this context, pancreatic organoids have been employed to investigate pancreatic ductal adenocarcinoma,^21,22 pancreas development,^23–25 or cystic fibrosis,²⁶ and to screen drugs targeting pancreatic diseases.^27,28 No information has yet been published on pancreatic organoids derived from diabetic patients in order to investigate the disease pathogenesis, or to screen candidate drugs for diabetes.²⁹ However, pancreatic organoids have been researched as the potential cell source for therapy of diabetes.

Diabetes is a chronic metabolic disease characterized by lost control over blood glucose. Type 1 diabetes is caused by an absolute deficiency of insulin. In type 2 diabetes, relative insulin deficiency results from an increased insulin resistance. At present, the type 1 diabetic patients are mostly managed by the administration of synthetic insulin.^30,31 Nevertheless, a small fraction of patients are already cured from diabetes by transplantation of beta cells within cadaver pancreas (more than 48,000 patients)³² or isolated pancreatic islets (more than 1900 allograft and autograft recipients).³³ Availability of the cell-based cure of diabetes is circumscribed by the limited amount cadaver pancreata. The demand–supply gap could potentially be bridged by beta cells derived from self-renewing stem cells. The three possible sources of SCs explored over the past two decades include the embryonic stem cells (ESC), induced pluripotent stem cells (iPSC), and adult pancreatic stem cells.

The first insulin-secreting cells derived from genetically modified ESCs were reported to normalize glycemia in streptozotocin-induced diabetic mice in 2000.³⁴ In the following years, the therapeutic potential of ESCs and iPSC has been extensively tested, and high efficacy multi-stage differentiation approaches were developed.^35,36 In 2014, Rezania and Pagliuca independently described protocols for in vitro differentiation of human pluripotent stem cells through pancreatic progenitors into beta cells that reversed diabetes when transplanted into mice.^37,38 However, the pluripotent stem cells carry the potential for mutagenesis, chromosomal aberrations, and carcinogenesis,³⁹ mainly due to the unphysiological number of cell divisions needed for meaningful therapeutic application.^40,41 The safety can be significantly improved by cell encapsulation or by molecules that degrade the remaining undifferentiated cells.⁴² While the encapsulation also avoids the requirement of immunosuppression, making it safer for patients, fibrosis of the capsule hinders proper vascularization (https://viacyte.com/press-releases/two-year-data-from-viacytes-step-one-clinical-trial-presented-at-ada-2018/ STEP ONE clinical trial). On the other hand, an open system capsule allows a better degree of vascularization at the cost of immunosuppression. Bioengineered pluripotent stem cell therapy and their bottlenecks were reviewed elsewhere.⁴³

Several biotechnology companies have already driven their successful research into pre-clinical and clinical trials.⁴⁴ Viacyte developed pancreatic progenitors, which differentiate from embryonic stem cells with almost 100% efficiency into islet cells, which are implanted into diabetic patients in closed or open capsule systems (Clinical trials NCT04678557, NCT03163511 https://www.clinicaltrials.gov/ct2/show/NCT04678557?term=viacyte&draw=1&rank=1 and https://www.clinicaltrials.gov/ct2/show/NCT03163511?term=viacyte&draw=1&rank=2).⁴⁵ Vertex recently registered a clinical trial to test SC-based islet cell therapy VX-880 in patients with hypoglycemia unawareness syndrome (NCT04786262 https://clinicaltrials.gov/ct2/show/NCT04786262?term=VX-880&draw=2&rank=1). Another clinical-stage company Kadimastem, which previously tested clinical grade ESC derived astrocytes (NCT03482050 https://clinicaltrials.gov/ct2/show/NCT03482050), has announced effectiveness of their microencapsulated islet-like clusters (IsletRx https://www.kadimastem.com/post/kadimastem-announces-successful-preclinical-results-of-its-cell-therapy-treatment-for-insulin-depend) in immunocompetent diabetic mice.

While pluripotent SCs appear to be a promising source of cells to cure diabetes, serious complications can still occur in clinical trials. It is therefore reasonable to consider alternative sources of insulin-producing cells, such as the adult pancreatic stem cells, which only recently became amenable to research thanks to the advancement of the organoid cell culture, as reviewed below.

2. Principles of pancreatic organoid culture

2.1. Basic principles of organoid culture

Organoid culture requires three key components: the extracellular matrix substitute, culture media, and source cells. The extracellular matrix substitutes provide specific attachment sites for cell adhesion molecules and three-dimensional support for the constituent cells. The media provide specific soluble factors to modulate signaling pathways or chromatin state. The cell source, from which pancreatic organoids are originated, can either be unselected, such as crude ductal fragments,^22,46–50 or highly selected single cells carrying putative stem cell markers.^{46,47,51–53} The capacity of organoids to expand over numerous passages proves the presence of stem/progenitor cells in the original preparation, as well as the ability of the culture conditions to sustain them.^46,47,54,55 By the same token, a failure to expand suggests either the absence of stem/progenitor cells or an inadequate media/matrix composition. When organoid cultures are initiated from unselected raw islet-depleted pancreatic tissue comprising multiple cell types, the organoids gradually prevail, while the other cell types diminish.⁵² The size of pancreatic organoids ranges between 50 and 2000 µm, dependent on the culture duration, as shown in Figure 1 top (we are grateful to Folia Biologica for the permission to reproduce data recently generated in our lab),⁵² and media composition. Once an organoid is established, some of the constituent cells can proliferate or differentiate, in response to the appropriate stimuli. Organoid cells are grown within a drop or a ring of Matrigel surrounded with medium, which is changed every 1–3 days. The passaging of organoids (generally every 7–10 days) consists of mechanical or enzymatic digestion of the basement membrane, releasing whole organoids that are further dissociated into small cell clusters or individual cells, that are subsequently replated in new Matrigel drop and culture medium. The three basic components of the organoid culture are discussed in the following chapters.

Figure 1.

