Method for producing cells having an acinar cell phenotype

Abstract

The present invention relates to a method for producing cells having an acinar cell phenotype comprising (a) exposing pancreatic progenitor cells derived from human pluripotent stem cells to Wnt pathway activator in the absence of fibroblast growth factor (FGF), or, if the pancreatic progenitor cells are cultured in a medium comprising FGF, in the presence of an FGF inhibitor, thereby differentiating said pancreatic progenitor cells into cells having an acinar cell phenotype.

Description

[0001] New PCT-Patent Application

[0002] Max-Planck-Gesellschaft zur Fbrderung der Wissenschaften e.V.

[0003] Vossius Ref.: AJ3280 PCT

[0004] Method for producing cells having an acinar cell phenotype

[0005] The present invention relates to a method for producing cells having an acinar cell phenotype comprising (a) exposing pancreatic progenitor cells derived from human pluripotent stem cells to Wnt pathway activator in the absence of fibroblast growth factor (FGF), or, if the pancreatic progenitor cells are cultured in a medium comprising FGF, in the presence of an FGF inhibitor, thereby differentiating said pancreatic progenitor cells into cells having an acinar cell phenotype.

[0006] In this specification, a number of documents including patent applications and manufacturer's manuals are cited. The disclosure of these documents, while not considered relevant for the patentability of this invention, is herewith incorporated by reference in its entirety. More specifically, all referenced documents are incorporated by reference to the same extent as if each individual document was specifically and individually indicated to be incorporated by reference.

[0007] Pancreatic ductal adenocarcinoma (PDAC) is a devastating type of cancer. Despite the usual ductal appearance of PDACs, most evidence from genetically engineered mice (GEMs) in the last 10-15 years has pointed to an acinar cell origin of a majority of these tumours in which acinar cells undergo ductal metaplasia at the onset of tumor formation.

[0008] Fully differentiated acinar cells are characterized by the expression of the transcription factors PTF1A, MISTI, GATA4, and NR5A2, which regulate, among others, the expression of digestive enzymes such as carboxypeptidases, trypsins, amylases, elastases, RNAses, and lipases. It was reported early on that mononucleated and binucleated acinar cells could be observed at the histological level, but it was not until single-cell RNA sequencing was introduced that acinar cell heterogeneity could be better characterized. Interestingly, in human studies, a subset of acinar cells (acinar-i cells) shows lower activation of acinar-specific regulatory networks, suggesting that they might have a higher capacity to convert into other cell types of the pancreas. A subpopulation of proliferative Stmn+ acinar cells with similar characteristics was also reported in mice hinting to certain similarities between the 2 species. Moreover, a recent study identified a subpopulation of human acinar cells (acinar-edge cells) with features of progenitor cells or dedifferentiation. New experimental approaches to study acinar cell heterogeneity will continue to shed light on acinar cell heterogeneity (Backx et al. Cell Mol Gastroenterol Hepatol 2022, volume 13, pages 1243-1253.).

[0009] The pancreas mainly consists of exocrine tissue, which is composed of acinar cells that produce digestive enzymes, and ductal cells that form the afferent system to the duodenum. In major pancreatic pathologies, namely pancreatitis and PDAC, acinar cells become replaced by ductal cells, a process called acinoductal metaplasia. Two different mechanisms may contribute to this metaplasia:

[0010] (1) elimination of acinar cells by apoptosis in combination with a selective expansion of ductal cells or

[0011] (2) trans-differentiation of acinar cells to ductal cells (Houbracken et al. Gastroenterology 2011, volume 141, pages 731-741), with more evidence for the latter. Hence, in vitro produced acinar or acinar-like cells also hold promise for the treatment of pancreatic pathologies, in particular PDAC.

[0012] In order to facilitate experimental approaches to better understand PDAC and for the discussed medial applications in pancreatic pathologies there is a need for a method for producing cells having an acinar cell phenotype, so that enough and suitable cells for experimental approaches and potential therapeutic interventions. This need is addressed herein.

[0013] Accordingly, the present invention relates in a first aspect to a method for producing cells having an acinar cell phenotype comprising (a) exposing pancreatic progenitor cells derived from human pluripotent stem cells to Wnt pathway activator in the absence of fibroblast growth factor (FGF), or, if the pancreatic progenitor cells are cultured in a medium comprising FGF, in the presence of an FGF inhibitor, thereby differentiating said pancreatic progenitor cells into cells having an acinar cell phenotype.

[0014] The pancreas has two distinct functional portions: the exocrine and the endocrine pancreas. The endocrine pancreas, consisting of pancreatic islet cells that produce insulin, glucagon, somatostatin, and pancreatic polypeptide serves to maintain the body's glucose homeostasis. The exocrine pancreas is structurally analogous to a bunch of grapes; this architecture contains microscopic, blind-ended tubules that are surrounded by polygonal acinar cells and these tubules are organized into lobules called acini. The primary function of the acini is to synthesize and secrete hydrolytic enzymes which empty into the duodenum for the digestion of our daily foodstuff. Each acinus consists of a small cluster of secretory epithelial cells that form a small central lumen, called intercellular canaliculi. Each acinar cell has a round pyramid-like shape. The acinar cell is highly polarized with two plasma membrane domains. The basolateral membrane is large and located at the acinar periphery; the apical membrane faces the acinar lumen that connects with a tiny intercalated duct. The digestive enzymes are stored in secretory granules which are concentrated near the apical membrane of the cell. The acini drain into the intercalated ducts, and multiple intercalated ducts drains into larger intralobular ducts, which in turn drain into much larger extralobular ducts; the latter form a main collecting duct which empties into the duodenum (Leung and Ip, The International Journal of Biochemistry & Cell Biology, 2006, volume 38, pages 1024-1030).

[0015] It is believed that the cells having an acinar cell phenotype as obtained or as being obtainable by the method of the invention are not identical to wild-type acinar cells. However, it is demonstrated in the examples herein below that cells having an acinar cell phenotype as obtained or as being obtainable by the method of the invention closely resemble wild-type acinar cells. Phenotypic properties that are shared by the cells having an acinar cell phenotype as obtained or as being obtainable by the method of the invention and wild-type acinar cells will be described herein below. The acinar cell phenotype is preferably a pancreatic acinar cell phenotype.

[0016] In accordance with step (a) of the method of the invention the source for obtaining cells having an acinar cell phenotype are pancreatic progenitor cells derived from human pluripotent stem cells.

[0017] Human pluripotent stem cells can be derived from human embryonic elements generally without the destruction of the human embryo. Human pluripotent stem cells can be derived either from early human embryos (4 - 7 days post-conception) or adult tissues. For example, induced human pluripotent stem cells can be used. Induced pluripotent stem cells (also known as iPS cells or iPSCs) are a type of pluripotent stem cell that can be generated directly from a somatic cell. The iPSC technology was pioneered by Shinya Yamanaka and Kazutoshi Takahashi in Kyoto, Japan, who together showed in 2006 that the introduction of four specific genes (named Myc, Oct3 / 4, Sox2 and Klf4), collectively known as Yamanaka factors, encoding transcription factors could convert somatic cells into pluripotent stem cells. The human pluripotent stem cells are preferably cells of the Hl, H7 or H9 human embryonic stem cell (hESC) lines, wherein the Hl and H9 hESC lines are preferred. Human induced PSC (hiPSC) cells are preferably from cells of the lines WTC11, CRTD1 and CRTD11A.

[0018] Means and methods for obtaining pancreatic progenitor cells from human pluripotent stem cells are know in the art and described in the appended examples. As examples of prior art methods reference is made to Rezania et al. Nature Biotechnology, 2014, volume 32, pages 1121-1133, Balboa et al. Nature Biotechnology, 2022, volume 40, pagesl042-1055, Chen et al., Nature Communications, 2024, volume 15, page6344 and Jiang et al., Stem Cell Reports, 2012, volume 16, pages 2395-2409. A preferred method is the method as used in the appended examples. Pancreatic progenitors can be amplified / expanded in culture, notably using media compositions containing FGF2. Large numbers of pancreatic progenitors can in this way be obtained, frozen, thawed and grown again while retaining their differentiation potential.

[0019] Pancreatic progenitor cells are multipotent cells which have the ability to differentiate into the lineage specific progenitors responsible for the developing pancreas. They can give rise to both, the endocrine and exocrine cells and this also to acinar cells or cells having an acinar cell phenotype (Larsen et al. Nature Communications, 2017 volume 8, Article number: 605 and Ma et al. Nature Communications, 2023, volume 14, Article number: 5354).

[0020] In accordance with step (a) the pancreatic progenitor cells are differentiated into cells having an acinar cell phenotype by (1) a Wnt pathway activator in the absence of fibroblast growth factor (FGF), or 2) a Wnt pathway activator in the presence of an FGF inhibitor, if the pancreatic progenitor cells are cultured in a medium comprising FGF.

[0021] In this connection it is of note that pancreatic progenitor cells are often cultured in a medium that comprises FGF, in particular FGF2, FGF7 or FGF10, preferably FGF2. This is because FGF, in particular FGF2, FGF7 and FGF10 promote pancreatic progenitor specification, maintenance and expansion (Ameri et al., Stem Cells 2010, volume28, pages 45-56, Rezania et al., Nature Biotechnology, 2014, volume 32, pages 1121-1133, Balboa et al., Nature Biotechnology, 2022, volume 40, pages 1042-1055, Nostro et al., Development, 2011, volume 138, pages 861-871 and the appended examples). However, FGFs can also inhibit the further differentiation of progenitors into acinar and endocrine cells (Hart et al., Dev Dyn 2023, volume 228, pages 185-93, Norgaard et al., Dev Biol 2003, volume264, pages :323- 3 , Ye at al., Diabetologia 2005, volume 48, pages 277-281, Kobberup et al., Meeh Dev 2010, volumel27 pages 220-34). Accordingly, option (2) is preferred, wherein (2) the pancreatic progenitor cells are cultured in a medium comprising FGF and Wnt pathway activator and in the presence of an FGF inhibitor. Instead of using an FGF inhibitor the FGF can also be removed, for example, by a medium exchange.

[0022] The Wnt pathway activator and the FGF inhibitor can but do not have to be added in step (a) at the same time. It is only required that both of them are present together for a time that is sufficient in order to differentiate the pancreatic progenitor cells into cells having an acinar cell phenotype. Is preferred that Wnt pathway activator and the FGF inhibitor are added at the same time or that the Wnt pathway activator is added first and then the FGF inhibitor.

[0023] A Wnt pathway activator is a compound that leads to Wnt signaling activation in cells. The nature of the Wnt pathway activator is not particularly limited and examples include Wnt mimics, antibodies targeting Wnt inhibitors, glycogen-synthase-3P inhibitors and indirubins; see for review Bonet et al., RSC Chem Biol. 2021, volume 2, pages 1144-1157. Preferred examples of Wnt pathway activators will described herein below. Wnt signaling represents one of the multiple conserved pathways, including Notch, Hedgehog, transforming growth factor p (TGF-P) / bone morphogenetic protein (BMP)3 and Hippo, essential for embryonic development, the maintenance of stem cell (SC) proliferation, SC selfrenewal, and tissue regeneration.

[0024] FGF (fibroblast growth factor) are a family of cell signalling proteins produced by multiple cell type; they are involved in a wide variety of processes, most notably as crucial elements for normal development in animal cells. In humans, 23 members of the FGF family have been identified, all of which are structurally related signaling molecules. Members FGF1 through FGF10 all bind fibroblast growth factor receptors (FGFRs). FGF1 is also known as acidic FGF, and FGF2 is also known as basic FGF. Hence, the FGF is preferably one or more of FGF1 to FGF10, and most preferably FGF2. Likewise, the FGF inhibitor is preferably an inhibitor of one or more of FGF1 to FGF10, and most preferably of FGF2.

[0025] It is to be understood that inhibiting FGF activity of the inhibitor might bind / inhibit an FGFR (fibroblast growth factor receptor) or FGF. The FGFRs are, as their name implies, receptors that bind to members of the fibroblast growth factor (FGF) family of proteins (FGFR1, FGFR2, FGFR3, FGFR4, and FGFR6). There are non-selective FGFR inhibitors that act on all of FGFR1-4 and other proteins, and some selective FGFR inhibitors for some / all of FGFR1-4. Selective FGFR inhibitors include, for example, AZD4547, BGJ398, JNJ42756493, and PD173074.

[0026] A plethora of selective FGFR / FGF pathway inhibitors are available in the art; see, for example, for review Repetto et al., Expert Rev Clin Pharmacol, 2021, volume 14, pages 1233-1252. Preferred examples of FGF inhibitors will be described herein below.

[0027] As can be taken form the appended examples by focusing on one family of compounds it was found that GSK3A / B inhibition via Wnt signaling has a global reversible effect on cell identity by repressing a series of pancreatic progenitor markers. This induces a poised state of progenitors transitioning to acinar cells and hinted additional signaling regulation that promotes further acinar differentiation. It was then further found that combining Wnt activation and FGF repression enables efficient generation of acinar cells organized in rosettes recapitulating pancreatic acini. The combined use of Wnt activation and FGF repression is advantageously a very simple two components only method for obtaining acinar- like cells. The culture conditions as provided herein are simpler than prior art methods and induce cells having the phenotype of mature acinar cells as evidenced in the example based on acinar marker induction, variety of markers observed, abundance of zymogen granules, and cellular organization into acini. While EP 3061808 A2 relates to a method for culturing epithelial stem cells, isolated tissue fragments comprising said epithelial stem cells, or adenoma cells, acinar cells are not mentioned therein. While in the method of the invention pancreatic progenitor cells are differentiated into cells having an acinar cell phenotype by exposing them to a Wnt pathway activator in the absence of FGF or in the presence of an FGF inhibitor, in EP 3061808 A2 pancreas organoids are made from adult stem cells within a medium comprising FGFs. EP 3061808 A2 therefore differs in several aspects and fundamental aspects from the method of the present invention. The ability to produce in vitro cells having the phenotype of acinar cells is valuable for future studies on pancreatic exocrine function and cancer initiation in human, as acinar cells are thought to be an important cell of origin for pancreatic adenocarcinoma (Backx et al., Cell Mol Gastroenterol Hepatol 2022, volume 13, pages 1243-1253, Houbracken et al. Gastroenterology 2011, volume 141, pages 731-741). In addition, and as discussed above, cells having an acinar cell phenotype hold promise for the treatment of pancreatic pathologies, in particular pancreatitis and PDAC.

[0028] In accordance with a preferred embodiment of the first aspect the Wnt pathway activator is a canonical Wnt pathway activator, more preferably an agonist of Frizzled or LRP5 and / or LRP6, even more preferably a Wnt activating the canonical Wnt pathway acting via GSK3, Norrin, the peptide p-catenin pathway agonist PG-008 or a GSK-inhibitor and is most preferably a GSK-inhibitor.

[0029] The Wnt signalling pathways include noncanonical and canonical pathways. The noncanonical Wnt pathways are independent of p-catenin-T-cell factor / lymphoid enhancer-binding factor (TCF / LEF), such as the Wnt / Ca2+pathway and noncanonical Wnt planar cell polarity. The canonical Wnt pathway, also known as the Wnt / -catenin pathway, involves the nuclear translocation of p-catenin and activation of target genes via TCF / LEF transcription factors. The canonical Wnt pathway mainly controls cell proliferation, whereas the noncanonical Wnt pathways regulate cell polarity and migration, and the two main pathways form a network of mutual regulation (Rim et al., Annu Rev Biochemm 2022, volume 91, pages 571-598.). Frizzled is a family of atypical G protein-coupled receptors that serve as receptors in the Wnt signaling pathway and other signaling pathways. When activated, Frizzled leads to activation of Dishevelled in the cytosol. LRP5 and LRP6 are single-pass transmembrane proteins with multiple domains and transmit Wnt signaling.