Example of organoid culture established from adult human pancreatic CD133⁺ cells. Top: Phase contrast microscopy visualizing the expansion of organoids over two weeks. Scale bar: 500 µm. Bottom: Immunohistochemistry visualizing the constituent cells at Day 15, indicating the proliferation activities (Ki67) and cell types (duct, KRT19; epithelial, ECAD; endocrine, CHGA; pancreatic progenitors PDX1, SOX9), including the original CD133⁺ cells. Scale bar: 200 µm

2.2. Three-dimensional extracellular matrices

Extracellular matrix in 3D organoid culture system permits the growth and expansion of cells in both horizontal and vertical planes, thus distinguishing it from the standard 2D culture system. The mammalian cells grown on a flat surface as opposed to the three-dimensional space are subjected to different mechano-chemical cues, causing differences in cytoskeleton rearrangement, gene expression, cell shape, and function. In two-dimensional culture, no gradients of nutrients and signal molecules are possible. Also, the cells are forced to apical-basal polarity, reducing lateral cell-to-cell adhesions, which are critical for development and function of cells,⁵⁶ including beta cells.⁵⁷ Sensitivity to such mechano-chemical cues was demonstrated for stem cells and pancreatic progenitors.^58,59 In 3D culture systems, the individual cells utilize a range of surface adhesion molecules to reaggregate among themselves and to interact with fibrillar proteins of the matrix; consequently, chemical gradients are generated.^60–62 3D organoid cultures can further be enhanced by co-culturing with additional cell types, e.g. endothelial or mesenchymal cells, to provide signals, which at present cannot be delivered by defined chemical components.^1,63

The most commonly used artificial extracellular matrix substitute is Matrigel/BME (Basement membrane extract), which generates 3D scaffold by rapid spontaneous gelification at 37°C. This assortment of gelatinous proteins of the extracellular matrix derived from Engelbreth-Holm-Swarm tumor cell line was shown to support the stem cells' self-renewal potential and preserve their undifferentiated state.⁶⁴ The major Matrigel components comprise ~60% laminin, ~30% collagen IV, ~8% nidogen/entactin, and Heparan sulfate.^65,66 A number of poorly defined growth factors naturally bound to the matrix were reduced in some commercial variants (e.g. BME2) to improve chemical definition and standardization. Matrigel has two major limitations: the xenogeneic cancer cell line origin, hindering its medical use; and the batch to batch variation, hampering research reproducibility.^67,68

Synthetic matrices are devoid of growth factors, which are substituted in defined media, thus advancing the organoid field toward Good manufacturing practice (GMP). Defined binding sites for integrins and for other cell adhesion molecules are attached to branched polymers interconnected by a crosslinker, thus forming a hydrogel network. Three such compounds were tested in pancreatic organoid culture. First, the poly(ethylene glycol) (PEG)-based hydrogel was covalently functionalized with laminin-1, and substrates of the FXIIIa enzyme, enabling the crosslinking by Thrombin-activated factor XIIIa. Similar to Matrigel, this hydrogel also supported the morphology, cluster formation, and progenitor maintenance of pancreatic embryonic organoids, but was less potent.⁵⁴ Second, another PEG-based hydrogel was functionalized with Integrin receptor binding motif containing sequence Arg-Gly-Asp (RGD) while using maleimide for crosslinking. It supported the growth of intestinal organoids derived from human embryonic/induced pluripotent stem cells,⁶⁹ but the morphology of human pancreatic organoids was altered.⁷⁰ Third, a dextran polymer was functionalized with RGD and crosslinked with hyaluronic acid. It supported pancreatic organoid morphology and simplified the passaging by digestion with dextranase. However, the expansion was slow and limited to only 5–6 passages in 100 days.⁴⁸ Additionally, apoly-isocyanopeptide-based hydrogel functionalized with human recombinant laminin-111 supported organoids derived from adult liver,⁷¹ making it promising for developmentally related pancreatic organoids. Novel variants of synthetic hydrogels were recently reviewed elsewhere.⁷²

2.3. Basic medium for expanding pancreatic organoids

Culture media for pancreatic organoids were developed from the medium originally established by Sato for intestinal organoids.⁷³ Similarly, Sato’s medium was adopted for derivation of organoids from normal or tumor tissue of other digestive organs, including colon,⁷⁴ stomach,⁷⁵ and liver.⁷⁶ Sato’s medium comprises Advanced DMEM/F12 medium and three key growth factors, EGF, Noggin, and R-Spondin-1, hence ENR medium.⁷³ The rationale for selecting these growth factors is following. EGF was shown to potentiate proliferation, while suppressing differentiation of pancreatic endocrine embryonic progenitors during development and in vitro.⁷⁷ Noggin, a member of the transforming growth factor superfamily, inhibits the bone morphogenetic protein signaling pathway, which is fundamental for solid organ development,⁷⁸ including pancreas.⁷⁹ R-Spondin-1 is an agonist of LGR5 receptor⁸⁰ of the Wnt-β-catenin signaling pathway (Wnt),⁸¹ which, in turn, is essential for development and for self-renewal of several types of adult stem cells.^82,83

The Basic Medium for pancreatic organoids is the ENR Medium enriched by additional factors, such as FGF10, Nicotinamide, N-acetylcysteine, and B27 (contains 21 ingredients, mostly antioxidants, and enzymes, such as catalase and superoxide dismutase). FGF10, a natural product of mesenchymal cells, supports the proliferative capacity of PDX1⁺ (Pancreatic and duodenal homeobox 1) pancreatic progenitors,⁸⁴ and helps to integrate the growth and the differentiation during pancreatic development.⁸⁵ Nicotinamide is an inhibitor of sirtuin (a regulator of epigenetic silencing), and poly(ADP-ribose) polymerase (PARP), which regulate protein deacetylation and DNA repair.^86,87 Nicotinamide was shown to promote survival and differentiation of human pluripotent stem cells after individualization.⁸⁷ Within this review, we call it “Basic Medium.”

Basic Medium was variously supplemented with Wnt pathway ligand, WNT3A (WENR medium);⁷⁴ hormone Gastrin; Rho-associated protein kinase (ROCK) inhibitor, Y-27632; inhibitors of TGFβ pathway, A83-01 or SB431542; Prostaglandin E2; Hepatocyte growth factor; an inhibitor of histone deacetylases, Trichostatin A; an inhibitor of glycogen synthase kinase 3β, CHIR99021; and an activator of adenylyl cyclase, Forskolin, Tables 1, Table 2. Gastrin is produced by G-cells of the developing⁸⁸ and neonatal pancreas,⁸⁹ where it has documented proliferative activity. Inhibitor of the ROCK signaling pathway diminished the dissociation-induced apoptosis in embryonic stem cells in vitro.⁹⁰ The contribution of the individual soluble factors to the expansion capacity of pancreatic organoids was systematically evaluated over 6 months (20 passages) in Basic Medium supplemented with WNT3A, Gastrin, A83-01, and Prostaglandin E2.²² Both human and mouse organoids failed to expand beyond the passages 3–5, when one of the following factors was omitted: EGF, Noggin, R-Spondin-1, WNT3A, Prostaglandin E2, or Nicotinamide. The omission of A83-01 or Gastrin allowed for at least ten passages.²²The protein factors can be derived either from specifically designed cell lines, that express these factors and secrete them into the medium,⁹¹ or these factors can be added into the medium in the form of pure recombinant proteins, which are amenable for GMP.

Table 1.