[0030] Agonist of Frizzled or LRP5 and / or LRP6. for high-efficiency Wnt / p-catenin signaling manipulation are know in the art and described, for example, in Dai et al, BioRxiv 2023, doi 10.1101 / 2023.06.21.545860 and Chuan et al., Pharmacological Research, 2024, volume 206 article 107286. Non-limiting but preferred examples are Frizzled6 agonist SAG1.3, FzM1.8, purmorphamine, RRP-pbFn and bivalent or tetravalent antibodies which activate the WNT signaling via inducing FZD-LRP5 / 6 heterodimerization

[0031] Wnt comprises a diverse family of secreted lipid-modified signaling glycoproteins that are 350-400 amino acids in length. The Wnt pathway is triggered by these secreted glycoproteins. In humans WNT1, WNT2, WNT2B, WNT3, WNT3A, WNT4, WNT5A, WNT5B, WNT6, WNT7A, WNT7B, WNT8A, WNT8B, WNT9A, WNT9B, WNT10A, WNT10B, WNT11, WNT16 are known. WNT3a is the most commonly used WNT and is also used in the appended examples. WNT3A is therefore preferred. R Spondin 1 can be and is preferably used in addition as a potentiator of WNT, in particular WNT3a.

[0032] Norrin also known as Norrie disease protein or X-linked exudative vitreoretinopathy 2 protein (EVR2) is a protein that in humans is encoded by the NDP gene. Norrin specifically binds to Frizzled-4 receptors and activates the canonical Wnt / p-catenin signaling pathway.

[0033] The peptide -catenin pathway agonist PG-008 is composed of a heterodimer of two cyclic peptides with selective binding affinity for Frizzled and LRP5 / 6. Similar to Wnt3a, it activates the p-catenin pathway of the Wnt signaling.

[0034] Glycogen synthase kinase 3 (GSK-3) is a multifunctional serine / threonine kinase consisting of two isoforms, alpha and beta. It is a highly conserved negative regulator of receptor tyrosine kinase, cytokine, and Wnt signaling pathways. Stimulation of these pathways inhibits GSK-3 to modulate diverse downstream effectors that include transcription factors, nutrient sensors, glycogen synthesis, mitochondrial function, circadian rhythm, and cell fate. GSK-3 also regulates alternative splicing in response to T-cell receptor activation, and recent phosphoproteomic studies have revealed that multiple splicing factors and regulators of RNA biosynthesis are phosphorylated in a GSK-3-dependent manner. GSK-inhibitors, in particular GSK-3P inhibitors have become a widely used chemical biology tool to study the canonical Wnt signaling pathway. A plethora of GSK-3 inhibitors are available and the MedChemExpress database currently lists 164 GSK-3 Related Products and thereof 81 GSK-3P inhibitors. Preferred examples of GSK-3 inhibitors are shown in Figure 11.

[0035] In accordance with a more preferred embodiment of the first aspect wherein the GSK3-inhibitor is CHIR99021 (6-[[2-[[4-(2,4-Dichlorophenyl)-5-(5-methyl-lH-imidazol-2-yl)-2-pyrimidinyl]amino] ethyl]amino]-3-pyridinecarbonitrile) (CAS No.: 252917-06-9), TWS119 (CAS No.: 601514-19-6), is preferably TWS119 and CHIR99021, and is most preferably CHIR99021.

[0036] It is shown in the appended examples that both, TWS119 and CH IR99021 worked well. While also other GSK3-inhibitors were tested in particular TWS119 and CHIR99021 resulted in the reduction of the progenitor marker PDX1. The use of TWS119 and CHIR99021 is therefore particularly advantageous for obtaining cells that closely resemble wild-type acinar cells.

[0037] CHIR99021 is most preferred since it is used for most experiments as the Wnt pathway activator in the appended examples.

[0038] In accordance with an even more preferred embodiment of the first aspect CHIR99021 is present in step (a) at a concentration of about 0.6 pM to about 7.5 pM, preferably about 1 pM to about 5 pM, more preferably about 2 pM to about 4 pM and most preferably about 3 pM.

[0039] CHIR99021 is used in the appended examples at a concentration of 3 pM and the ranges according to the above more preferred embodiment are also suitable concentration ranges for CHIR99021.

[0040] The term "about" has used herein means for each occurrence independently with increasing preference ±20%, ±10% and ±5%.

[0041] In accordance with a preferred embodiment of the first aspect the FGF inhibitor is an FGF receptor inhibitor.

[0042] The FGF receptor inhibitor can be a non-selective FGFR inhibitors that act on all of FGFR1-4 and other proteins, and some selective FGFR inhibitors for some / all of FGFR1-4. Selective FGFR inhibitors for some / all of FGFR1-4 are preferred. Preferred examples of FGF receptor inhibitors are shown in Figure

[0043] 12. In accordance with a more preferred embodiment of the first aspect the FGF receptor inhibitor is SU5402 (CAS No.: 215543-92-3) or Infigratinib (CAS No. 872511-34-7), wherein SU5402 is preferably at a concentration in step (a) of about 5 pM to about 15 pM, more preferably at about 8 pM to about 12 pM and most preferably at about 10 pM, or wherein Infigratinib is preferably at a concentration in step (a) of about 0.5 pM to about 100 nM.

[0044] SU5402 is most preferred since it is used in the appended examples.

[0045] SU5402 is used in the appended examples at a concentration of 10 pM and the ranges according to the above more preferred embodiment are also suitable cconcentration ranges for SU5402 (CAS No.: and Infigratinib, respectively.

[0046] In accordance with a preferred embodiment of the first aspect step (a) is carried out for at least 5 days, preferably at least 6 days and most preferably at least 7 days, and / or up to 30 days, preferably up to 20 days and most preferably up to 10 days.

[0047] The preferred upper and lower time limits for carrying out step (a) can be combined into ranges and these ranges are also described herein. Increasingly preferred examples of such ranges are 5 to 30 days, 6 to 20 days and 7 to 10 days. For the 2D conditions as described herein step (a) can also be carried out for at least 2 days and the time ranges (with increased preference of) 2 to 30 days, 2 to 20 days and 2 to 10 days.

[0048] In accordance with another preferred embodiment of the first aspect in step (a) the pancreatic progenitor cells are in or on an extracellular matrix (3D conditions).

[0049] In accordance with yet other preferred embodiment of the first aspect in step (a) the pancreatic progenitor cells are kept under 2D conditions, preferably on a fibronectin coated surface (e.g. labware such as culture dish or bottle).

[0050] While it is shown in the appended examples that the extracellular matrix is not essential for the method of the invention the extracellular matrix is deemed advantageous because the matrix supports cellular growth and the formation of cell layers and organoids. The matrix also positively affects a variety of cellular mechanisms, such as adhesion, migration, proliferation, and differentiation. In accordance with a more preferred embodiment of the first aspect the extracellular matrix (i) comprises one or more and preferably all of collagen (preferably collagen type IV), entactin, perlecan (preferably heparan sulfate proteoglycan), fibronectin and laminin; and / or (ii) is a reconstituted basement membrane derived from extracts of mammalian cells, preferably mouse cells, and most preferably Engelbreth-Holm-Swarm mouse tumor cells.

[0051] An extracellular that comprises one or more and preferably all of collagen (preferably collagen type IV), entactin, perlecan (preferably heparan sulfate proteoglycan), fibronectin and laminin is known as the product Matrigel. Matrigel a reconstituted basement membrane derived from extracts of Engelbreth-Holm-Swarm mouse tumor cells.

[0052] A common laboratory procedure is to dispense small volumes of chilled (4°C) liquid Matrigel onto plastic tissue culture labware. When incubated at 37°C (body temperature) the Matrigel proteins polymerize (solidify) producing a basement membrane that covers the labware's surface.

[0053] In accordance with a further more preferred embodiment of the first aspect the method further comprises before step (a) (a') adding the pancreatic progenitor cells, preferably in the form of pancreatic progenitor organoids into the extracellular matrix.

[0054] According to this preferred embodiment the method of the first aspect comprises the additional step (a'). According to this step the pancreatic progenitor cells, preferably in the form of pancreatic progenitor organoids into the extracellular matrix. The organoids may form in the extracellular matrix by culturing and propagating pancreatic progenitor cells in the extracellular matrix.

[0055] An organoid is a miniaturised and simplified version of an organ produced in vitro in three dimensions that mimics the key functional, structural, and biological complexity of that organ. The organ is in the present case the pancreas.

[0056] In accordance with a preferred embodiment of the first aspect the method further comprises after step (a) (b) isolating the cells having an acinar cell phenotype, wherein the cells having an acinar cell phenotype are preferably in the form of organoids.

[0057] According to this preferred embodiment the method of the first aspect comprises the additional step (b). As discussed above while cells having an acinar cell phenotype can ale be produced in 2D cultures, in particular for the formation of organoids the extracellular matrix is of help. The isolated cells (2D) or organoids (3D) may be stored until further use, for example, by freezing them. Cryopreservation of cells and organoids is art established and freezing media are commercially available or can be prepared by standard methods.

[0058] In accordance with another preferred embodiment of the first aspect the method further comprises before step (a) and if present after step (a') (a") culturing the pancreatic progenitor cells in a medium comprising (a) FGF, preferably FGF2, and / or (b) a Rock inhibitor, preferably Y27632 (Cas No. 146986- 50-7) and / or (c) insulin, preferably in the form of B27 medium supplement.

[0059] A medium comprising (a) FGF, preferably FGF2, (b) a Rock inhibitor, preferably Y27632 (Cas No. 146986-50-7) and (c) insulin, preferably in the form of B27 medium supplement is used in the appended examples to culture the pancreatic progenitor cells before their differentiation into cells having an acinar cell phenotype.

[0060] This medium in particular FGF, preferably FGF2 ensures that the pancreatic progenitor cells do not further differentiate and retain their capacity to form cells having an acinar cell phenotype in step (a) of the method of the invention.

[0061] In accordance with a further preferred embodiment of the first aspect the method further comprises before step (a) and if present also before steps (a') and (a") (al) differentiating human pluripotent stem cells into pancreatic progenitor cells, and (a2) optionally expanding and / or maintaining the pancreatic progenitor cells, wherein pancreatic progenitor cells are preferably in the form of pancreatic progenitor organoids.

[0062] As mentioned, above means and methods for differentiating human pluripotent stem cells into pancreatic progenitor cells are known in the art.

[0063] Step (al) is preferably carried out as described in the section "differentiation of human pluripotent stem cells into pancreatic progenitor cells" in the appended examples, wherein the protocol for differentiation as originally published by Rezania et al (Rezania et al., Nature Biotechnology, 2014, volume 32, pages 1121-1133), modified and adapted by us (Goncalves et al„ Nature Communications 2021, volume 12, article 3144) was used. Accordingly, it preferred that the following differentiation media and stages are used in step (al): Human pluripotent stem cells are preferably expanded and maintained in mTeSRl (Stem Cell Technologies) medium. The mTeSRl medium is a complete, serum-free, defined formulation that contains recombinant human basic fibroblast growth factor (rh bFGF) and recombinant human transforming growth factor p (rh TGF ).

[0064] Differentiation:

[0065] Stage 1: 1.5 g / L sodium bicarbonate, lx Glutamax, 10 mM Glucose and 0.5% fatty acid free bovine serum albumin in MCDB 131 basal medium. The stage medium was supplemented with 3 pM CHIR and 100 ng / ml Activin A for day 1, 0.3 pM CHIR and 100 ng / ml Activin A for day 2 and 100 ng / ml Activin A for day 3.

[0066] Stage 2: 1.5 g / L sodium bicarbonate, lx Glutamax, 10 mM Glucose, 0.5% fatty acid free bovine serum albumin, 0.25 mM ascorbic acid and 50 ng / ml FGF 7 in MCDB 131 basal medium for 2 days.

[0067] Stage 3: 2.5 g / L sodium bicarbonate, lx Glutamax, 10 mM Glucose, 2% fatty acid free bovine serum albumin, 0.25 mM ascorbic acid, 1:200 ITS-X, 1 pM retinoic acid, 0.25 pM Sant-1, 100 nM LDN193189, 200 nM TPB and 50 ng / ml FGF 7 in MCDB 131 basal medium for 2 days.

[0068] Stage 4: 2.5 g / L sodium bicarbonate, lx Glutamax, 10 mM Glucose, 2% fatty acid free bovine serum albumin, 0.25 mM ascorbic acid, 1:200 ITS-X, 0.1 pM retinoic acid, 0.25 pM Sant-1, 200 nM LDN193189, 100 nM TPB and 2 ng / ml FGF 7 in MCDB 131 basal medium for 3 days.

[0069] The human pluripotent stem cells are preferably differentiated into pancreatic progenitor cells in or on an extracellular matrix as described herein above.

[0070] Step (a2) is preferably carried out as described in the sections "Expansion and maintenance of hPSC- derived pancreatic progenitor organoids" and "Expansion and maintenance ofhPSC-derived pancreatic progenitors in 2D". Accordingly, the pancreatic progenitor cells are preferably expanded and maintained in a DMEM / F12 GlutaMAX medium (Thermo Fisher Scientific) containing FGF2 (Peprotech), lx B-27 (Thermo Fisher Scientific) and the TGF-beta / Smad inhibitor SB431542 (Santa Cruz Biotechnology-only in 2D) supplemented with ROCK inhibitor Y-27632 (only in the day after plating).

[0071] Also the pancreatic progenitor cells are preferably expanded and maintained in or on an extracellular matrix as described herein above.

[0072] It is to be understood that after step (al) or (a2) the pancreatic progenitor cells can but do not need to be directly processed to step (a), (a') or (a"). The pancreatic progenitor cells may be stored until further use, for example, by freezing them.

[0073] In accordance with a preferred embodiment of the first aspect the cells having an acinar cell phenotype are characterized by the expression of AMY, CPA2, PRSS1, MISTI and SPINK1 and / or the reduced expression of progenitor markers, preferably the progenitor markers PDX1 and / or NKX6-1 and / or the reduced expression of the ductal marker SOX9 and the and endocrine progenitor marker CHGA.

[0074] In accordance with a related preferred embodiment of the first aspect the cells having an acinar cell phenotype are characterized by the expression of AMY, CPA2, PRSSl / Trypsin, MISTI, CEL, GATA6, UEA1, REGIA and SPINK1 and / or the reduced expression of progenitor markers, preferably the progenitor markers PDX1 and / or NKX6-1 and / or the reduced expression of the ductal markers SOX9, CFTR, KRT19 and endocrine progenitor marker CHGA.

[0075] The marker genes AMY, CPA2, PRSS1 / Trypsin MISTI and SPINK1 and optionally in addition MISTI, CEL, GATA6, UEA1 and REGIA are marker genes of acinar cells. Hence, their expression shows that the cells as obtained or as being obtainable by the method of the invention indeed have an acinar cell phenotype.

[0076] The progenitor markers genes, in particular PDX1 and / or NKX6-1 characterize pancreatic progenitor cells. Hence, a reduced expression of progenitor markers, preferably the progenitor markers PDX1 and / or NKX6-1 also shows that the cells as obtained or as being obtainable by the method of the invention indeed have an acinar cell. The same hold true for the endocrine progenitor marker CHGA.

[0077] As explained above, pancreatic progenitor cells can differentiate into ductal cells or acinar cells. For this reason, reduced expression of the ductal marker SOX9 and optionally addition CFTR and KRT19 likewise shows that the cells as obtained or as being obtainable by the method of the invention indeed have an acinar cell phenotype.

[0078] In the context of the above preferred embodiment the "and / or" is preferably "and".

[0079] In accordance with a related preferred embodiment of the first aspect the cells having an acinar cell phenotype are characterized by zymogen granules and / or cellular organization into acini and / or digestive enzyme production, wherein the digestive enzymes are one or more of amylase, trypsin and carboxypeptidase (such as CPA2). Wild-type acinar cells are characterized by zymogen granules, cellular organization into acini, digestive enzyme production, including the digestive enzymes amylase, trypsin and carboxypeptidase. If follows that one or more of zymogen granules, cellular organization into acini, digestive enzyme production, wherein the digestive enzymes are one or more of amylase, trypsin and carboxypeptidase further indicate that the cells as obtained or as being obtainable by the method of the invention indeed have an acinar cell phenotype.

[0080] Also in the context of the above preferred embodiment the "and / or" is preferably "and".

[0081] The present invention relates in a second aspect to a cell having an acinar cell phenotype obtained by the method of the first aspect.

[0082] The definitions and preferred embodiments of the first aspect apply mutatis mutandis to the second aspect as far as being amenable for combination with the second aspect. For example, also the cells having an acinar cell phenotype or being obtainable by the method of the first aspect are preferably characterized (1) by the expression of AMY, CPA2, PRSS1, MISTI and SPINK1 and / or the reduced expression of progenitor markers, preferably the progenitor markers PDX1 and / or NKX6-1 and / or the reduced expression of the ductal marker SOX9; and / or (2) zymogen granules and / or cellular organization into acini and / or digestive enzyme production, wherein the digestive enzymes are one or more of amylase, trypsin and carboxypeptidase. Also in this context "and / or" is preferably "and".

[0083] As discussed herein above, the cells as obtained or as being obtainable by the method of the invention have an acinar cell phenotype that closely resemble but is not identical to wild-type acinar cells. It is also believed that it was not possible to obtain these cells having an acinar cell phenotype by methods other than the method of the invention.

[0084] For this reason, it is believed that the cells having an acinar cell phenotype or being obtainable by the method of the first aspect are a novel type of cells

[0085] The present invention relates in a third aspect to the cell of the second aspect for use in treating or preventing a pancreatic pathology, preferably being selected from pancreatic adenocarcinoma, acinar cell carcinoma, mucinous cystic pancreatic neoplasms, adenosquamous carcinomas, Shwachman Diamond Syndrome or acinar cell damage, preferably caused by pancreatitis. The definitions and preferred embodiments of the first and second aspect apply mutatis mutandis to the third aspect as far as being amenable for combination with the third aspect.

[0086] As discussed herein above, pancreatic pathologies can be characterized or even caused by deficient acinar cells. For example, aberrant acinar cells may cause cancer or the death (apoptosis) of acinar cells may cause pancreatic misfunction and / or inflammation. Such pancreatic pathologies are pancreatic adenocarcinoma, acinar cell carcinoma, mucinous cystic pancreatic neoplasms, adenosquamous carcinomas, Shwachman Diamond Syndrome or acinar cell damage, preferably caused by pancreatitis.

[0087] As regards the embodiments characterized in this specification, in particular in the claims, it is intended that each embodiment mentioned in a dependent claim is combined with each embodiment of each claim (independent or dependent) said dependent claim depends from. For example, in case of an independent claim 1 reciting 3 alternatives A, B and C, a dependent claim 2 reciting 3 alternatives D, E and F and a claim 3 depending from claims 1 and 2 and reciting 3 alternatives G, H and I, it is to be understood that the specification unambiguously discloses embodiments corresponding to combinations A, D, G; A, D, H; A, D, I; A, E, G; A, E, H; A, E, I; A, F, G; A, F, H; A, F, I; B, D, G; B, D, H; B, D, I; B, E, G; B, E, H; B, E, I; B, F, G; B, F, H; B, F, I; C, D, G; C, D, H; C, D, I; C, E, G; C, E, H; C, E, I; C, F, G; C, F, H; C, F, I, unless specifically mentioned otherwise.

[0088] Similarly, and also in those cases where independent and / or dependent claims do not recite alternatives, it is understood that if dependent claims refer back to a plurality of preceding claims, any combination of subject-matter covered thereby is considered to be explicitly disclosed. For example, in case of an independent claim 1, a dependent claim 2 referring back to claim 1, and a dependent claim 3 referring back to both claims 2 and 1, it follows that the combination of the subject-matter of claims 3 and 1 is clearly and unambiguously disclosed as is the combination of the subject-matter of claims 3, 2 and 1. In case a further dependent claim 4 is present which refers to any one of claims 1 to

[0089] 3, it follows that the combination of the subject-matter of claims 4 and 1, of claims 4, 2 and 1, of claims

[0090] 4, 3 and 1, as well as of claims 4, 3, 2 and 1 is clearly and unambiguously disclosed.

[0091] The figures show:

[0092] Fig. 1: Image-based morphological screening assay for PP-organoids. a, Schematic representation of the experimental protocol for the image-based screening of PP- organoids. Seeded PP-cells formed organoids that expanded in pancreatic epithelium (PE) medium, and the compounds were treated from day 3 onward. Fixed organoids on day 10 were stained with DAPI and Phall oidin . Images were acquired with an automated spinning disc microscope. Images were segmented and features of interest were extracted for quantification and hit selection. Created with BioRender.com. b-j, Four pipelines for image analysis, b, Flow-chart of well-mean analysis. Parameters for nuclear and organoid measurements were averaged per well and compared to the DMSO-treated negative control. c, An example hit from well-mean analysis shows CHIR99021 treatment decreased the mean GFP intensity in organoids. N=3, *****p<0.00001, data represented as mean ± SD. d,e, Binning analysis. Graphical representation of the binning analysis for a parameter (d). Distribution for each parameter of the DMSO control was divided into three bins of equal number of objects, and the same binning thresholds were applied for the treatment conditions, z-factors for each bin were calculated to quantify increase or decrease in number of objects. An example hit (e) by Myriocin Mycelia sterilia treatment shows enriched objects with smaller area compared to the control. N=3, **p=0.0027, data represented as mean ± SD. f,g, Network analysis by the principle of Delaunay triangulation of building non-intersecting triangles. Nuclear arrangement inside a organoid was quantified by network analysis (f). An example hit (g) by Myriocin Mycelia sterilia treatment resulted in organoids with closely arranged nuclei with lower distance (less than 35 pixels) between the neighbors. N=3. *****p<0.00001, data represented as mean ± SD. h-j, Clustering of organoids in the morphospace. PP-organoids with lumen from the primary screen are grouped into 6 clusters (h), and representative images of organoids taken from the center of each cluster (i). Nuclei (DAPI), PDX1:H2B-GFP and Actin. Scale bars 30 pm. An example hit profile of GW1929 treatment shows enriched objects in cluster 1 (j). k, Overview of hits selected from the primary screen. A total of 54 hits were identified from primary screen, and some hits were identified in multiple analyses.

[0093] Fig. 2: GSK3 inhibition reduces PDX1 expression and changes the morphology of PP-organoids. a-e, Representative images of wells treated with negative control DMSO (a) and 4 different GSK3 inhibitors (b-e) identified as hits from the primary screen. PDX1:H2B-GFP, nuclei (DAPI) and F-actin ( Phall oidin). Scale bar 100 pm. f, IC50 dosage response curve showing the % GFP+ nuclei at different concentrations of CHIR. Dotted line represents the level of the DMSO control and line denotes the IC50. g, Lumen occupancy quantified by binning analysis for CHIR at different concentrations. Bin 1 has lowest lumen occupancy, which corresponds to a smaller lumen size, whereas bin 5 has the highest lumen occupancy corresponding to a bigger lumen size, z-score = | 3 | corresponds to p = 0.012419, and the higher the z-score, the lower is the p value. Data represented as mean of all wells with a given treatment ± SD. h-i, UMAP representation of the morphology clustering of all organoids with lumen, irrespective of their treatments, from the IC50 validation experiment (h). Each dot represents an organoid. Organoids corresponding to the center of each cluster are shown with nucleus and Actin. Organoids are scaled to the same size for representation. In DMSO condition, organoids are distributed across all clusters (dots), while in 3 pM CHIR organoids are enriched in cluster 2 (dots) with a stronger phenotype, corresponding to objects with larger lumen (i). Scale bars in (H), 30 pm. In (I) Data represented as mean of all wells with a given treatment ± SD.

[0094] Fig. 3: Effect of CHIR on pancreatic progenitors and partial rescue of phenotype after withdrawal a, Schematic representation of experiment for CHIR treatments on PP-organoids. b-c, Flow cytometry analysis for the expression of PP markers PDX1 (b) and NKX6-1 (c) of the day 10 PP-organoids treated with CHIR. d, Mean fluorescence intensity of the progenitor markers from flow cytometry normalized to the DMSO control. PDX1:H2B-GFP (N=3) is native GFP signal of the reporter, while PDX1 (N=5) and NKX6- 1 (N=6) are signals from antibodies used for the analysis of multiple cell lines. Data represented as mean ± SD. ****p<0.00001, **p<0.01. e, Representative time course images of PP-organoids derived from PDX1:H2B-GFP reporter. Images of were acquired following the same organoids at days 3, 7, and 10. CHIR-7days organoid exhibits increased growth rate and decreased GFP signal, compare to DMSO control or CHIR-4days. Scale bar 50 pm. PDX1:H2B-GFP. f, Schematic representation of experimental design for CHIR treatment on PP-organoids for bulk RNA sequencing. g, Principal component analysis plot of the CHIR-treated samples from the bulk RNA-sequencing data. h, Heatmap showing centered rlog-normalized expression levels of pancreatic progenitor markers after treatment with CHIR.

[0095] Fig. 4: WNT3A and RSPO1 treatments supplemented by WNT conditioned medium reproduce the phenotypes of CHIR. a, GO term enrichment for dlO CHIR-7days and dlO DMSO conditions. Arrows indicate enriched WNT- related pathways. b, Heatmap showing centered rlog-normalized expression levels of selected canonical WNT candidate genes after CHIR treatment. c, Mean fluorescence intensity levels of PDX1:H2B-GFP, PDX1 and NKX6-1 analysed by flow cytometry. N=3, WCM: WNT conditioned medium. Data represented as mean ± SD. ***p<0.001, **p<0.01. d, Representative images of PDX1:H2B-GFP PP-organoids treated with DMSO, CHIR, WNT3A+RSPO1+WCM for 7 days and fixed on day 10. Scale bars 100 pm. PDX1:H2B-GFP, F-actin, and DAPI (nuclei).

[0096] Fig. 5: Combination of WNT activation and FGF inhibition induces acinar differentiation in PP- organoids. a, Heatmaps showing centered rlog-normalized expression levels of acinar markers from the set of DEGs in the transcriptomics data. b, Schematic representation of the experiment for acinar differentiation of PP-organoids. c, Gene expression analysis by qPCR for PP and acinar markers. d, Images of optical sections from whole-mount immunostained organoids for acinar markers. Concurrent CHIR treatment and FGF2 withdrawal from day 3 upregulate multiple acinar markers such as Amylase, Trypsin, CPA2 and UEA1. Acinar marker positive cells exhibit strong apical actin and are organized in rosette-like shape. Scale bar 50 pm. e, Images by transmission electron microscopy. CHIR+noFGF2-treated organoid contains cells with zymogen vesicles (arrow) typically observed in the acinar cells. Scale bar 10 pm.

[0097] Fig. 6: Differentiation of hPSCs into PP-Organoids (Related to Fig. 1). a, Constructs for generating the reporter line PDX1:H2B-GFP and NEUROG3:tagRFP-T. b, Schematic representation of differentiation of hPSCs into PP-organoids and maintenance of PP- organoids. c, Flow cytometry analysis of the Stage 4 day 3 cells from differentiation of PDX1:H2B-GFP ESCs, expressing PDX1:H2B-GFP and NEUROG3:tagRFP-T. Gating set for ~10,000 live cells. d, Optical section image of PDX1:H2B-GFP PP-organoids at day 10. PDX1:H2B-GFP, F-Actin, and DAPI (nuclei). Scale bar 100 pm.

[0098] Fig. 7: Hits from the screening assay (Related to Fig. 1). a, An example hit from well-mean analysis for organoid eccentricity. b, An example hit from binning analysis for organoid compactness. c, An example hit from network analysis for median of edge length. d, PP-organoids without lumen from the screen grouped into 5 clusters. e, Representative images of organoids taken from the center of each cluster. Nucleus, PDX1 and Actin. f, Profile of an example hit identified from clustering analysis. g, Images of example hits identified from the screen. Fig. 8: Dosage response validation of the hits (Related to Fig. 2). a-b, Toxicity curves showing % nuclei at different concentrations of CHIR (a) and GSK3b Inhibitor XII TWS119 (b). Number of nuclei in the DMSO condition was normalized to 100% and represented with the dotted line. c, IC50 dosage response curves showing the % GFP+ nuclei at different concentrations for GSK3b Inhibitor XII TWS119. Dotted lines represent the DMSO condition. d, Lumen occupancy quantified by binning analysis for GSK3b Inhibitor XII TWS119 at different concentrations. Bin 1 has lowest lumen occupancy, which corresponds to a smaller lumen size, whereas bin 5 has the highest lumen occupancy corresponding to a bigger lumen size, z-score = 12.51 corresponds to a p = 0.012419, and the higher the z-score, the lower is the p value.

[0099] Fig. 9: Supporting transcriptomics data and endocrine differentiation (Related to Figs. 3 and 4) a, Live cells at Stage 6 day 14 for each condition observed by flow cytometry. b, Endocrine differentiation (S6dl4) by Rezania et al 2014 protocol and quantification by flow cytometry analysis for NEUROG3:tag-RFP-T, C-PEPTIDE and GLUCAGON. Data represented as mean ± SD. c, Heatmap showing centered rlog-normalized expression levels of proliferation and cell cycle markers from transcriptomics data. d, Mean fluorescence intensity levels of PDX1:H2B-GFP, PDX1 and NKX6-1 analysed by flow cytometry. N=3, WCM: WNT conditioned medium. Data represented as mean ± SD. **p<0.01, *p<0.05. e, Heatmap showing centered rlog-normalized expression levels of markers for foregut derived tissues from transcriptomics data.

[0100] Fig. 10:

[0101] Expression of acinar markers under various treatment conditions (Related to Fig.5) a, Expression of SOX9 and CHGA analyzed by qPCR. qPCR data on pancreas lineage markers, N=5. Data represented as mean ± SD. b and c, Combinations of Wnt activation by CHIR99021 and FGF inhibition, either by removing FGF2 or adding the FGF2 inhibitor (10 pM SU5402), were tested between day 3 to day 10. b, schematic representation of the experiment for different combinations of WNT activation and FGF inhibition, c, Gene expression levels analyzed by qPCR on acinar lineage markers upon treatment with FGFR inhibitor in the PE medium, N=3. Data represented as mean ± SD. Fig. 11:

[0102] Examples of GSK-3 inhibitors from MedChemExpress.

[0103] Fig 12:

[0104] Preferred examples of FGF receptor inhibitors from MedChemExpress.