Effect on organoid proliferation and differentiation by selected proteins and small molecules

Factor*	When the factor is absent from media	Reference no.
EGF (133 kDa, 1207 AA) promotes cell proliferation	• In adult organoids, human: reduced size and expansion (2–5 passages) • In fetal organoids, mouse/human: reduced proliferation, improved differentiation	46, 22 55
Noggin (58 kDa, 232 AA) inhibits Bone morphogenic protein pathway	• In adult organoids, human: reduced size and expansion (4 passages) • In adult organoids, mouse: reduced expansion (2 months) • In embryonic organoids, mouse: increased cystic morphology	22 46 54
R-Spondin-1 (~28kDa, 263 AA) activates Lgr5 in Wnt/β-catenin pathway	• In adult organoids, human: reduced size and expansion of (3 passages) • In adult organoids, mouse: reduced expansion (2–5 passages)	22 46
WNT3A (39 kDa, 352 AA) activates Wnt/β-catenin signaling pathway	• In adult organoids, human: reduced expansion (approx. 3 passages)	22
FGF10 (23 kDa, 208 AA), mesenchymal factor	• In mouse adult organoids: reduced expansion (2–5 passages) • In fetal organoids, mouse/human: slower expansion • In mouse embryonic organoids: reduced acinar diff. and Pdx1 expression • In mouse embryonic organoids: no effect after 4 days	46 55 54 54
FGF1 (17 kDa, 155 AA), mesenchymal factor	• In embryonic organoids, mouse: improved endocrine differentiation • In embryonic organoids, mouse: diminished Pdx1 expression	54
VEGF (27 kDa, 232 AA), vascular endothelial growth factor	• In adult organoids, human: reduced engraftment efficiency (<3 months)	48
Nicotinamide (MW 122.12) a vitamin B3 form	• In adult organoids, human: reduced expansion (4 passages) • In adult organoids, mouse: reduced expansion (<2 months)	22 46
Y-27632 (MW 247.34) inhibits ROCK (Rho-associated protein kinase)	• In fetal organoids, mouse/human: cell proliferation dramatically decreased • In embryonic organoids, mouse: reduced org. formation, none Pdx1+	55 54
A83-01 (MW 421.52) inhibits TGFβ pathway	• In adult organoids, human: reduced expansion (from 20 to 10 passages)	22
Prostaglandin E2 (MW 352.47)	• In adult organoids, human: reduced expansion (5 passages)	22
*) UniProtKB/Swiss-Prot: https://genecards.weizmann.ac.il; MW, molecular weight.

Table 2.

Expansion of pancreatic organoids; cell origin, media composition, and results

Cell selection	Basic Medium modifications*	Cultivation time, No. of passages	Doubling time [hrs]	Reference no.
Mouse, hand-picked duct fragments	Gastrin	10 months	~60	46
Mouse, Lgr5+	Gastrin, ROCK inhibitor	>4 months	-
Mouse, Ptf1A+	Gastrin	3–4 passages	-
Mouse, EpCAM+ TSQ+	Gastrin, ROCK inhibitor	1 month, no proliferation	-
Mouse, hand-picked single ducts	Gastrin	-	-	49
Human, islet-depleted fragments	Gastrin, A83-01; w/o Nicotinamide	10 passages	67	47
Human, ALDHhigh (from organoids)	Gastrin, A83-01, ROCK inhibitor; w/o Nicotinamide	-	-
Human, hand-picked duct fragments	WNT3A, Gastrin, A83-01, Prostaglandin E2	20 passages, 6 months	-	22
Human, CD133+	w/o: B27, N-acetylcysteine	>3 months	-	53
Human, CD133+	Prostaglandin E2, HGF, Trichostatin A, CHIR99021, SB431542	>5 months	72	52
Human, islet-depleted fragments	Gastrin, A83-01, Prostaglandin E2, Forskolin	~5 passages, 70 days	73	50
Mouse, Procr+ cells from islets	B27, ITS, EGF, FGF2, heparin, endothelial cells; w/o:Basic Medium	20 passages, 6 months	-	161
*Basic Medium: Advanced DMEM/F12, B27, N-acetylcysteine, EGF, Noggin, R-Spondin-1, Nicotinamide, FGF10

2.4. Media for pancreatic organoid differentiation

Unlike the intestinal organoids, no universal medium has yet been developed for differentiation of pancreatic organoids toward beta cells. Several approaches were tested. The simplest one is the exclusion or reduction of growth factors that stimulate cell proliferation, such as EGF or R-Spondin-1.^{46,47,52,53,55} Nicotinamide, known to induce differentiation and maturation of fetal pancreatic endocrine cells in non-organoid culture,⁹² is commonly used in organoid culture.^47,52,53,93 The genes required for differentiation of organoid stem cells are presumably inaccessible for the developmental transcription factors. The accessibility of such genes can be improved by chromatin state modulation, using small molecules inhibiting nuclear epigenetic modifiers, such as DNA methyltransferases.⁵² More direct approach employed in some pancreatic organoid studies is transdifferentiation in vitro by introducing the key transcription factors by viral^49,53 or non-viral⁵² means. Detailed information about differentiation approaches are described in the chapter Organoids derived from Adult Pancreas.

3. Lessons from Embryogenesis and adult organ regeneration

Organoids derived either from embryonic or adult pancreatic tissue could potentially serve as a model system to accelerate the study of some aspects of pancreas and beta cell development. This is particularly important for research of human pancreas, given the limited accessibility. The designing of organoid culture systems draws from the knowledge obtained from studies of the development and regeneration of pancreas and other epithelial organs.

Pancreas develops from dorsal and ventral buds of the foregut endoderm in close contact with endothelium.^94–97 The early cellular development as well as the late maintenance of the differentiated state are orchestrated by a hierarchical cascade of stage-specific combinations of transcription factors.^98–100 The development of pancreas is divided into three major stages: a primary transition (E9.5–12.5), a secondary transition (E12.5-birth), and the early postnatal period until weaning.¹⁰¹ During the primary transition period, the pancreatic progenitors proliferate within thickening endoderm, which comes into a close proximity of endothelial cells of dorsal aortae.¹⁰² The endothelial cells initiate evagination of endoderm,⁹⁴ which leads to tubulogenesis and branching. During this period, bipotent pancreatic progenitors form the stalk/trunk domains, which give rise to ductal and endocrine cells, while multipotent progenitors present in the tip domains generate endocrine, ductal, and acinar cells.¹⁰³ Most of the endocrine cell differentiation occurs within the secondary transition period, hand in hand with the extensive exocrine differentiation. Within the secondary transition period, extensive exocrine and most of the endocrine cell differentiation occurs. Multipotent progenitors of the tip domains lose their multipotency and differentiate into acinar cells.⁹⁹ In the original model of islet formation, the differentiating Ngn3⁺ endocrine progenitors, derived from bipotent trunk progenitors, undergo epithelial-mesenchymal transition allowing for cell delamination from the ducts, migration toward blood vessels and aggregation within newly formed islets.¹⁰¹ However, recent model omits both the epithelial-mesenchymal transition and cell migration. In the course of islet formation, the cell contacts between the precursors are maintained, bi-layered nascent ”peninsulas” are formed, where alpha cells first develop in the outer layer (E13.5), while beta cells gradually appear beneath them, attached to the epithelial cord producing differentiating cells (E14.5). This spatiotemporal collinearity gradually leads to the core-mantel architecture typical for mature mouse islets.¹⁰⁴ Functionally, mouse beta cells mature only after birth, coincident with weaning.¹⁰⁵ In the unique human islet architecture, all endocrine cells are attached to blood vessels, heterologous alpha–beta cell contacts are favored, and homologous beta–beta cell contacts are permitted.¹⁰⁶ In the formation of human islets, a coalescence of few small ”peninsulas” was proposed.¹⁰⁴