[0105] Fig. 13:

[0106] Acinar cell differentiation in pancreatic progenitors cultured in 2D in CHIR without FGF2

[0107] Cells cultured for 1 day in ePE medium and for 2 additional days in ePE with CHIR and without FGF2 as compared to control conditions with ePE (containing FGF2) and DMSO. Immunostainings with the color-encoded antibodies (trypsin, amylase) as well as DAPI and Phal loidin.

[0108] Fig. 14:

[0109] Quantifications of acinar marker induction based on immunofluorescence images and qPCR (Related to Fig. 5) a, b Quantifications of the average intensity of CPA2 (D) and Amylase (F) in organoids upon concurrent CHIR treatment and FGF2 withdrawal from day 3 to day 10 relative to DMSO controls. Quantifications of the positive cell proportions with detectable CPA2 (D) and Amylase (F). N=3, >10 organoids averaged per biological replicate. Data represented as mean ± SD. ***p<0.001, **p<0.01. c, d, e Images of optical sections from whole-mount immunostained organoids for acinar markers. Concurrent CHIR treatment and FGF2 withdrawal from day 3 upregulate multiple acinar markers such as Trypsin (c), CPA1 (d), and UEA1 (e). Nuclei (Hoechst 33342) and F-actin ( Phall oidin) staining are also shown in conditions. Scale bars, 50 pm. N=5. f Gene expression analysis by qPCR for PP and acinar markers, across 3 hPSC lines (N=3 - 8). Fold change indicates relative fold change normalized against DMSO control. Data represented as mean ± SD. ****p<0.0001, ***p<0.001, **p<0.01, *p<0.05.

[0110] Fig. 15:

[0111] Combination of WNT activation and FGF inhibition induces acinar functional features

[0112] (A) Representative image of optical sections from whole-mount immunostained organoids in CHIR+noFGF2 condition. A subset of cells express the acinar marker Amylase with basal nuclei forming a bulging rosette-like acinar structure exhibiting apical constriction, rich in F-actin and abundant apical Golgi marked by GM130. Scale bar, 20 pm.

[0113] (B) Quantification of Amylase activity measured in supernatants from organoids cultured in CHIR+noFGF2 and control DMSO conditions. N=3. Data represented as mean ± SD. *p<0.05. (C) Quantification of trypsin concentration in supernatants, measured by ELISA from organoids cultured in CHIR+noFGF2 and control DMSO conditions. N=10. Data represented as mean ± SD.

[0114] *p<0.05.

[0115] (D) Quantification of Carboxypeptidase A2 concentration in supernatants, measured by ELISA from organoids cultured in CHIR+noFGF2 and control DMSO conditions. N=7. Data represented as mean ± SD.

[0116] Fig. 16:

[0117] Single-cell transcriptome reveals efficient but partial production of acinar cells upon combined WNT activation and FGF inhibition

[0118] (A) Experimental workflow of the single-cell sequencing experiments.

[0119] (B) Umap representation of clustering of the single-cell samples into 6 clusters.

[0120] (C) Annotation of the three treatment conditions on the Umap shown in B.

[0121] (D) Acinar marker enrichment upon CHIR and further enrichment upon CHIR+noFGF2.

[0122] (E) Progenitor marker global repression (PDX1, NKX6-1, HNF1B) upon CHIR and CHIR+noFGF2 treatment. No effect on HES1. Reduced HES1 in cluster 3 upon CHIR and CHIR+noFGF2 treatment.

[0123] (F-G, J) Violin plot showing the expression of specific marker expression in the 3 conditions, focusing on acinar (F), progenitor (G), ductal / progenitor (J) markers.

[0124] (H) Ductal / Progenitor marker repression upon CHIR and CHIR+noFGF2 treatment.

[0125] (I) Gene expression in the endocrine cluster comprising cells from all conditions.

[0126] (K) Quantitative dot plot representation of enrichment of specific markers across conditions.

[0127] Fig. 17:

[0128] Wnt activation induces tip progenitors rather than liver, foregut or posterior gut

[0129] (A) Heatmaps showing centered rlog-normalized expression levels of acinar markers from the set of DEGs in the transcriptomics data.

[0130] (B) Gene expression analysis by qPCR for foregut and liver markers across 3 pluripotent stem cell lines comparing CHIR treatment to control DMSO condition. N=4-8. Data represented as mean ± SD.

[0131] (C) Gene expression analysis by qPCR for posterior gut markers across 3 pluripotent stem cell lines comparing CHIR treatment to control DMSO condition. N=3-8. Data represented as mean ± SD.

[0132] Fig. 18:

[0133] Representative plots and gating strategies for FACS analysis (Related to Figures 3D and 4C)

[0134] (A) From a dissociated, stained cell population, single cells were selected, and then Ghost dye- negative population of live cells was selected for analysis. (B-E) From single live cells, positively stained cell populations were gated above the levels of fluorescence from antibody-specific isotype controls (B), as a negative control. Examples of positive populations of PDX1 and NKX6-1 are shown for DMSO (C), dlO CHIR-7days (D)), and dlO CHIR-4days (E) conditions.

[0135] Fig. 19:

[0136] Polarization and rosette organization of acinar and centroacinar / ductal / progenitor cells (Related to Figure 5)

[0137] (A-C) Representative images of optical sections from whole-mount immunostained organoids for acinar progenitor and polarity markers. Concurrent CHIR treatment and FGF2 withdrawal from day 3 upregulate multiple acinar markers such as Amylase (A, C, D) and Trypsin (B) expressed in polarized rosettes marked by PKCzeta (A), ZO-1 (B), and Ezrin (C).

[0138] (D) SOX9 appears enriched in some cells (indicated with white arrows) and down-regulated in others. Scale bars, 50 pm. N=4.

[0139] The examples illustrate the invention.

[0140] Example 1 - Results

[0141] Organoids have emerged as a valuable model system to investigate the mechanisms of human development, organ regeneration, function, and disease progression (Lewis et al 2021, Zhao et al 2022). Organoids derived from patient tissues or generated by genetic engineering enable studying how specific proteins or their variants affect these processes (Corsini & Knoblich 2022). Going beyond single genes, the ease of producing large amounts of material by the expansion of organoids enables multiplexing and studying many genes at a time (Lampart et al 2023, Ringel et al 2020, Ungricht et al 2022). Combined with single-cell sequencing, global molecular understanding of the effects of perturbations can be inferred. While extremely powerful, this approach does not provide information on morphological or spatial consequences of gene perturbations. For this purpose, high-content image-based screens are more suitable (Keshara et al 2022, Lukonin et al 2020, Lukonin et al 2021, Mead et al 2022, Suppinger et al 2023). However, they pose challenges in the scale of their implementation and analysis and have so far been underexplored, particularly to study pathways affecting human development.

[0142] Here, we conducted a high-content screen to investigate the mechanisms of human pancreas development focusing on both differentiation and morphogenesis. The adult pancreas is composed of acinar and ductal exocrine cells executing digestive functions and 5 types of endocrine cells releasing hormones involved in digestion and glucose homeostasis. These cells form developmentally from pancreatic progenitors (PPs) (Jennings et al 2015). PPs can be produced from human pluripotent stem cells (hPSCs) and expanded as organoids (Goncalves et al 2021, Hohwieler et al 2017). With readouts for cell identity and morphology, we screened the PP-organoids against a library of small molecules targeting kinases and pathways important in stem cells. We developed analysis methods that would identify changes in individual cell identity or more global identity shifts in organoids, as well as readouts of individual cell shape and organoid shape. We identified 54 compounds affecting at least one of the 439 phenotypic features and focused on inhibitors of glycogen synthase kinase 3 (GSK3) among the validated pathways. We found that GSK3 inhibition via WNT signaling has a global reversible effect repressing multiple PP markers and initiating acinar differentiation. We also identified an additional control, fibroblast growth factor (FGF) suppression that promotes further acinar differentiation uncovering the multistep mechanisms of human acinar cell specification.

[0143] A screening pipeline identifies morphological and differentiation extremes in a progenitor landscape

[0144] To easily assess cell identity, we generated a double reporter H9 human embryonic stem cell (h ESC) line for PDX1:H2B-GFP and NEUROG3:tagRFP-T to mark PPs and endocrine progenitors, respectively (Fig. 7a-c). We have previously reported a method to differentiate hPSCs in 2-dimensions (2D) into PPs and expand them as organoids (PP-organoids) in 3D that we adapted to 384-well plates (Fig. 7d) (Goncalves et al 2021). Following initial expansion for 3 days, these PP-organoids were treated against a library of 548 annotated small molecules (Fig. la) encompassing compounds that are cell-permeable and annotated for inhibition of protein kinases or regulation of stem cell biology targeting a diverse set of known biological pathways (Methods, Example 3). After 7 days of treatment, the PP-organoids were fixed and stained for DAPI (as a nuclear marker), and Phall oidin (an F-actin marker). Images were acquired with an automated spinning disc microscope for four channels (PDX1:H2B-GFP, NEUROG3:tag-RFP-T, DAPI, and Pha lloidin-AF647) and four planes in z at a distance of 10 pm (Fig. la). Individual nuclei were segmented based on DAPI signals whereas F-actin combined with DAPI was used to demarcate organoid boundaries. A set of 439 multivariate features for fluorescence intensity, radial intensity distributions, and shape was extracted for all four channels based on the segmented nuclei and organoids. The primary screen was conducted on three independent experiments generating about 1.38xl05images, including 603,902 profiled organoid objects and 19,883,057 nuclei. We set up 4 complementary analysis pipelines to data-mine the feature-rich dataset (as described in Methods, Example 3), leading up to the identification of 54 hit compounds. Prior to hit selection we excluded 212 compounds with more than 30% reduction in cell number - possibly due to toxicity or effects on proliferation.

[0145] The first pipeline - referenced as "well-mean analysis" (Fig. lb) - aims to find compounds changing the phenotypic profile based on well-averaged features, compared to the control treated with dimethyl sulfoxide (DMSO) based on well-averaged features. The profiles have been built based on a subset of features with low redundancy. Finally, 22 compounds were flagged as a hit as they showed reproducible feature profiles between the biological replicates. Example hits from the well-mean analysis are shown in Fig. lc and Fig. 7a.

[0146] The second pipeline, binning analysis, tackled the problem of heterogeneous morphologies that PP- organoids exhibited (Fig. 6d). Using average measurements of organoid objects per well has a limited discriminating power to identify phenotypes due to this heterogeneity, which led us to perform a subpopulation analysis for parameters accounting for the shape and size of objects, the fluorescence distribution as well as the lumen presence and relative dimension to the organoid (size). A change of the percentage of organoid in the low-, medium- or high-values bin for each feature reveals population shifts (Fig. ld,e and Fig. 7b). This analysis, binning analysis, led to the identification of 43 hit compounds.

[0147] The focus of the third pipeline, network analysis, was to capture compounds changing the spatial arrangement of cells within organoids. For each organoid we retrieved statistical features based on the network morphology built with the locations of nuclei as nodes and edges calculated by applying a Delaunay triangulation (Delaunay 1934) (Fig. If). We retrieved 8 hits as exemplified in Fig. lg and Fig. 7c.

[0148] The second and third analysis pipelines (both, binning and network analysis pipelines) do not use multiparametric profiles, because each feature have been analyzed independently. Therefore, in a fourth analysis, namely morphological clustering, we decided to use a clustering approach on a phenotypic fingerprint generated by a curated selection of features describing the morphology of each organoid. To determine clusters based on these fingerprints we first reduced the high dimensional dataset, including both control and experimental organoids, to 2D by applying the Uniform manifold approximation and projection (UMAP) to overcome the curse of dimensionality (Mclnnes et al 2018). A spectral clustering was then applied and split the projected dataset into several phenotypic clusters sharing similar fingerprints. The initial analysis of the whole (entire) dataset discriminated (distinguished) two big (main) clusters: objects with lumen and objects without lumen. Then we further refined the analysis on each of these independent clusters - organoids with and without lumen. 22 compounds showing increased or decreased presence of organoids within such a cluster were flagged as hits. The cluster enrichment of objects is shown in Fig. lj for an example hit. We visualized 6 clusters for the objects with lumen on the UMAP (Fig. lh,i) and 5 clusters for the objects without lumen (Fig. 7d-f). For objects with lumen, we observed distinct segregation based on the lumen morphology (Fig. li). Clusters 0 and 1 consisted of organoids with concentrated apical actin, with cluster 1 having the highest intensity levels. Clusters 2, 3 and 4 had organoids with single lumen, with cluster 2 being closer to cluster 1 and consisting of higher levels of apical actin levels. Cluster 3 consisted of objects (organoids) with mostly a monolayer of cells lining an enlarged lumen. Cluster 5 consisted of objects (organoids) with relatively smaller lumen size as well as relatively higher levels of actin on the basal side. The cluster enrichment of organoids is shown for an example hit in Figure lj.

[0149] Using collectively the four analysis pipelines outlined above, we identified a total of 54 hits from the primary screen, 6 of which were identified by all four analysis pipelines (Fig.lk). 11 compounds were selected for validation, prioritizing those identified by multiple analyses and the phenotypes observed. For those we performed dosage response validations with concentrations ranging from 0.1 to 10 mM (see examples in Fig 2f and h and Fig. 8). It is worth noting that no compounds specifically affected the endocrine differentiation readout NEUROG3:tagRFP-T.

[0150] GSK3 inhibition increases lumen occupancy and decreases PDX1 expression in PP-organoids

[0151] The targets with the most hits in the screen were glycogen synthase kinase 3a and b (GSK3A / B), two serine / threonine kinases involved in multiple biological pathways (Beurel et al 2015, Cohen & Frame 2001). Six GSK3 inhibitors were identified as hits from multiple analysis pipelines (Fig. 2a-e). Unlike the selective GSK3B inhibitors among the hits, CHIR99021 (CHIR), a widely used potent inhibitor of both GSK3A and GSK3B, displayed the strongest phenotype exhibiting changes in cell identity as well as the object (organoid) morphology; i.e., decrease in PDXl-proxy GFP intensity and increase in lumen occupancy of organoids (i.e. the ratio of lumen area to the total organoid area) (Fig. 2b,f,g). Upon validation, CHIR was not toxic in the tested range of concentrations (Fig. 8a), and its IC50 for GFP+ nuclei was 1 pM (Fig. 2f), whereas other inhibitors of both GSK3A and GSK3B such as TWS119 had an effect closer to their toxic dose (Figure 8b, d). To quantify the effect on morphologies across different concentrations, we performed binning analysis with five bins for morphological parameters. A parameter of particular interest was the lumen occupancy, which is the ratio of lumen area to the total organoid area. The higher the lumen occupancy, the larger is the lumen size. At concentrations higher than 1 pM, CHIR treatment significantly enriched organoids in bin 5 with larger lumens (Fig. 2g). Therefore, treatment with CHIR affected the cell identity as well as the morphology of PP-organoids. Additionally, morphological clustering was also performed on the organoids treated with different dosages of compounds during validation experiments, and this time 7 clusters were visualized for objects with lumen (Fig. 2h). Organoids treated with CHIR at 3 pM were enriched in cluster 2, exhibiting enlarged lumen (Fig. 2h,i), while at 0.1 pM they were distributed across all clusters (Fig. 2h, Fig. 8e). These observations thus validated the morphological phenotype observed in the screen. Taken together, our experiments show that GSK3 activity regulates PP identity marked by PDX1 expression and the morphology of PP-organoids.

[0152] GSK3 inhibition-induced phenotype is partially reversible upon removal of the compound

[0153] Using flow cytometry to examine the expression of PP marker proteins, we confirmed a reduction of PDX1:H2B-GFP intensity as well as of PDX1 and NKX6-1 expression, after 7 days of CHIR treatment (referred to as CHIR-7days) on organoids derived from multiple hPSC lines (Fig. 3a-d). To test whether CHIR triggered a permanent loss of PP identity or a temporary effect during drug treatment, PP- organoids were treated with CHIR for 4 days, i.e., from day 3 to day 7 and then with culture medium with DMSO for the last three days until day 10 (Fig. 3a). Hereafter, this treatment is referred to as CHIR- 4days. Expression of PDX1:H2B-GFP, PDX1, as well as NKX6-1 was restored to control levels within 3 days of withdrawal in the CHIR-4days condition (Fig. 3b-d). This was confirmed with multiple cell lines. Imaging of the PP-organoids also confirmed that the CHIR-4days samples had decreased PDX1:H2B- GFP and an enlarged lumen after 4 days of treatment at day 7. However, after 3 days of withdrawal at day 10, these organoids had regained PDX1:H2B-GFP expression but still retained a single enlarged lumen (Fig. 3e). Hence, we could rescue the cell identity phenotype, but not the morphological phenotype of CHIR treatment by removing the compound from the culture medium. This indicates that PPs maintain their cellular plasticity during temporary treatment of CHIR for 4 days, but the PP- organoid morphology is not as plastic as their gene expression, possibly owing to physical properties of organoids.

[0154] To identify underlying biological pathways and mechanisms of the phenotypes induced by CHIR on PP- organoids, we performed bulk RNA-sequencing on the CHIR-treated organoids. To examine transcriptional profiles of the phenotypes observed in the screen, day 10 samples of CHIR-7days and CHIR-4days conditions as well as DMSO control were collected. Additionally, to investigate short-term responses, samples at day 4, i.e. 1 day after CHIR treatment (d4 CHIR-lday) and DMSO control were also included (Fig. 3f). Principal component analysis revealed that d4 CHIR-lday and dlO CHIR-7days clustered separately from the d4 DMSO and dlO DMSO respectively (Fig. 3g). However, dlO CHIR-4days samples clustered closer to d 10 DMSO, whereas the dlO CHIR-7days samples clearly segregated distant from them (Fig. 3g), supporting the observed results of the partial rescue of the phenotype upon withdrawal of CHIR (Fig. 3b-e).

[0155] Transcriptome analysis further supported the loss of PP markers observed by microscopy (live imaging) and flow cytometry showing a reduction of PDX1 and NKX6.-1 at day 4 and 10, along with some PP markers such as ONECUT1, HNF1B, GATA4, and NR5A2 (Fig. 3h). This effect, however, was not general, as other PP markers such as SOX9, GATA6, TEAD1, and FOXA2 remained unchanged. Differentially expressed genes (DEGs) from the transcriptome were analyzed and tested for GO term enrichment (Fig. 4a).

[0156] Since a subset of PP markers was decreased by CHIR and were restored upon withdrawal of the compound, we investigated if a hallmark of PPs, namely their ability to differentiate into endocrine cells, was affected. To investigate their functionality, PP-organoids treated with DMSO, CHIR-7days or CHIR-4days were differentiated into pancreatic endocrine cells using the stage 5 and stage 6 media from the Rezania et al. 2014 protocol (Petersen et al 2017, Rezania et al 2014). Analysis of stage 6 day 14 cells by flow cytometry showed no differences in the percentage of C-PEPTIDE and GLUCAGON expressing cells, though the number of live cells at S6dl4 were reduced after CHIR treatment (Fig. 9a, b). Importantly, the differentiation media do not contain CHIR, supporting the observation that progenitors largely revert to their original potency and functionality after CHIR removal.

[0157] Canonical WNT signaling underlies the phenotype induced by GSK3 inhibition

[0158] In order to identify the biological pathways affected by CHIR treatment, we performed GO term enrichment on the differentially expressed genes (DEGs) from the transcriptome to identify the biological pathways affected upon CHIR treatment and it highlighted WNT signaling (Fig. 4a). While the first categories indicated changes in cellular processes, including gland development, the first pathway highlighted was WNT signaling. Identifying the most significantly changing genes in the transcriptome underlined the canonical WNT signaling (Fig. 4b), including the up-regulation of LGR5, TCF7, AXIN2, NKD1, NOTUM, and SP5 in d4 CHIR-lday and dlO CHIR-7days samples (Fig. 4a). Several WNT receptors such as FZ7 and FZ8 were down-regulated, possibly due to negative feedback, which is suggested by the substantial increase of the DVL negative regulator, NKD1 (Larraguibel et al 2015). However, little difference was observed in expression levels of these WNT genes between dlO DMSO and dlO CHIR- 4days samples (Fig. 4b), supporting again the rescue upon CHIR withdrawal (Fig. 3d,e). Another set of genes that were flagged by the DEGs were the proliferation and cell cycle regulators. Upregulation of these genes was observed at d4 CHIR-lday but not at dlO CHIR-7days (Fig. 9c), suggestive of a temporary proliferative effect in early phase upon CHIR treatment. Similar to phenotypes, little difference was observed in the expression levels of proliferation and progenitor markers between dlO DMSO and dlO CHIR-4days samples (Fig. 9c). Since the proliferative effect was temporary (temporary proliferative signature) and the lumen expansion was observed upon temporary treatment of CHIR between days 4 and 7 (Fig. 3e), it is likely that the lumen enlargement is due to increased proliferation rate at early time points after CHIR treatment.

[0159] The results from transcriptomics suggested a role of the canonical WNT signaling in driving the phenotypes of CHIR treatments. To test this hypothesis, PP-organoids were treated with WNT3A and RSPO1 proteins and a WNT-conditioned medium in different combinations (Fig. 4c and Fig. 9d). Analysis by flow cytometry showed little or no effect on PP expression upon WNT3A and RSP01 treatments (Fig. 9d). Nonetheless, upon supplementation with WNT-conditioned medium, a significant decrease in PDX1:H2B-GFP, PDX1 as well as NKX6-1 was confirmed (Fig. 4c). The morphological phenotypes were also reproduced after treatment with WNT3A and RSP01 in WNT-conditioned medium (Fig. 4d). Therefore, our results support that canonical WNT signaling may be responsible for the phenotypes induced by CHIR.

[0160] WNT activation in combination with inhibition of FGF signaling induces acinar differentiation of PP- organoids

[0161] To investigate the biological relevance of the phenotypes observed, we considered three hypotheses. First, a decrease in PDX1 and other PP markers would be expected if the PPs were converted into liver intestine or stomach (Fujitani et al 2006, Horb et al 2003, Offield et al 1996, Shen et al 2003). However, no (significant) increase was observed in marker expression of liver, intestine, stomach from the transcriptomics data (Fig. 9e). Secondly, differentiation of PPs into endocrine cells was also ruled out as these organoids were not positive for the reporter NEUROG3:tag-RFP-T. The third hypothesis was the induction of acinar development as a decrease in PDX1 expression is expected in the acinar lineage based on studies in mice (Ebrahim et al 2022, Miyatsuka et al 2006). In support of this hypothesis, our bulk seq (transcriptomics) data suggested an upregulation of BHLHA15, the gene encoding MISTI protein, a key transcription factor driving the pancreatic acinar cell differentiation program, along with the genes for digestive enzymes CPA4 and PRSS2, upon CHIR treatment (Fig. 5a). This was confirmed by qPCR, which also revealed an up-regulation of BHLHA15 and the digestive enzymes Carboxypeptidase A2 (CPA2) and Trypsinogen (PRSS1) (Fig. 5c, 14f). However, the fold increase was rather moderate (3-15 fold), and many other digestive enzymes were not affected by CHIR alone (Fig. 5c, 14f). We reasoned that even though acinar cells are first detected at E14.5 in mice, the first acinar- committed cells emerge in mice at Ell.5 (Larsen et al 2017) coinciding with the cessation of FGF10 production by the mesenchyme surrounding the pancreas epithelium (Bhushan et al 2001). We therefore hypothesized that a decrease in FGF signaling may be important for acinar cell induction. To test this hypothesis, we treated PP-organoids with different combinations of WNT activation via CHIR treatment and inhibition of FGF either by removal of FGF2 from the medium or addition of SU5402, an inhibitor of FGFR (Fig. 10a). Gene expression analysis confirmed that treatment with CHIR along with FGF inhibition either by removal of FGF2 or by addition of the FGFR inhibitor increased the expression of acinar markers such as MISTI by over 10-fold, CPA2 by over 100 fold, and PRSS1 by over 10 fold (Fig 5c, 14f and Fig. 10b). However, AMY2B was not up-regulated by CHIR with the removal of FGF2 (Fig. 5c, 14f), while it was highly upregulated by CHIR combined with both adding FGFR inhibitor and withdrawing FGF2 (Fig. 10b).

[0162] To further examine spatial expression and differentiation of acinar markers, we performed immunofluorescence staining. Immunostaining indeed confirmed the formation of rosette-like acinar structures, with concentrated F-apical actin, Ezrin, ZO-1, and aPKC, and larger nuclei at the basal side in organoids (Fig. 5d, Fig. 14c, d, Fig. 15a and Fig. 19a-c). These rosette-like cells expressed acinar enzymes such as Amylase, Trypsin, CPA1, CPA2, as well as being labelled with the acinar marker UEA- 1 (Baldan et al 2019) on the membrane (Fig. 5d, Fig. 14 and Fig 19a-c). Quantifications of images showed that while the average Amylase and Carboxypeptidase intensity was globally increased, the proportions of cells expressing enzymes were different, with an estimate of 15% cells for Amylase and 40% cells for Carboxypeptidase (Fig. 14a, b). Notably, cells expressing high levels of SOX9 were devoid of enzymes (Fig. 19D). Moreover, the enzyme-expressing cells exhibited an abundant apically- positioned Golgi, as expected for an exocrine cell (Fig. 15a) Additionally, electron microscopy showed the presence of abundant zymogen vesicles (Fig. 5e), a typical signature of pancreatic acinar cells, thereby confirming the presence of acinar cells in these organoids. Moreover, we examined the functionality of exocrine cells, assessing enzyme secretion and activity. Upon carbachol stimulation, organoids treated with CHIR and no FGF2 exhibited significantly increased Amylase activity, though to a modest level, and Trypsin secretion by 15-fold, although an increase in Carboxypeptidase A2 secretion was not statistically significant (Fig. 15b, d). We noted that there was rather a large variability across lines on these assays.

[0163] From the immunostainings, while Carboxypeptidase staining seemed to be observed in many cells, the cells marked by Amylase seemed more contrasted and yet more sparse, accompanied by distinctive proportions of cells expressing these enzymes (Fig. 14). To further explore these proportions and obtain a more unbiased readout of the effect of CHIR and CHIR in the absence of FGF2, we conducted single-cell RNA-sequencing on these conditions as compared to DMSO controls (Fig. 16a-c). This confirmed that acinar enzymes were induced in the conditions with CHIR and more so in CHIR without FGF2 (Fig. 16b-d, f, k). Moreover, we observed that only a subpopulation of cells expressed CPA1, CPA2, PRSS1, PRSS2, CEL, and up-regulated GATA6 (Fig. 16b-d, f, k). An unbiased assessment of GO categories enriched in each cluster also revealed an enrichment of genes in categories corresponding to zymogen, secretion and digestion in cluster 1 containing mostly organoids grown with CHIR without FGF2. Moreover, we observed that these enzymes were not always co-expressed in the same cells. While this may be explained by the detection limits of the method, this may also be interpreted as a partial or unsynchronized acinar program being induced in each cell. We confirmed that progenitor and ductal / progenitor markers were also down-regulated by CHIR (with and without FGF2) (Fig. 16e, g, h, k). A very small population of endocrine cells was found in all conditions (Fig. 161, k).

[0164] The culture medium established also induces acinar cell differentiation in 2D cultures of pancreas progenitors (Nakamura et al., 2022 doi: 10.1016 / j.stemcr.2022.03.013), as evidenced by the presence of Amylase and trypsin 2 days after CHIR treatment and retrieval of FGF2 (Fig. 13).

[0165] Taken together, the screen led us to identify multiple signaling molecules affecting human pancreas differentiation and morphogenesis in organoids in different ways. Particularly, concurrent activation of canonical WNT signaling and inhibition of FGF signaling are sufficient for pancreatic acinar lineage cell differentiation as well as terminal acinus-like morphogenesis and function during human pancreas development.

[0166] Example 2- Discussion

[0167] In this study, we developed image-based high-content screening in the context of human organogenesis, challenged it with a library of small molecules and established new readouts for morphological features by considering the innate heterogeneity in the organoids. In addition to the object morphology, our assay included a progenitor marker proxy, thereby giving a readout for cell fate. Though the classical well-mean analysis pipeline identified some hits, we show that a subpopulation binning analysis is a critical tool to reveal hits in a heterogeneous population of objects for a given parameter. In addition, we show that morphological clustering can classify different populations of organoids and identify how certain small molecules skew their distribution. Furthermore, network analysis quantifies nuclear arrangements inside organoids, providing morphological features of individual objects. Overall, our multi-modal assay exploits the potential of high-content analysis to explore cell identity as well as organoid morphology and could be applied to all parameters of interest in any organoid model systems. Furthermore, our assay may be extended to CRISPR-based genetic screens to screen for morphological features beyond genetic screening by sequencing.

[0168] Among the numerous hits identified in the screen, we show one example where the screen led to a biological discovery. Through GSK3 inhibition, we identified canonical WNT signaling as a modulator of PP identity as well as morphogenesis. CHIR, an inhibitor of GSK3A as well as GSK3B, showed the strongest phenotype among all the GSK3 inhibitors, suggesting that GSK3A and GSK3B may be functionally redundant for canonical WNT signaling in the pancreas, similar to what has been reported in mouse embryonic stem cells (Doble et al 2007). Studies from model organisms have shown that canonical WNT signaling is crucial for pancreas development at various stages. Its inhibition is critical for pancreas specification from the posterior foregut (Heller et al 2002, Mclin et al 2007, Nadauld et al 2004), a process that is already completed in our human PP-organoid cultures when we add CHIR (Funa et al 2024). Following pancreas specification, activation of WNT is essential for progenitor expansion in mice (Dessimoz et al 2005, Heiser et al 2006, Murtaugh et al 2005, Papadopoulou & Edlund 2005, Wells et al 2007) a process that is elicited in our human organoids through the presence of FGF (Goncalves et al 2021), another known cytokine promoting mouse PP proliferation (Bhushan et al 2001). The role of canonical WNT signaling on differentiation has been controversial but greatly clarified by Baumgartner et al. (Baumgartner et al 2014), who showed that inhibition of beta-catenin reduces the formation of pro-acinar tip progenitors and thereby has secondary effects on acinar and beta cells (Baumgartner et al 2014, Dessimoz et al 2005, Murtaugh et al 2005, Wells et al 2007). It may also have an inhibitory effect on the endocrine lineage (Pedersen & Heller 2005, Scheibner et al 2019, Sharon et al 2019). It was previously shown in mice that beta-catenin activation reduces the number of PDX1+ cells and that its inhibition increases PDX1 in the duodenum (Heiser et al 2006), suggesting that there is a conserved pathway by which WNT signaling via GSK3B and beta-catenin regulates PDX1. Here, we show that active WNT signaling in human PPs regulates the expression of a specific set of progenitor markers, as well as the morphology of organoids. We also show that WNT / CHIR do not trigger a transition but need to continuously signal to maintain this state as the removal of CHIR partially reversed the phenotype, suggesting a plasticity of this cell state. By triggering differentiation in this state, we show that the cells are still progenitors endowed with the ability to give rise to endocrine and acinar cells, alike the tip cells in mice (Zhou et al 2007).