Since the 1880s, adult pancreatic cells were considered undividing, but the capacity of pancreas to regenerate remained tenable, due to the early observations of mitotic figures in pancreatic cells, after partial pancreatectomy.¹⁰⁷ In humans, beta cells noticeably increase their numbers only in the first few years of life and during pregnancy.^108,109 In 1993, Bonner-Weir proposed two pathways of pancreas regeneration after 90% pancreatectomy in mice, when observing a) the replication of preexisting endocrine and exocrine cells, and b) the proliferation and differentiation of ductal cells, forming new lobules, which contained new beta cells, reminiscent of embryonic development.¹¹⁰ Two decades later, the regeneration of adult beta cells remains unsettled. Some genetic cell tracing studies support the original hypothesis¹¹⁰ of beta cell neogenesis from adult ductal progenitors after injury inflicted by duct ligation^111,112 or genetic cell targeting.¹¹³ Other fate mapping studies failed to find evidence for beta cell neogenesis from adult pancreatic progenitors.^{111,114–119} Additional genetic cell tracing studies combined with pancreatic duct injury in rodents convincingly demonstrated the replication of beta cells as the dominant regenerative mechanism during adult life^120,121 even after partial pancreatectomy¹²⁰ or targeted beta cell depletion.¹²² Significance of these complex experiments has been thoroughly reviewed elsewhere.^{107,123–131} Finally, transdifferentiation of alpha to beta cells was identified as a source of beta cell regeneration after near-total beta cell ablation in adult mice.¹³² After puberty, this capacity of alpha cells was not altered by age until senescence.¹³³ In juvenile age, dedifferentiation of delta cells was observed, accompanied by subsequent proliferation and differentiation into beta cells after almost complete beta cell ablation.¹³³

An unexpected heterogeneity of regeneration strategies was discovered among epithelial organs, ranging from the presence of multiple types of adult stem cells within adult tissues to the natural plasticity of differentiated cells, which de-differentiate into stem cells and in turn are capable of tissue regeneration.¹³⁴ The existence of facultative adult progenitors was proposed in some tissues, including liver and pancreas. A picture is emerging, where the Lgr5⁺ crypt collumnar cells residing at the bottom of the intestinal crypt serve as the intestinal stem cells under normal conditions, while under other conditions, such as radiation injury, the radiation-resistant +4 cells replenish not only the differentiated intestinal cell population, but also lost Lgr5⁺ crypt collumnar stem cells.¹³⁵ LGR5 was also used as a positive selection marker to generate adult pancreatic organoids.⁴⁶

4. Organoids derived from embryonic and fetal pancreatic progenitors

Unlike in adult pancreas, the existence of pancreatic progenitors in the embryonic and fetal tissues is undoubted. The following three studies^54,55,93 represent the first attempts to apply the principles of organoid culture in order to replicate the main features of pancreas development in vitro.

Greggio⁵⁴ optimized culture conditions and achieved expansion and partial differentiation of pancreatic organoids. The organoids originated from single cells or small groups of multipotent pancreatic progenitors isolated from embryonic (E10.5) mice. The mice were genetically modified (e.g. Ngn3-EYFP, Pdx1-nGFP, Sox2-Cre x R26R-lacZ, Sox2-Cre x R26R-YFP, Neurog3 knockout) in order to visualize the lineage tracing with fluorescent or histochemistry labels, in time lapse experiments. The efficiency of organoid structure formation was found to depend on the number of Pdx1⁺ cells per cluster, giving rise to the organoid (100% efficiency was achieved when at least 12 Pdx1⁺ cells were present per cluster),⁵⁴ suggesting the community effect.¹³⁶ Within seven days, the largest organoids formed lobulated tubes with duct cells marked with mucin. The periphery of terminal buds was crowned with PTF1A⁺/SOX9⁺/PDX1⁺ cells, giving rise to acinar cells (15–20%). In the organoid core bipotent progenitor cells (HNF1B⁺) differentiated into polarized ductal cells with rare endocrine differentiation. The frequency of mono-hormonal C-peptide⁺ cells increased up to ~0.7% by omission of FGF1, which however was necessary in the first four days for the proliferation of Pdx1⁺ progenitor cells. Similarly, ROCK inhibitor was indispensable for the survival of progenitors expressing Pdx1. When the organoids were dissociated and transplanted into E13.5 pancreatic explants (a transplantation assay), the endocrine differentiation increased within 10 days up to 4%, indicating an incompleteness of the artificial niche provided by the in vitro culture conditions.⁵⁴

Sugiyama⁹³ described an alternative protocol for partial reconstitution of pancreas development, using spherical organoids. Transgenic mice (e.g. Sox9-eGFP/Ngn3-tdTomato, MIP-GFP) were used for cell selection and tracking. FACS-sorted fetal (E11.5) NGN3⁻/SOX9⁺ multipotent progenitors (expressing also Pdx1, Hnf6, Tcf2 and Hes1, but not endocrine markers) were seeded at clonal density and cultured in Matrigel mixed with mesenchymal cells in PrEBM medium supplemented with FGF10, IGF1, Retinoic acid, insulin, and Transferrin. While the organoids expanded for three passages only, the progenitors here differentiated into multilayered spheres, with the inner layer of mucin⁺ duct-like cells, and the periphery with SOX9⁻ cells. A subset of SOX9⁻ cells expressed the endocrine markers, e.g. Ngn3, insulin, C-peptide, and glucagon. The endocrine cells on the outer surface were separated from the inner lining of mucin⁺ cells by another cell layer. The differentiation into glucose-responsive insulin-secreting cells was enhanced by a combination of Nicotinamide, physiological level of oxygen, reduced concentration of FGF10, and in the presence of mesenchymal cells. Uniquely, twofold increase in glucose-stimulated (3 vs. 20 mM) C-peptide secretion was observed in vitro.⁹³

Bonfanti⁵⁵ took advantage of transgenic mice (Pdx1-eGFP x Ins1-mRFP) to prospectively select and monitor fetal (E12 or E13) pancreatic progenitors in order to study the differentiation dynamics in organoids derived from a single cell.⁵⁵ Unlike the previous group, the authors demonstrated a high efficiency of organoid formation from individual progenitor cells, however with only partial endocrine differentiation.

In the same paper, Bonfanti investigated the impact of EGF on human fetal pancreatic organoids with the conclusion that EGF potentiates the organoid expansion while suppressing the differentiation toward endocrine fate.⁵⁵ Human progenitors were isolated after abortion 8–11 weeks post conception; mesenchymal cells were removed and the remaining epithelial tissue was digested, dissociated, mixed with Matrigel and cultured in Basic Medium supplemented with Gastrin and ROCK inhibitor. In the presence of EGF, the organoids expanded for at least 5 months, without changes in their cystic morphology. The organoid architecture contained ductal structures with cell polarization, as detected by MUC1 expression. In the absence of EGF, the organoids grew smaller, some assumed dense rather than cystic morphology, and the possible passage number dropped to ten. Significantly, a spontaneous differentiation toward endocrine fate (e.g. chromogranin A, insulin, glucagon, somatostatin) and acinar fate (e.g. PTF1A) was observed. The authors also observed a decreased proliferation rate after excluding ROCK inhibitor, FGF10, or R-Spondin-1. Their findings demonstrated a similar response between human and mouse fetal progenitors to the same environmental cues.⁵⁵

5. Organoids derived from adult pancreas

Enzymatic digestion of adult pancreas releases endocrine islets, and non-endocrine fragments comprising the acinar, centro-acinar, and ductal components, which somewhat differ in the tissue density.¹³⁷ Enrichment of the starting material with the prospective progenitor cells was achieved using techniques, such as tissue fragment separation on density gradient, hand-picking of duct fragments, and single cell sorting using putative markers of the proposed pancreatic stem/progenitor cells, as summarized in Table 2. The mouse and human organoid studies discussed below explored the utility of cell surface markers LGR5, CD133, and PROCR, as well as intracellularly expressed cell markers ALDH and PTF1A. Beta cell differentiation was induced by media composition and/or manipulating key transcription factors, as summarized in Table 3.

Table 3.

Differentiation of expanded organoids from adult murine and human pancreata

Cell source	In vitro differentiation	In vivo differentiation	Maturation by histology	C-peptide secretion	Reference no.
Mouse, EpCAM+ TSQ− CAG+	-	Re-aggregated with embryonic cells, transplanted under kidney capsule	2.5% insulin+ (all mono-hormonal)	In mice: detectable, response not tested	46
Mouse, hand-picked single ducts	Lentiviral delivery: Pdx1, MafA, Ngn3-WT	-	7% insulin+	-	49
	Lentiviral delivery: Pdx1, MafA, Ngn3-phosphomutant	-	28% insulin+ (33% coexpressed somatostatin)	-
	Plus medium modification: ROCK inhibitor; IWP-2, w/o R-Spondin-1, EGF for the last 2 days	-	61% insulin+ (41% coexpressed somatostatin)	-
	Plus last 2 days the three genes turned off	-	61% insulin+ (87% mono-hormonal)	-
Human, islet-depleted fragments	Nicotinamide containing medium	Transplanted under kidney capsule in normoglycemic or hyperglycemic mice	1.5% insulin+ (mono-hormonal)	In mice: detectable, not glucose-responsive	47
Human, CD133+	Adenoviral delivery: PDX1, MAFA, NGN3, PAX6; Medium: Retinoic acid; w/o R-spondin	Transplanted under kidney capsule	7–11% insulin+ (all mono-hormonal)	In vitro: not glucose-responsive; In mice: detectable*	53
Human, CD133+	mRNA delivery: NGN3; Medium: RepSox, PP2, ISX-9, GSK126, 5-aza-2´-deoxycytidin, Forskolin	-	5% insulin+ (some coexpressed somatostatin)	In vitro: not glucose-responsive	52
Mouse, Procr+ cells from islets	B27, ITS, EGF, FGF2, heparin, endothelial cells	Organoid transplanted under the kidney capsule were already differentiated	in vitro: 30% mono-hormonal insulin+cells, 70% bi-hormonal insulin+ cells	In vitro: glucose-responsive, In mice: glucose-responsive	161
*After 3 g/kg of glucose, after o/n starving, there was a trend to incresed C-peptide levels, but statistical significance is not reported.

Azzarelli⁴⁹ transdifferentiated mouse ductal organoids in vitro using lentiviral-mediated overexpression of transcription factors Pdx1, MafA, and Neurogenin3. The organoid culture was initiated from hand-picked ducts. Neurogenin3 was administered in two forms: the wild type and a more stable phospho-mutant. The best percentage of transdifferentiation into insulin positive cells (28% versus 7% in the wild type) was observed in organoid cells treated with the stable mutant of Neurogenin3. Significantly, the differentiation further increased up to 61% of insulin positive cells, when Wnt pathway was inhibited (addition of IWP-2, omission of R-Spondin-1) and EGF was removed for the last two days of culture. Moreover, when viral expression of all the transcription factors was turned off, the proportion of mono-hormonal insulin-producing cells increased. In spite of this success, the measurement of glucose responsiveness was inconclusive due to the small number of organoid cells.⁴⁹

A potential selection marker for adult pancreatic progenitors is LGR5, a plasma membrane receptor for R-Spondin-1, involved in the signaling of canonical Wnt pathway.¹³⁸ Its deletion is neonatally lethal.¹³⁹ It was identified in stem cells of the small intestine, colon,⁸³ and in long-lived cycling stem cells in hair follicle.¹⁴⁰ Lgr5⁺ cells in the intestinal crypt were shown to generate all epithelial lineages over more than a year of follow up in vivo study.¹⁴¹ Additionally, liver injury induced the expression of Lgr5 on proposed facultative liver progenitors, as documented by their capacity to expand in vitro in a Wnt-pathway-dependent organoid culture.⁷⁶ Huch⁴⁶ stimulated the emergence prospective facultative progenitors in vivo by applying the partial duct ligation model in genetically modified adult mice (e.g. ECad-CFP, CAG-EGFP, Lgr5-LacZ). After pancreatic duct ligation, a population of Lgr5⁺ stem/progenitor cells appeared in the ducts. Huch⁴⁶ concluded that adult Lgr5⁺ cells are bipotential progenitors capable of endocrine and ductal differentiation. In this study, hand-picked ductal fragments were embedded in Matrigel and grown in Basic Medium in Matrigel and grown in Basic Medium supplemented with Gastrin. Ductal fragments formed budding cyst-like organoids composed of duct-like progenitors (expressing Sox9, Pdx1, Muc1, Krt19). The complete medium allowed the organoids expansion for more than ten months, while the omission of EGF, Noggin, R-Spondin-1, FGF10, or Nicotinamide grossly reduced the number of possible passages. The capacity of FACS-sorted single Lgr5⁺ cells to form organoids was demonstrated with a colony-forming efficiency of 16%, that was in agreement with organoid studies of other digestive organs.⁷⁵ The potential of the Lgr5⁺ pancreatic progenitors to differentiate into ductal and endocrine lineages was demonstrated in vivo using organoids derived from single epithelial non-endocrine cells (EpCAM⁺ TSQ⁻). After 6 weeks of expansion, the organoids were dissociated, re-aggregated, and mixed with late embryonic pancreatic cells (E13.5). Additional stem cell niche signals were provided by transplanting this cell mixture under the kidney capsule of immunodeficient mice. Mainly, differentiated ductal (KRT19⁺) were observed. Histology also revealed 5% of endocrine cells (synaptophysin⁺), half of which were mono-hormonal insulin⁺ and C-peptide⁺. However, C-peptide in the plasma was not reported.⁴⁶