[0169] We have previously shown that FGF signaling is essential for expansion of the progenitor pool during human pancreas development (Goncalves et al 2021). Here we demonstrate that in addition to WNT activation, removal of FGF signaling allows the multipotent PPs to progress to acinar fate. Studies in mice have shown that the level of FGF10 decreases drastically from Ell.5, coinciding with the initiation of acinar differentiation (Bhushan et al 2001, Elghazi et al 2002, Greggio et al 2013, Kobberup et al 2010). While previous studies have reported the initiation of acinar differentiation of hPSC-derived organoids (Hohwieler et al 2017, Huang et al 2021, Huang et al 2015, Merz et al 2023), the culture conditions we propose are simpler and induce more mature cells based on the levels of acinar marker induction, variety of markers observed, abundance of zymogen granules, and cellular organization into acini. In vivo, acinar enzymes in acinar cells reach 100- or 1000-fold levels detected in progenitors or other cell types, in a gradual way. Previously reported enrichments were in the order of 2- to 10-fold in those organoids, while our organoid acinar differentiation reaches 10- to 100-fold increases. Moreover, the medium composition we use is extremely simple. Thus, this system could provide a useful tool to model exocrine pancreas and acinar pathologies in humans, notably the initiation of adenocarcinoma. Since acinar cells are thought to be the cell of origin of a majority of pancreatic adenocarcinoma, this provides an easily accessible PSC-derived model to study the initiation of adenocarcinoma.

[0170] Example 3 - Methods

[0171] Pancreatic progenitor organoids derived from human pluripotent stem cells

[0172] Expansion and maintenance of human pluripotent stem cells

[0173] H9 and Hl hESC lines were obtained from WiCell. Reporters of PDX1:H2B-GFP and NEUROG3:tagRFP- T in H9 line (Fig. 6a) and NEUROG3-TagRFP-T:EGFP in Hl line (Beydag-Tasoz et al 2023) were generated in house. The hESC lines were approved for use in this project by the Robert Koch Institute (AZ: 3.04.02 / 0148). Human induced PSC (hiPSC) lines CRTD1 and CRTD11A were obtained from Center for Regenerative Therapies Dresden (CRTD), and wild-type and H2B-mEGFP reporter WTC11 hiPSC lines were obtained from Allen Institute for Cell Science. The PSCs were expanded and maintained in mTeSRl (Stem Cell Technologies) medium on hESC-qualified Matrigel (Growth Factor Reduced (GFR))(Corning). Cells were maintained at 37°C and 5% CO2, and the medium was changed daily. Upon reaching around 80% confluency, cells were dissociated with TrypLE (ThermoFisher), counted with a cell counter (Countess II FL) and seeded at a density of 40,000 cells / cm2. While passaging, the cells were supplemented with 10 pM ROCK inhibitor Y27632 (Cell Signaling) for the first 24 hours. The cells were regularly tested for Mycoplasma-free status.

[0174] Generation of PDXl and NEUROG3 dual reporter hESC line

[0175] To label PP with H2B-GFP on PDX1 locus and endocrine progenitor with tagRFP-T on NEUROG3, each locus was targeted by CRISPR / Cas9-mediated homology-directed repair. Reporter coding sequences of P2A-H2B-GFP and P2A-tagRFP-T-NLS were inserted before stop codon of PDX1 and NEUROG3 coding sequence, respectively (Fig. 6a). We used previously reported guide RNA sequences for PDX1 (Zhu et al 2015) and NEUROG3 (Beydag-Tasoz et al 2023). PDX1 locus was targeted first by electroporation with 2 plasmids, one containing guide RNA and CAS9 gene and the other containing the targeting construct with homology arms, and then a clone, mono-allelic targeted, was selected for second targeting of NEUROG3 locus. After the second targeting, a clone (L3-H9) was selected after differentiation validation. NEUROG3 targeting was also mono-allelic.

[0176] Differentiation of human pluripotent stem cells into pancreatic progenitor cells

[0177] The protocol for differentiation was originally published by Rezania et al (Rezania et al 2014), modified and adapted by us (Goncalves et al 2021). Briefly, cells were passaged and cultured in 2D on the plates coated with growth factor reduced Matrigel (Corning) at 1 / 30 dilution in DMEM / F12 Glutamax (ThermoFisher). Cells were seeded at a density of 350,000 cells / cm2in mTeSRl medium supplemented with 10 pM ROCK inhibitor for 24 hours before starting treatment with the differentiation media. Media changes were performed daily. The following differentiation media were used for the treatments: Stage 1: 1.5 g / L sodium bicarbonate, lx Glutamax, 10 mM Glucose and 0.5% fatty acid free bovine serum albumin in MCDB 131 basal medium. The stage medium was supplemented with 3 pM CHIR and 100 ng / ml Activin A for day 1, 0.3 pM CHIR and 100 ng / ml Activin A for day 2 and 100 ng / ml Activin A for day 3. Stage 2: 1.5 g / L sodium bicarbonate, lx Glutamax, 10 mM Glucose, 0.5% fatty acid free bovine serum albumin, 0.25 mM ascorbic acid and 50 ng / ml FGF 7 in MCDB 131 basal medium for 2 days. Stage 3: 2.5 g / L sodium bicarbonate, lx Glutamax, 10 mM Glucose, 2% fatty acid free bovine serum albumin, 0.25 mM ascorbic acid, 1:200 ITS-X, 1 pM retinoic acid, 0.25 pM Sant-1, 100 nM LDN193189, 200 nM TPB and 50 ng / ml FGF 7 in MCDB 131 basal medium for 2 days. Stage 4: 2.5 g / L sodium bicarbonate, lx Glutamax, 10 mM Glucose, 2% fatty acid free bovine serum albumin, 0.25 mM ascorbic acid, 1:200 ITS-X, 0.1 pM retinoic acid, 0.25 pM Sant-1, 200 nM LDN193189, 100 nM TPB and 2 ng / ml FGF 7 in MCDB 131 basal medium for 3 days. At stage 4 day 3, 97% of these cells expressed PDX1:H2B-GFP, a PP proxy, and 15% expressed NEUROG3:tagRFP-T, an endocrine progenitor proxy (Fig. 6c, d).

[0178] Expansion and maintenance of hPSC-derived pancreatic progenitor organoids

[0179] Pancreatic progenitor cells at Stage 4 day 3 were harvested using TrypLE and seeded at a density of 40,000 cells in a 40 pl dome of 75% growth factor-reduced (GFR) Matrigel (Corning) per well on a 24- well plate. Upon solidification, the Matrigel domes were incubated in pancreatic epithelium (PE) medium (Goncalves et al 2021) consisting of lx B27 (ThermoFischer), 64 ng / ml FGF2 (Peprotech), and 10 pM ROCK inhibitor in DMEM / F12-Glutamax at 37°C with 5% CO2. Regular medium changes were performed every three days. Embedded in GFR Matrigel, PP cells cluster, proliferate, and self-organize to form PP-organoids. Organoids were passaged every ten days by dissociating into mostly single cells using TrypLE and re-seeding in 75% GFR Matrigel. Using this protocol, PP- organoids can be maintained in long-term cultures including freeze-and-thaw cycles. For quality controls, PP-organoids were regularly tested for expression of PDX1 (>99%) and NKX6-1 (>70%) by flow cytometry and also for Mycoplasma-free status.

[0180] Expansion and maintenance of hPSC-derived pancreatic progenitors in 2D

[0181] Polystyrene cell culture plates (Corning) were coated with FN human plasma solution (Sigma-Aldrich) for 1-2 h at room temperature at the concentration of 100 pg / mL. Differentiated PE cells or sorted PE cells were re-suspended in ePE medium (DMEM / F12 GlutaMAX (Thermo Fisher Scientific), 64 ng / mL FGF2 (Peprotech), lx B-27 (Thermo Fisher Scientific), and 10 pM SB431542 (Santa Cruz Biotechnology) supplemented with 10 pM ROCK inhibitor Y-27632 (Sigma). The cells were plated at a density of 1.58 x 105 / cm2at P0. The medium was changed to ePE medium without ROCK inhibitor 24 h after re-plating and then the cells were fed every 2-3 days until passaging after 5-7 days in culture for cells at P0-P5 or 3-4 days cells at later passages.

[0182] Image-based high-content screening of pancreatic progenitor organoids

[0183] High-content screening assay on pancreatic progenitor organoids

[0184] The screening was carried out in the High Throughput Technology Development Studio (TDS) in MPI- CBG. Dissociated cells from PDX1 and NEUROG3 dual reporter PP-organoids at day 10 were seeded in 384-well plates (Greiner, cat. 788986) as 3 pl suspensions in 75% GFR Matrigel at a density of 1,000 cells / pl using an automated dispenser Multidrop (ThermoFischer). After 10 min of solidification at 37°C, 18 pl of PE medium was added per well.

[0185] We used in total 548 annotated small molecules form a set of compound libraries (StemSelect™ Small Molecule Regulators Library I, cat. 569744, InhibitorSelect™ Protein Kinase Inhibitor Library I, II, and III, cat. 539744, 539745, and 539746, Calbiochem, EMD Bioscience, Merck). These libraries contain compounds that are cell-permeable and annotated for inhibition of protein kinases or regulation of stem cell biology and provided a good source targeting a diverse set of known biological pathways. These compounds are 10 mM stock solutions in Dimethyl Sulfoxide (DMSO). Each screening plate consisted of 28 wells of negative control DMSO, and the library compounds were distributed in three plates as one well per compound using an acoustic dispenser (Echo® 550, Labcyte). For compound treatments on day 3, the plates were first washed three times with DMEM / F12-Glutamax. Compounds were diluted to 2 pM in 2x concentrated PE medium without ROCK inhibitor and 10 ml of PE medium was added into wells containing 10 pl of DMEM / F12-Glutamax to achieve a final concentration of 1 pM. Compound-containing media were exchanged on days 3, 5, 7 and 9 using an automated liquid handling workstation (Fluent 1080 and HydroSpeed, Tecan). The screen was performed as three biological replicates. The total duration of culture was 10 days, including 7 days with compounds to account for the slow development in human (considering cell cycle length of approximately 40 hours based on live imaging and differentiation of endocrine cells taking several days from progenitors to hormone expression).

[0186] Fixation, Immunostaining and image acguisition

[0187] On day 10, plates were briefly centrifuged until 1000 rpm was reached to prevent wash away of objects, and the PP-organoids were fixed with 3.7% formaldehyde (FA) for 15 min at room temperature (RT) by addition of the 2x concentrated fixation solution (7.4% FA in PBS) using the Well Mate (ThermoScientific). After seven washes with lx PBS using the HydroSpeed (Tecan), organoids were permeabilized with 0.5% Triton X-100 in PBS for 5 min at RT. Organoids were then incubated with 165 nM Phalloidin and 1.5 mg / ml DAPI in PBS containing 1% bovine serum albumin for 3 hours at RT and stored in PBS containing 0.02% sodium azide at 4°C until imaging. Images were acquired for 4 fields and 4 z-positions (Z step 10 pm) for each well of the 384-well plates for 4 channels (DAPI: excitation 405 nm and emission 445 / 45 nm, PDX1:H2B-GFP: 488 nm and 525 / 50 nm , NEUROG3:tagRFP-T: 561 nm and 600 / 37 nm and Phalloidin : 640 nm and 676 / 29 nm) with a 20X objective (NA 0.75) of an automated spinning disc microscope (Yokogawa CellVoyager CV7000).

[0188] Image segmentation and feature extraction

[0189] Images from 4 fields were stitched together with a FIJI macro (Schindelin et al 2012) to generate one image per well for each z-position, 4 images per well. Cell nuclei were then segmented from this image using StarDist in the DAPI channel (Schmidt et al 2018). The nuclei labels and the images of all 4 channels were then imported into a CellProfiler (Carpenter et al 2006) image analysis pipeline. We used a combination of the DAPI and Phalloidin signals to segment organoids in 2D for each z-plane. Nuclei objects identified by the StarDist segmentation were then related to the parent organoid. Finally, features describing shape, fluorescence intensities and radial intensity distributions of the segmented organoid and nuclei objects have been measured on each channel and exported as a csv file. Altogether we extracted 351 features per organoid and 88 features per nuclei. All macros, scripts and pipelines have been developed in the TDS.

[0190] Hit-selection KNIME (version 4.5) (Berthold et al 2007) was used for the following data analysis. Organoid objects touching the border as well as containing less than 7 nuclei were excluded from the analysis. Additionally, objects from the lowest z-plane were not taken into account. Cell counts per well have been expressed as percent of control (POC) with DMSO-only treated wells as control. Treatments resulting in more than 30% cell reduction have been discarded from all hitlists.

[0191] Well-mean analysis: Out of 218 features, scores of each feature have been averaged per well. These well means were then normalized per plate by applying a POC normalization using control wells treated with DMSO only as reference. A second normalization step was then applied on the whole screen by calculating z-values based on statistical descriptors of the distribution of DMSO-treated controls (Z- score; Z = (x - m) / s with x is the experimental value, m and s are respectively the average and standard deviation of the normalizing control). We then applied a feature selection by first calculating a correlation matrix based on the Pearson correlation coefficient (PCC) between all features to eliminate redundancy. We clustered highly correlating parameters into 65 groups with | PCC | > 0.5 by using hierarchical clustering with a distance measure being 1 - | PCC | . A curated selection of maximum one parameter per group led us to 64 features which built well-based multiparametric profiles. We measured reproducibility between the profiles of the three runs by computing the PCC between the different biological run for each compound generating three sets of values for each comparison: run A and B, run A and C, run B and C. A compound was flagged as a hit when at least 2 out of 3 PCCs > 0.5. Subpopulation Analysis / binning analysis: the distributions of the values of the DMSO-treated control for a curated set of 34 different parameters related to size, shape, and morphology of the organoid objects were divided into three bins using the intervals [minimum value; 33rdpercentile], [33rdpercentile; 66thpercentile] and [66thpercentile; maximum] (Stoter et al 2019). Using these bins for each feature, we counted the number of organoids and determined their percentage in comparison to the DMSO reference values per well and bin. Finally, these values have been expressed as z-scores (see Well-means analysis). For each feature bin combination, it was checked whether a compound showed an increase or decrease with z-score > 131 in at least two runs to consider it as a hit.

[0192] Network analysis: To determine shape and cell arrangement of each organoid object we used the xy locations of all corresponding nuclei as nodes. Then we performed a Delaunay triangulation to build a network (Barber et al 1996). 4 parameters describing this network have been extracted for each organoid and then averaged per well. We applied the same normalization steps as in the well-means analysis to retrieve z-scores per feature and compound and the same hit selection criteria of the binning analysis to build the hit list.

[0193] Morphological clustering: We built a multi-parametric profile using a curated set of 34 morphology features for each organoid, adding a binning analysis with 10 bins on the edge distances of all nuclei. All features were normalized as POC first followed by z-score. We then applied a UMAP projection of the resulting normalized morphology profiles, into a 2D space to facilitate visualization. As a few features are only available on organoids with the presence of a lumen, the dataset has been split into organoids with and without lumen. The following analysis has been carried out on both sets and the hitlist have been merged. We clustered the UMAP-features of the organoid objects by applying spectral clustering (von Luxburg 2007) with k=6 for organoids with lumen and k=5 for organoids without lumen. For each cluster the percentage of organoids per well assigned to it has been calculated and z-score normalized. The hit selection followed the criteria described in the binning analysis but applying z-score > 12.51 as a threshold.