Adult Ptf1a⁺ acinar cells in vivo were shown to regain aspects of embryonic multipotentiality under injury, and subsequent conversion into mature beta cells was revealed.¹⁴² PTF1A (Pancreas Transcription Factor 1a) is selectively expressed in pancreas, retina, spinal cord, brain, and enteric nervous system. It is indispensable in controlling the expansion of multipotent progenitor cells as well as the specification and maintenance of the acinar cells.¹⁴³ Huch⁴⁶ took advantage of the organoid culture in order to substantiate the hypothesis of the acinar cells as the adult pancreatic progenitors. The authors used PTF1A marker to select single cells for the organoid formation, but their expansion ceased four passages later.

Loomans⁴⁷ identified in adult human pancreatic organoids an abundant cell population characterized by a high aldehyde dehydrogenase (ALDH) activity, and proposed it as a new marker for adult pancreatic progenitors. ALDH/RALDH (Aldehyde dehydrogenase) participates in All-trans Retinoic acid synthesis, which regulates gene expression by activating specific nuclear receptors during development in various tissues, including pancreas.^144,145 In the organoid study by Loomans et al., the transcription profile of ALDH⁺ cells (CPA1, PDX1, MYC, and PTF1A) corresponded with multipotent embryonic progenitors described by Zhou.¹⁰³ ALDH^high cells constituted a quarter of the cells in the primary organoids, which were derived from fragments of islet-depleted pancreatic tissue and cultured in Matrigel and Basic Medium supplemented with Gastrin and TGFβ inhibitor (A83-01), with Nicotinamide omitted. ALDH^high cells were predominantly localized at the tips of the budding organoids, which expressed mucin-1 at the luminal side and were maintained for at least 10 passages. ALDH^high cells selected from the primary organoids formed secondary cyst-like colonies, which in turn also expanded, suggesting stemness of the original cells. No endocrine differentiation of the secondary colonies was reported. A small cell fraction (0.5%) in the primary organoids differentiated in vitro into insulin positive cells, when Nicotinamide containing differentiation medium was used.¹⁴⁶ No glucagon positivity was reported. One month after the primary organoids were transplanted under the kidney capsule of either normoglycemic or hyperglycemic immunodeficient mice, 1.5% of organoids cells were insulin positive. Human C-peptide was detectable in mouse plasma, but was not glucose-responsive, indicating failed maturation in vivo. Glucagon was detected in mono-hormonal cells only.⁴⁷ While ALDH⁺ cells were scarcely detectable in normal adult pancreas,¹⁴⁷ they become abundant in regenerating conditions in human (patients with pancreatitis or T1D) as well as in mice (early postnatal, pregnancy),¹⁴⁸ corroborating their facultative progenitor status. In this context, Rovira,¹⁴⁷ by means of the centro-acinar/terminal ductal cells expressing ALDH1, published the first attempt to utilize the principles of mouse organoid culture in order to identify the putative adult pancreatic progenitor cell. Sorted ALDH positive cells were used to generate clonal pancreatospheres. The spheres contained cells co-expressing ALDH and SOX9, suggesting self-renewal capacity, which however was supported by only miniscule expansion (three passages). Interestingly, spontaneous C-peptide secretion occurred, but glucose responsiveness was tested at 0 vs. 11 mM glucose levels (the standard glucose testing levels are 3 vs. 20 mM), nor was it statistically evaluated.¹⁴⁷

Another potentially useful selection marker for adult pancreatic progenitors is CD133 (AC133, Prominin-1), a plasma membrane protein with a large extracellular loop. It was identified in undifferentiated embryonic stem cells,¹⁴⁹ and organ-committed stem cells,¹⁵⁰ such as hematopoietic¹⁵¹ and neural¹⁵² stem cells. Histology of pancreas demonstrated the presence of CD133⁺ cells within adult ducts.^153,154 Several groups utilized it for positive selection of the prospective bipotent or multipotent pancreatic progenitors, employing antibody-based FACS^53,155–157 or MACS^52,158 cell sorting.

Lee⁵³ transdifferentiated human CD133⁺ cells derived pancreatic organoids into insulin-secreting cells, using ectopic overexpression of four principal islet transcription factors (PDX1, MAFA, Neurogenin3, and PAX6). The CD133⁺ cells, selected by FACS from islet-depleted pancreatic tissue, were positive for the ductal marker cytokeratin-19, and negative for acinar and endocrine markers. Single-layer organoids were expanded for up to 3 months in Basic Medium with the omission of B27 supplement. The transdifferentiation protocol involved adenoviral vectors expressing the transcription factors inspired by Zhou.¹⁵⁹ After a few days of R-Spondin-1 withdrawal and supplementation with Retinoic acid, the organoids were cultured for subsequent two weeks in a differentiation medium, composition of which was less important than the timing. The differentiated spheres comprised 7–11% insulin positive cells, which secreted C-peptide into the medium, irrespective of the glucose level in the range 2–11 mM. A trend toward glucose-responsiveness (2.4×) was observed only when glucose was increased from starvation level (0.1 mM). However, C-peptide response to non-glucose secretagogues (KCl, sulphonylurea) was observed. After transplantation of transdifferentiated spheres under the kidney capsule of non-diabetic immunodeficient mice, human C-peptide was detectable in plasma for two weeks and a trend toward glucose responsiveness was observed (statistics for these functional observations was not provided).⁵³

Koblas⁵² achieved reprogramming of human CD133⁺ organoid cells into insulin-producing cells by the means of non-viral non-integrative introduction of Neurogenin3 in a combination with small molecules altering signaling pathways and epigenetic state. CD133⁺ cells were immunomagnetically separated from islet-depleted pancreatic tissue and cultivated in Matrigel with Basic Medium replenished with HGF, Prostaglandin E2, Trichostatin A, CHIR99021, and ALK5 inhibitor (SB431542). Single-layer organoids were expanded for at least 5 months. Neurogenin3 was introduced in the form of synthetic mRNA. Differentiation medium contained RepSox, PP2, ISX-9, GSK126, 5-aza-2´-deoxycytidin, and Forskolin. After the differentiation, the organoids comprised ~40% chromogranin A positive cells, including 5% insulin positive cells. C-peptide was detected in the medium at low levels (1/10³ of that produced by the same amount of islets), which failed to respond to glucose challenge, possibly due to insufficient expression of MAFA and PAX6. However, the secretion of C-peptide was KCl-responsive. A double-hormonal subpopulation co-expressing insulin and somatostatin was observed, further corroborating the immature character of insulin-producing cells. Glucagon positive cells were not observed.⁵² The non-integrative approach taken in this study avoids virus-induced inflammatory response and oncogenic transformation.¹⁶⁰

The above discussed studies achieved various degrees of expansion of pancreatic organoids derived from different subpopulations of pancreatic duct cells. The maximum length of the growth of human organoids reached six months (20 passages), while ten months for mice. Cyst-like oval shape organoids generally comprised a single layer of predominantly undifferentiated cells oriented around a central lumen (rather than elongated ducts),^22,46–53 as exemplified in Figure 1. The culture conditions and the results of the presented studies are summarized in Table 1–3.

Wang et al.¹⁶¹ uniquely searched for the adult endocrine progenitors within the pancreatic islets. The authors identified a novel population of Procr⁺ progenitors within adult mouse pancreatic islets, using single-cell RNA sequencing (scRNA-seq), and demonstrated their capacity to ameliorate hyperglycemia in diabetic mice. PROCR (EPCR, CD201) is an endothelial receptor for protein C, which was previously identified on hematopoietic stem cells,¹⁶² cultured cord blood cells,¹⁶³ blood vascular endothelial stem cells,¹⁶⁴ and breast cancer cells.¹⁶⁵ More than 7000 individual cells isolated from islet-enriched preparation were analyzed by scRNA-seq and mapped to clusters representing ten known cell types (alpha, beta, delta, PP, duct, acinar, endothelial, immune, mesenchymal, and stellate cells), and an additional cluster of previously unknown cells was identified. Unique signature genes in this cluster included Wnt pathway agonist gene R-spondin-1 and Wnt target gene Procr, suggesting stem cell character. These Procr⁺ islet progenitors had the capability to form expanding organoids, to differentiate into four endocrine cell types including insulin-producing beta-like cells, which ameliorated diabetes after transplantation. In Procr-mGFP-2A-LacZ mouse model, Procr⁺ cells were identified from all pancreas exclusively in the islets, each containing only few such cells. The lineage tracing in adult Procr-CreERT2; RosaConfetti mice several months after tamoxifen pulse revealed clones comprising approximately seven cells. In 70% of the clones, all endocrine cell types were present, while 30% of the clones contained beta cells only. One cell in each clonal population was hormone negative, which was the proposed progenitor cell. When FACS-isolated Procr⁺ cells were plated at a clonal density in serum-free medium supplemented with B27, ITS, EGF, FGF2, and heparin, one out of 15 cells formed a colony (Procr⁻ cells failed to form colonies). However, Procr⁺ cells could not be maintained for more than 7 days. Co-culture with endothelial cells was revealed as the key factor allowing long-term culture (more than 20 passages) of Procr⁺ derived organoids. At 15th passage, the organoids contained all hormone⁺ cell types, and secreted insulin and C-peptide (approximately tenfold and fivefold, respectively) into the medium in the glucose-regulated manner. When 1000 organoids were transplanted under the kidney capsule of diabetic mice, glucose level was ameliorated below 10 mM, which was similar to the control group of diabetic mice transplanted with 300 natural islets. One month after transplantation, intraperitoneal glucose tolerance test (IPGTT) demonstrated the improvement of glucose tolerance to a similar degree as 300 islets transplanted to a control group. Graftectomy after 125 days led to an abrupt blood glucose increase to 25 mM. The clustering of scRNA-seq data from explanted graft revealed alpha, beta, delta, and PP cells matching well with the primary islet cell clusters.¹⁶¹

6. Potential clinical application of adult pancreatic organoids

The first and the only reported attempt to develop a large-scale GMP-level manufacturing procedure for the expansion of organoids for clinical purposes comes from Dossena et al.⁵⁰ The enzymatic digestion was replaced by a mechanical one, thus avoiding manual duct picking. The organoids were expanded using BME2 matrix and Basic Medium supplemented with Gastrin, Prostaglandin E2, Forskolin, and A83-01. The cystic morphology and the expression profile (positivity for SOX9, PDX1, and MUC) typical for adult ductal organoids were observed. The authors were able to expand the islet-depleted pancreatic tissue, derived from a single cadaver donor, up to 250 × 10⁶ pancreatic organoids with the prospect of future differentiation into insulin-secreting cells.⁵⁰

The total number of organoid cells reached approximately 250 × 10⁹ cells.⁵⁰ So far, the best published human differentiation rate achieved in vitro by clinically amenable means reached insulin positivity in 5% of organoid cells.⁵² If these two approaches were combined, one could hope for 13 × 10⁹ of cells, which however would not yet be functional (steady secretion of 1/10³ of the amount of insulin secreted by isolated islets).⁵² Drawing from the experience with clinical islet transplantation, the minimum number of beta cells to achieve insulin independence is estimated to be 10⁹ of functional beta cells per recipient, given the estimated 1140 beta cells per islet equivalent.¹⁶⁶ Some authors demonstrated up to 60% differentiation rate into insulin positive cells, in the case of mouse organoids.⁴⁹ While the means were unsuitable for clinical practice, these results suggest that a highly efficient though partial transdifferentiation is possible.

The safety of pancreatic organoids derived from adult stem cells for clinical application has not yet been properly evaluated. Only a few pilot studies performed transplantation of expanded or differentiated organoids into the mice without detection of any malignant transformations after maximal observation time for 3 months.^48,53

7. Perspective

The existence of adult pancreatic progenitors or facultative progenitors has been contested for decades. Historically, three major assays have been devised to prove the existence of a stem cell in a given tissue: lineage tracing in vivo, clonogenic growth in vitro, and cellular transplantation.¹²⁴ Here, we reviewed the attempts to resolve the issue by the means of pancreatic organoid culture, which was applied to all four histological compartments of pancreas, comprising the acinar, and centro-acinar/terminal ductal, ductal, and islet cells.