[0194] Though we observed hits for features extracted based on the RFP channel, no significant changes in the expression of NEUROG3:tagRFP-T were observed in the nuclei. Several compounds with signals in this channel were excluded because they caused autofluorescence. These were generally detected in multiple channels (though some compounds fluoresced specifically in the red channel), and when checked in HeLa cells, which don't have the potential to differentiate into pancreatic endocrine cells, the same fluorescence was observed. Therefore, we did not pursue the hits from the RFP channel.

[0195] Dose response assay for hit validation

[0196] Eleven hit compounds were selected for validation. Ten concentrations were chosen from a wide range of 0.1 pM to 10 pM, and assay was performed in two independent experiments with three replicates each. For the assay, imaging, and analyses, same protocols were applied as described above for the screen. We observed similar results from the two independent experiments, and data from one experiment are shown in Figs. 2f,g,i and 8.

[0197] Bulk RNA-sequencing

[0198] PP-organoid culture and sample preparation for RNA-sequencing

[0199] PDX1:H2B-GFP PP-organoids were treated with DMSO or 3 pM CHIR from day 3 onward with media changes performed on every alternate day. For CHIR-4days samples, CHIR was treated only from day 3 to day 7, followed by DMSO from day 7 to day 10 to observe persistent effects after drug removal. Organoids were harvested as mostly single cells using TrypLE on day 4 and day 10 and snap frozen in dry ice for RNA extraction. For CHIR-treated conditions, three biological replicates while four for DMSO were collected. Total RNA was extracted from the samples using the RNeasy micro kit (QIAGEN) by following manufacturer's guidelines, and quality assessment was done with Bioanalyzer (Agilent). Samples of RNA were sequenced with Smart-Seq2 at Dresden Concept Genome Center in CRTD. Bulk RNA-sequencing data analysis

[0200] 29 paired-end (2x101) Illumina read data sets of sizes between 26.4 Mio and 37.9 Mio were processed. Illumina universal adapters were trimmed from both ends of the reads with Cutadapt vl.16 discarding reads trimmed to a length shorter than 19 nt. Quality of reads was assessed using FastQC 0.11.9 (Andrews 2010). Reads were mapped against the Homo sapiens genome reference assembly GRCh38, and genes of the Ensembl release v99 (Cunningham et al 2019) were quantified using STAR 2.7.3a (Dobin et al 2013). Genes with at least 10 reads in at least one sample were input into the analysis of differential gene expression with DESeq2 vl.22.1 (Love et al 2014). Genes considered to be differentially expressed had to have an FDR<1% considering p-values from the Wald test implemented in DESeq2 and a log2FC of min 2 (or max -2). The sets of differentially expressed genes were tested for Gene Ontology (GO) term (Kegg, GO, Reactome) enrichment with clusterProfiler v3.10.1 reporting hits at FDR<1%, and enriched Kegg pathways were visualized with pathview vl.20.0. Gene-set-enrichment analyses were performed with fgsea vl.8.0 and MSigDB v7.2.1 with the hallmark gene set (H), curated gene sets (C2), and ontology gene sets (C5) ranking all genes according to their logl0-DESeq2-p-value (multiplied by -1 if the FC<1), reporting hits at FDR<1%. DESeq2, fgsea, and clusterProfiler were run in R 3.5.1.

[0201] Endocrine differentiation of PP- organoids (based on Stage 5 and 6 of Rezania et al. protocol)

[0202] PP-organoids were treated with CHIR from day 3 onward, changing medium on alternate days. On day 10, the organoids were washed with lx PBS and treated with stage 5 medium (Rezania et al 2014) for three days, changing medium daily. On day 13, organoids were washed with lx PBS and treated with stage 6 medium for 14 days, changing medium on alternate days. Cells were harvested, and analyzed by flow cytometry for C-PEP and GCG after stage 6 (on day 14).

[0203] Stage 5 medium: 1.5 g / L sodium bicarbonate, lx Glutamax, 10 U / ml penicillin-streptomycin, 20 mM Glucose, 2% fatty acid free bovine serum albumin, 1:200 ITS-X, 10 pM Zinc sulfate, 0.05 pM retinoic acid, 0.25 pM Sant-1, 100 nM LDN193189, 10 pM ALK5 inhibitor II, 1 pM T3 and 10 pg / ml Heparin in MCDB 131 basal medium. Stage 6 medium: 1.5 g / L sodium bicarbonate, lx Glutamax, 10 U / ml penicillin-streptomycin, 20 mM Glucose, 2% fatty acid free bovine serum albumin, 1:200 ITS-X, 10 pM Zinc sulfate, 100 nM LDN193189, 10 pM ALK5 inhibitor II, 1 pM T3, 100 nM GSiXX and 10 pg / ml Heparin in MCDB 131 basal medium.

[0204] WNT3A and RSPO1 treatment

[0205] From day 3 onward, PDX1:H2B-GFP PP-organoids were treated with PE medium supplemented with 30% WNT-conditioned medium along with WNT3A and RSP01 (250 ng / ml and 750 ng / ml respectively). WNT-conditioned medium was a kind gift from Meritxell Huch (Barker et al 2010, Willert et al 2003). PE medium, DMSO and 3 pM CHIR treatments were used as controls and media changes were performed as explained in the protocol for the screening assay. Samples were analyzed on day 10 by flow cytometry and whole-mount immunostaining.

[0206] Flow cytometry

[0207] Dissociated cells from PP-organoids were washed in lx PBS and stained with Ghost Dye in PBS (1:1000) (Cell Signaling Technology) for 10 min at 4°C to stain dead cells before fixation. Cells were then washed with lx PBS and fixed with 4% FA for 10 min at RT. Fixed cells were washed with PBS and permeabilized for 20 min at 4°C with 0.2% Triton X-100 and 5% donkey serum in PBS. After permeabilization, cells were incubated with primary antibodies in the blocking buffer consisting of 0.1% Triton X-100 and 5% donkey serum in PBS overnight at 4°C. For the unconjugated antibodies, the cells were further incubated with the secondary antibodies in the blocking buffer for 45 min at RT. Cells were then analyzed using FACSAria III (BD Biosciences), FCS Express 7 software (De Novo Software) and FlowJo 10.10.0 (BD Biosciences). Analysis was performed for at least 10,000 live cells.

[0208] Whole-mount immunostaining of PP- organoids

[0209] PP-organoids were fixed with 4% FA for 20 min at RT and washed three times with lx PBS. Organoids were permeabilized with 0.5% Triton X-100 in PBS for 15 min at RT and incubated at RT with the blocking buffer made of 0.5% Triton X-100 and 3% BSA in PBS. Primary antibodies were added in the blocking buffer for 48 hours, followed by PBS washes for three times and secondary antibodies in the blocking buffer for 48 hours. For clearing, 60% (vol / vol) glycerol and 2.5 M fructose was added and incubated overnight at 4°C. Images were acquired using the Yokogawa CellVoyager CV7000 or Zeiss LSM 780 with Zen Black software and analyzed with ImageJ software.

[0210] Immunostaining of 2D cultures

[0211] The cells were permeabilised with 0.5% Triton X-100 in PBS (PBT) for 5 min at RT, followed by the Blocking solution (PBS + 1% BSA + 0.5% Triton X-100) for lh at RT. They were incubated with primary antibodies in Blocking solution for 24 hrs at 4°C, followed by 3x washes with PBS and secondary antibodies in Blocking solution for 24 hrs at 4°C. Phall oidin and DAPI were added after 3x washes for 3 hrs at RT. After 3x washes with PBS, the samples are ready for imaging.

[0212] Live imaging of PP-organoids

[0213] PP-organoids were cultured in 5 pl domes of 75% GFR Matrigel with 200 pl PE media in a 96-well plate. CHIR treatments were performed as explained above from day 3 to day 10. PP-organoids were imaged using Yokogawa CellVoyager CV7000 immediately after the compound addition on day 3, followed by every 24 hours until day 10. Images were acquired with the 20x objective in green channel as well as bright field, covering a z-distance of 200 pm with a step-size of 10 pm.

[0214] Acinar fate induction protocol

[0215] PP-organoids were seeded on day 0 in 75% GFR Matrigel domes and treated with PE medium as described previously. On day 3, the organoids are treated with the modified PE medium containing 3 pM CHIR without FGF2 for seven days. 10 pM of SU5402, an inhibitor of FGFR, was also added along with CHIR in PE medium in the appropriately mentioned conditions. In some experiments windows of treatment with the different compounds from day 3 to 7 or day 7 to 10 were also tested. Media changes were performed on days 5, 7 and 9. The spheres were harvested and fixed on day 10 for analysis. Experiments were performed with hESC- and hiPSC-derived PP-organoids.

[0216] For induction in 2D, ePE cells were seeded on Fibronectin- (100 pg / ml) coated plates on day 0 and treated with ePE medium (64 ng / ml FGF2, lx B27 and 10 pM SB431542 in DMEM / F12-Glutamax) supplemented with 10 pM Rock Inhibitor Y-27632 (refer Nakamura et al 2022 doi: 10.1016 / j.stemcr.2022.03.013. ). On day 1, the cells are treated with the modified ePE medium containing 3 pM CHIR without FGF2 and without Rock Inhibitor for two days. The cells were fixed on day 3 for immunostaining.

[0217] Gene expression analysis by qPCR

[0218] The organoids were harvested as single cells after TrypLE treatment as described before and snap frozen as pellets in -80°C until processing. RNA was extracted using the QIAGEN micro kit by following the manufacturer instructions. First-strand cDNA synthesis was performed with Superscript III system (Thermo Fisher) using random primers (Thermo Fisher). Quantitative PCR was performed using SYBR- Green (Thermo Fisher) on LightCycler 480 II instrument (Roche) for 96 well plates. PCR primers were selected based on published data or designed using NCBI primer design tool and validated for efficiency ranging between 95 and 105%. Expression values for each gene were normalized against ACTB and then the DMSO control, using the delta CT and delta-delta CT methods, respectively.

[0219] Transmission Electron Microscopy

[0220] Human pancreas organoids were collected and fixed in modified Karnovsky's fixative, in 1% glutaraldehyde and 2% paraformaldehyde in lOOmM phosphate buffer (PB, pH 7.4) lhr at room temperature and stored overnight at 4°C. Samples were washed and further postfixed in 1% aqueous OSO4 solution containing 1.5% potassium ferrocyanide. After washing samples were incubated and en bloc contrasted with 0.5% uranyl acetate / water overnight at 4°C, washed, dehydrated in a graded ethanol series, infiltrated into EMBed 812 resin and placed in embedding molds (Science Services GmbH). The embedded Organoids were polymerized at 60°C for 2 days. Ultrathin sections were cut with a Leica UCT ultramicrotome (Leica Microsystems) and collected on formvar-coated slot grids. The 70 nm sections were poststained with 1% uranyl acetate and 0.4% lead citrate. Electron micrographs were obtained at a 100 kV Tecnail2 (ThermoFisher) TEM with a digital camera (TVIPS TemCamF416).

[0221] Enzyme secretion stimulation and enzyme activity assays

[0222] PP-organoids were differentiated toward the acinar lineage as described above. On day 10, organoids were collected from Matrigel domes by removing the growth medium, washing three times with PBS, and incubating with TrypLE for 4 minutes at 37°C. TrypLE was neutralized by adding growth medium containing 10% FBS, followed by three PBS washes (centrifugation at 500 g for 3 minutes between washes). To stimulate enzyme secretion, 200 pl of 10 pM carbachol in HEPES buffer was added to each well, and cells were incubated for 2 hours at 37°C in a humidified incubator. Following stimulation, cell supernatants were collected and centrifuged at maximum speed for 2 minutes to remove debris. Cleared supernatants were snap-frozen on dry ice and stored at -80°C until further analysis. Concentrations and activities of enzymes (Amylase, Trypsin, and Carboxypeptidase A2) in the supernatants were determined by ELISA and enzyme activity assays according to the manufacturers' instructions. The source of assay kits are listed in Table S3.

[0223] Image segmentation and analysis of acinar-differentiated organoids

[0224] All analyses of images immunostained for acinar markers, amylase, and CPA2 were performed on CentOS Linux 7.4.1708 running Python 3.10.7. For nuclei segmentation, we used a pretrained 3D Cellpose model ('CP'), and trained it for an additional 100 epochs on an annotated subset of images from our own data (14 slices total from 3 different image volumes). During model prediction, we first smoothed the image with a Gaussian filter (sigma=2) and used a flow threshold of 0.4. After segmentation with Cellpose, we additionally removed segmented objects with an area less than 5000 pixels. For organoid segmentation, we first applied a Gaussian filter (sigma =10) for each image volume and then used the Triangle threshold from the scikit-image library to obtain an initial organoid mask. We then kept only the largest thresholded volume and applied post-processing steps of binary closing and erosion to further refine the mask. Image analysis was performed to quantify marker expression at the single-cell level. Individual cells were segmented based on nuclear staining, and mean fluorescence intensity was measured for each marker channel within the segmented regions. For each marker, mean intensity values were inspected and thresholded to classify cells as positive or negative. The number of positive and negative cells was then determined for each organoid. For each replicate, the proportion of marker-positive cells per organoid, and then the average value across organoids, was calculated. Single-cell RNA-sequencing and analysis

[0225] PP-organoids (Hl hESC-derived) were differentiated toward the acinar lineage as described above. On day 10, organoids were dissociated into single cells using TrypLE, and cells were sorted into single cells by FACS based on cell viability staining by DAPL The viability of single-cell suspensions was 90% (DMSO), 80% (CHIR), and 95% (CHIR noFGF2). For each condition, 27,000 cells were loaded onto a lOx Genomics Chromium GEM-X microfluidic chip (Single Cell 3' GEM-X, v4 chemistry), targeting "20,000 captured cells per sample. cDNA amplification was performed for 10 cycles, followed by 0.6x SPRI purification. Libraries were sequenced on an Illumina NovaSeq 6000 (NovaSeq S4 vl.5, 4XP, 200 cycles). Base calling was performed with bcl2fastq2 (v2.20.0). Sequencing yielded 930,222,396 (DMSO), 857,521,610 (CHIR), and 699,688,065 (CHIR noFGF2) reads per sample.

[0226] Reads were demultiplexed and processed with Cell Ranger (v9.0.1, lOx Genomics) using parameters - -chemistry=auto and -include-introns=true, aligned to the GRCh38 lOx Genomics reference (refdata- gex-GRCh38-2024-A). Cell Ranger invoked Martian Runtime (v4.0.13), Python 3.10.8, and dependencies numpy (vl.26.4), scipy (vl.10.1), pysam (v0.21.0), h5py (v3.9.0), pandas (v2.1.4), and STAR (v2.7.2a). The minimum fraction of valid barcodes was 97.3%, and the fraction of valid UMIs reached 100%. Barcode-rank plots were inspected, and a second Cell Ranger run was executed with manually set expected cell numbers (--force-cells): 21,200 (DMSO), 16,000 (CHIR), and 20,200 (CHIR noFGF2). Mapping quality was high, with >93.3% of reads confidently mapped to the genome, >85% to the transcriptome, <3.6% mapping to intergenic regions, and >93.2% reads in cells. The median number of detected genes per cell was >4,163.