The acinar and centro-acinar/terminal ductal cells failed to expand in long-term organoid culture, suggesting the absence of the stem cells. Although the ductal cells expanded in the long term and differentiated into insulin-producing cells, suggesting the presence of stem/progenitor cells, the differentiated cells remained unresponsive to glucose. Unlike the insulin-secreting cells derived from pluripotent stem cells,¹⁶⁷ the insulin positive cells derived from the organoid culture failed to mature in situ after transplantation under the kidney capsule. These findings need to be interpreted with caution, because under the current culture conditions even the embryonic/fetal stem-cells-derived organoids failed to completely differentiate unless co-cultured with mesenchymal cells. One explanation for failed differentiation can be the epigenetic state of the organoid cells preventing sufficient induction of key transcription factors involved in the differentiation process. The beta cell apparatus for metabolically regulated insulin release is quite complex,¹⁶⁸ and beta cells naturally mature only after weaning from fat-based to carbohydrate-based food.¹⁰⁵ This suggests that current adult non-islet derived organoid cultures might have failed to provide the adequate developmental and maturation cues.

Only Procr⁺ cells isolated from adult mouse islets constituted organoid culture capable of the long-term expansion as well as the differentiation into glucose-responsive insulin-producing cells, which were uniquely able to ameliorate blood glucose in diabetic mice after transplantation.¹⁶¹ Repeated supplementation of fresh endothelial cells to the organoid co-culture was necessary. The essential role of endothelial cells not only for embryogenesis, but also for proper beta cell function in adulthood was previously described,^106,169,170 with the dependence of beta cell on signaling cues from basal membrane, where beta cell cannot synthesize.

Clinical application of adult pancreatic progenitors for stem cell therapy of diabetes is challenged not only by doubtful existence of the adult pancreatic bipotential progenitors, but also by the inaccessibility of the proper pancreatic niche. The unparalleled success of clinical transplantation of hematopoietic stem cells was facilitated by the accessibility of the appropriate niche to the natural adult stem cells. Experimentally, hyperglycemia was indeed ameliorated in mice by transdifferentiation of pancreatic acinar cells in situ, taking advantage of the natural pancreatic niche.¹⁵⁹ While current evidence supports other sources of beta cell regeneration,¹³¹ the existence of adult bipotential pancreatic progenitors has not yet been abandoned.^129,130 The co-expression of Pdx1, Sox9, Nkx6.1, and Hnf6 (Onecut-1) is characteristic for the bipotent endocrine/duct progenitors during embryogenesis, and the same transcription factors are also detected within the pancreatic organoids.^46,47,52

Although a large-scale expansion of human adult pancreatic organoids was recently developed, partially at GMP level for putative clinical application, more research would be necessary to develop truly efficient and safe approach. Another drawback of the adult pancreas as the source for cell-based therapy seems to be rather complicated access to this potential cell source, the inherent variability among different donors, and technical difficulties associated with the up-scaling of organoid culture.

Nevertheless, pancreatic organoid culture represents a potentially helpful tool bringing otherwise inaccessible and complex organ in a dish. Hypothetically, a pancreatic biopsy could be used for this purpose, similar to the intestinal organoids from cystic fibrosis patients.¹⁶ Diabetes-associated genetic variants could also be introduced into organoids via CRISPR/Cas9 approach. At present, however, the identification of prospective pancreatic progenitor appears to be the most attractive goal for the use of pancreatic organoid system.

At the moment, adult pancreatic organoids are far from becoming a cell source for clinical application, but they remain an invaluable research tool for the investigation of beta cell development and regeneration, which makes more refinement effort worthwhile. It is intriguing whether the recently discovered Procr⁺ islet progenitor cells in the adult mice exist also in human islets, constituting a novel source of potentially curative beta cell replacement therapy or in situ regeneration. The lessons learned from the differentiation approach so successful in Procr⁺ islet progenitors might be applicable for the putative extra-islet pancreatic progenitors, thus moving the ten years old field forward.

Abbreviations

ALDH	Aldehyde dehydrogenase
ALK5	Activin receptor-like kinase 5
BME	Basement membrane extract
CD133	Prominin 1
ECAD	E-cadherin
ECM	Extracellular matrix
EGF	Epidermal growth factor
ENR	EGF+Noggin+R-Spondin-1
EpCAM	Epithelial cell adhesion molecule
ESCs	Embryonic stem cells
EZH2	Enhancer of zeste homolog 2
FACS	Flourescent activated cell sorting
FGF1	Fibroblast growth factor 1
FGF10	Fibroblast growth factor 10
GMP	Good manufacturing practices
HGF	Hepatocyte growth factor
HNF1B	Hepatocyte nuclear factor-1 beta
HNF6	Hepatocyte nuclear factor 6, Onecut-1
CHGA	Chromogranin A
CHIR99021	Inhibitor of glycogen synthase kinase 3ß
IGF-1	Insulin-like growth factor 1
IPGTT	Intraperitoneal glucose tolerance test
iPSC	Induced pluripotent stem cells
ISX-9	Izoxazole 9
ITS	Insulin-Transferrin-Selenium
IWP-2	Inhibitor of WNT Production-2
Ki67	Proliferation marker
KRT19	Cytokeratine 19
LGR5	Leucine-rich-repeat-containing G-protein-coupled receptor 5
MACS	Magnetic-activated cell sorting
MAFA	C-maf musculoaponeurotic fibrosarcoma oncogene homolog A
MUC1	Mucin 1
NGN3	Neurogenin3
NKX6.1	Nirenberg and Kim homeobox 6.1
PAX6	Paired box gene 6
PDX1	Pancreatic and duodenal homeobox 1
PEG	Poly(ethylene glycol)
PP2	4-amino-5-(4-chlorophenyl)-7-(t-butyl) pyrazolo[3,4-d]pyrimidine
PrEBM	Prostate epithelial cell growth basal medium
Procr	Protein C receptor
PTF1A	Pancreas associated transcription factor 1A
RepSox	Inhibitor of TGFβ type I activin like kinase receptor (ALK5)
RFD	Tripeptide Arg-Gly-Asp
R-Spondin-1	Roof plate-specific Spondin-1
SC	Stem cell
scRNA-seq	Single-cell RNA sequencing
SOX9	SRY-Box Transcription Factor 9
TGFβ	Transforming growth factor β superfamily
TSQ	Fluorescent chelator for Zn2+ ionts (6-methoxy-8-p-toluenesulfonamidoquilone)
VEGF	Vascular endothelial growth factor (A)
Wnt	Wingless-INT-#x003B2;-catenin signaling pathway
WNT3A	Wnt family member 3A
Y-27632	ROCK inhibitor

Funding Statement

This work was supported by the Grantová Agentura České Republiky [19-07661S].

Disclosure statement

No potential conflict of interest was reported by the authors.