[0227] Spliced and unspliced transcript counts were derived using Velocyto (vO.17.17) applied to the Cell Ranger-generated BAM files. Output matrices were stored in MatrixMarket format (v2). Filtered count matrices (filtered_feature_bc_matrix) from the second Cell Ranger run were used for all downstream analyses.

[0228] Downstream analyses were performed in R (v4.0.5) using Seurat (v4.1.1), SeuratWrappers (v20210208), SingleCellExperiment (vl.12.0), SCTransform (vO.3.5), Scater (vl.18.6), VeloCyto.R (v0.6), scDblFinder (vl.15.1), and DoubletFinder (v2.0.3). Cells were filtered to retain those with >750 reads, >350 detected genes, and <15% mitochondrial read fraction; genes were required to be expressed in >3 cells. The mitochondrial gene fraction per cell was calculated from gene annotation. Normalization and variance stabilization were performed with SCTransform, regressing out (1) mitochondrial read fraction, (2) S. Score, and (3) G2M. Score computed by Seurat's CellCycleScoring function. Normalized data were centered and scaled. No data integration was applied as it erased all biological variation previously observed in bulk sequencing. We note that data integration has been developed with the intent of comparing different populations with overlapping populations of cells where anchors can be easily extracted but not for perturbations leading to large shifts in cell identity. As all three conditions were processed in parallel, low technical noise is expected. This was confirmed by the fact that all endocrine cells from all conditions formed a distinct cluster and that part of the cells in CHIR+No FGF cluster with CHIR alone. Principal component analysis was performed on the 3,000 most variable genes. Cells were clustered using FindClusters (50 iterations, k = 75, entrees = 100, resolution = 0.1), and marker genes were identified with FindAIIMarkers (only.pos = TRUE, min. pct = 0.25, logfc. threshold = 0.25, test, use = "wilcox", ba e = 2) after running PrepSCTFindMarkers. Visualization was performed in Seurat (v5.3.3) and ggplot2 (v3.5.2).

[0229] Quantification and statistical analysis

[0230] Statistical tests were done using KNIME 4.5 or GraphPad Prism 10. All p-values were calculated using a one sample t-test or two way-ANOVA. For the screen data, z-scores were converted to p-values based on two-tailed hypothesis. Significance was defined as *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001 and *****p<o.00001. Experimental repeats are indicated in the figure legends. "N" denotes the number of independent experiments (biological repeats), and "n" the total number of measurements. All data are presented as mean ± standard deviation.

[0231] References

[0232] Andrews S. 2010. FastQC: A Quality Control Tool for High Throughput Sequence Data

[0233] Backx E, Coolens K, Van den Bossche JL, Houbracken I, Espinet E, Rooman I. 2022. On the Origin of Pancreatic Cancer: Molecular Tumor Subtypes in Perspective of Exocrine Cell Plasticity. Cell Mol Gastroenterol Hepatol 13: 1243-53

[0234] Baldan J, Houbracken I, Rooman I, Bouwens L. 2019. Adult human pancreatic acinar cells dedifferentiate into an embryonic progenitor-like state in 3D suspension culture. Sci Rep 9: 4040

[0235] Barber CB, Dobkin DP, Huhdanpaa H. 1996. The Quickhull algorithm for convex hulls. Acm T Math Software 22: 469-83

[0236] Barker N, Huch M, Kujala P, van de Wetering M, Snippert HJ, et al. 2010. Lgr5(+ve) stem cells drive selfrenewal in the stomach and build long-lived gastric units in vitro. Cell Stem Cell 6: 25-36

[0237] Baumgartner BK, Cash G, Hansen H, Ostler S, Murtaugh LC. 2014. Distinct requirements for beta- catenin in pancreatic epithelial growth and patterning. Dev Biol 391: 89-98

[0238] Berthold MR, Cebron N, Dill F, Gabriel TR, Kotter T, et al. 2007. KNIME: The Konstanz Information Miner. Stud Class Data Anal: 319-26

[0239] Beurel E, Grieco SF, Jope RS. 2015. Glycogen synthase kinase-3 (GSK3): regulation, actions, and diseases. Pharmacol Ther 148: 114-31

[0240] Beydag-Tasoz BS, D'Costa JV, Hersemann L, Lee BH, Luppino F, et al. 2023. Integrating single-cell imaging and RNA sequencing datasets links differentiation and morphogenetic dynamics of human pancreatic endocrine progenitors. Dev Cell Bhushan A, Itoh N, Kato S, Thiery JP, Czernichow P, et al. 2001. FgflO is essential for maintaining the proliferative capacity of epithelial progenitor cells during early pancreatic organogenesis. Development 128: 5109-17

[0241] Carpenter AE, Jones TR, Lamprecht MR, Clarke C, Kang IH, et al. 2006. CellProfiler: image analysis software for identifying and quantifying cell phenotypes. Genome Biol 7

[0242] Cohen P, Frame S. 2001. The renaissance of GSK3. Nat Rev Mol Cell Biol 2: 769-76

[0243] Corsini NS, Knoblich JA. 2022. Human organoids: New strategies and methods for analyzing human development and disease. Cell 185: 2756-69

[0244] Cunningham F, Achuthan P, Akanni W, Allen J, Amode MR, et al. 2019. Ensembl 2019. Nucleic Acids Res 47: D745-D51

[0245] Delaunay B. 1934. Sur la sphere vide. A la memoire de Georges Voronoi. Rlseecmun POCCUUCKOU QKadeMuu HayK. Cepun MameMamu^ecKan 793-800

[0246] Dessimoz J, Bonnard C, Huelsken J, Grapin-Botton A. 2005. Pancreas-specific deletion of p-catenin reveals Wnt-dependent and Wnt-independent functions during development. Curr Biol 15: 1677-83

[0247] Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, et al. 2013. STAR: ultrafast universal RNA-seq aligner. Bioinformatics B 15-21

[0248] Doble BW, Patel S, Wood GA, Kockeritz LK, Woodgett JR. 2007. Functional redundancy of GSK-3alpha and GSK-3beta in Wnt / beta-catenin signaling shown by using an allelic series of embryonic stem cell lines. Dev Cell 12: 957-71

[0249] Ebrahim N, Shakirova K, Dashinimaev E. 2022. PDX1 is the cornerstone of pancreatic -cell functions and identity. Front Mol Biosci 9

[0250] Elghazi L, Cras-Meneur C, Czernichow P, Scharfmann R. 2002. Role for FGFR2lllb-mediated signals in controlling pancreatic endocrine progenitor cell proliferation. P Natl Acad Sci USA 99: 3884-89 Fujitani Y, Fujitani S, Boyer DF, Gannon M, Kawaguchi Y, et al. 2006. Targeted deletion of a cis- regulatory region reveals differential gene dosage requirements for Pdxl in foregut organ differentiation and pancreas formation. Genes Dev 20: 253-66

[0251] Funa NS, Mjoseng HK, de Lichtenberg KH, Raineri S, Esen D, et al. 2024. TGF-beta modulates cell fate in human ES cell-derived foregut endoderm by inhibiting Wnt and BMP signaling. Stem Cell Rep 19: 973-92

[0252] Goncalves CA, Larsen M, Jung S, Stratmann J, Nakamura A, et al. 2021. A 3D system to model human pancreas development and its reference single-cell transcriptome atlas identify signaling pathways required for progenitor expansion. Nat Common 12: 3144

[0253] Grapin-Botton A, Kim YH. 2022. Pancreas organoid models of development and regeneration. Development 149

[0254] Greggio C, De Franceschi F, Figueiredo-Larsen M, Gobaa S, Ranga A, et al. 2013. Artificial three- dimensional niches deconstruct pancreas development in vitro. Development 140: 4452-62

[0255] Heiser PW, Lau J, Taketo MM, Herrera PL, Hebrok M. 2006. Stabilization of p-catenin impacts pancreas growth. Development 133: 2023-32

[0256] Heller RS, Dichmann DS, Jensen J, Miller C, Wong G, et al. 2002. Expression patterns of Wnts, Frizzleds, sFRPs, and misexpression in transgenic mice suggesting a role for Wnts in pancreas and foregut pattern formation. Dev Dyn 225: 260-70

[0257] Hohwieler M, 11 ling A, Hermann PC, Mayer T, Stockmann M, et al. 2017. Human pluripotent stem cell- derived acinar / ductal organoids generate human pancreas upon orthotopic transplantation and allow disease modelling. Gut 66: 473-86 Horb ME, Shen CN, Tosh D, Slack JMW. 2003. Experimental conversion of liver to pancreas. Curr Biol 13: 105-15

[0258] Houbracken I, de Waele E, Lardon J, Ling Z, Heimberg H, et al. 2011. Lineage tracing evidence for transdifferentiation of acinar to duct cells and plasticity of human pancreas. Gastroenterology 141: 731-41, 41 el-4

[0259] Huang L, Desai R, Conrad DN, Leite NC, Akshinthala D, et al. 2021. Commitment and oncogene-induced plasticity of human stem cell-derived pancreatic acinar and ductal organoids. Cell Stem Cell 28: 1090-+

[0260] Huang L, Holtzinger A, Jagan I, BeGora M, Lohse I, et al. 2015. Ductal pancreatic cancer modeling and drug screening using human pluripotent stem cell- and patient-derived tumor organoids. Nat Med 21: 1364-71

[0261] Jennings RE, Berry AA, Strutt JP, Gerrard DT, Hanley NA. 2015. Human pancreas development. Development 142: 3126-37

[0262] Keshara R, Kim YH, Grapin-Botton A. 2022. Organoid Imaging: Seeing Development and Function. Anna Rev Cell Dev Biol 38: 447-66

[0263] Kobberup S, Schmerr M, Dang ML, Nyeng P, Jensen JN, et al. 2010. Conditional control of the differentiation competence of pancreatic endocrine and ductal cells by FgflO. Meeh Develop 127: 220-34

[0264] Lampart FL, Iber D, Doumpas N. 2023. Organoids in high-throughput and high-content screenings. Front Chem Eng 5

[0265] Larraguibel J, Weiss AR, Pasula DJ, Dhaliwal RS, Kondra R, Van Raay TJ. 2015. Wnt ligand-dependent activation of the negative feedback regulator Nkdl. Mol Biol Cell 26: 2375-84

[0266] Larsen HL, Martin-Coll L, Nielsen AV, Wright CVE, Trusina A, et al. 2017. Stochastic priming and spatial cues orchestrate heterogeneous clonal contribution to mouse pancreas organogenesis. Nature Communications 8

[0267] Lewis A, Keshara R, Kim YH, Grapin-Botton A. 2021. Self-organization of organoids from endoderm- derived cells. J Mol Med (Berl) 99: 449-62

[0268] Love Ml, Huber W, Anders S. 2014. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 15

[0269] Lukonin I, Serra D, Challet Meylan L, Volkmann K, Baaten J, et al. 2020. Phenotypic landscape of intestinal organoid regeneration. Nature 586: 275-80

[0270] Lukonin I, Zinner M, Liberali P. 2021. Organoids in image-based phenotypic chemical screens. Exp Mol Med 53: 1495-502

[0271] Mclnnes L, Healy J, Melville J. 2018. Umap: Uniform manifold approximation and projection for dimension reduction. arXiv preprint arXiv:1802.03426

[0272] Mclin VA, Rankin SA, Zorn AM. 2007. Repression of Wnt / p-catenin signaling in the anterior endoderm is essential for liver and pancreas development. Development 134: 2207-17

[0273] Mead BE, Hattori K, Levy L, Imada S, Goto N, et al. 2022. Screening for modulators of the cellular composition of gut epithelia via organoid models of intestinal stem cell differentiation. Nat Biomed Eng 6: 476-94

[0274] Merz S, Breunig M, Melzer MK, Heller S, Wiedenmann S, et al. 2023. Single-cell profiling of GP2- enriched pancreatic progenitors to simultaneously create acinar, ductal, and endocrine organoids. Theranostics 13: 1949-73 Miyatsuka T, Kaneto H, Shiraiwa T, Matsuoka TA, Yamamoto K, et al. 2006. Persistent expression of PDX-1 in the pancreas causes acinar-to-ductal metaplasia through Stat3 activation. Gene Dev 20: 1435-40

[0275] Murtaugh LC, Law AC, Dor Y, Melton DA. 2005. Beta-catenin is essential for pancreatic acinar but not islet development. Development 132: 4663-74

[0276] Nadauld LD, Sandoval IT, Chidester S, Yost HJ, Jones DA. 2004. Adenomatous polyposis coli control of retinoic acid biosynthesis is critical for zebrafish intestinal development and differentiation. Journal of Biological Chemistry 7B 51581-89

[0277] Offield MF, Jetton TL, Labosky PA, Ray M, Stein RW, et al. 1996. PDX-1 is required for pancreatic outgrowth and differentiation of the rostral duodenum. Development 122: 983-95

[0278] Papadopoulou S, Edlund H. 2005. Attenuated Wnt signaling perturbs pancreatic growth but not pancreatic function. Diabetes 54: 2844-51

[0279] Pedersen AH, Heller RS. 2005. A possible role for the canonical Wnt pathway in endocrine cell development in chicks. Biochem Bioph Res Co 333: 961-68

[0280] Petersen MBK, Azad A, Ingvorsen C, Hess K, Hansson M, et al. 2017. Single-Cell Gene Expression Analysis of a Human ESC Model of Pancreatic Endocrine Development Reveals Different Paths to beta-Cell Differentiation. Stem Cell Rep 9: 1246-61

[0281] Rezania A, Bruin JE, Arora P, Rubin A, Batushansky I, et al. 2014. Reversal of diabetes with insulinproducing cells derived in vitro from human pluripotent stem cells. Nat Biotechnol 32: 1121- 33

[0282] Ringel T, Frey N, Ringnalda F, Janjuha S, Cherkaoui S, et al. 2020. Genome-Scale CRISPR Screening in Human Intestinal Organoids Identifies Drivers of TGF-P Resistance. Cell Stem Cell 26: 431-+

[0283] Scheibner K, Bakhti M, Bastidas-Ponce A, Lickert H. 2019. Wnt signaling: implications in endoderm development and pancreas organogenesis. Curr Opin Cell Biol 61: 48-55

[0284] Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, et al. 2012. Fiji: an open-source platform for biological-image analysis. Nat Methods 9: 676-82

[0285] Schmidt U, Weigert M, Broaddus C, Myers G. 2018. Cell Detection with Star-Convex Polygons. Leet Notes ComputSc 11071: 265-73

[0286] Sharon N, Vanderhooft J, Straubhaar J, Mueller J, Chawla R, et al. 2019. Wnt Signaling Separates the Progenitor and Endocrine Compartments during Pancreas Development. Cell Rep U 2281-91 e5

[0287] Shen CN, Horb ME, Slack JMW, Tosh D. 2003. Transdifferentiation of pancreas to liver. Meeh Develop 120: 107-16

[0288] Stoter M, Janosch A, Barsacchi R, Bickle M. 2019. Cell Profiler and KNIME: Open-Source Tools for High- Content Screening. Methods Mol Biol 1953: 43-60

[0289] Suppinger S, Zinner M, Aizarani N, Lukonin I, Ortiz R, et al. 2023. Multimodal characterization of murine gastruloid development. Cell Stem Cell 30: 867-84 ell

[0290] Ungricht R, Guibbal L, Lasbennes MC, Orsini V, Beibel M, et al. 2022. Genome-wide screening in human kidney organoids identifies developmental and disease-related aspects of nephrogenesis. Cell Stem Cell B-. 160-+ von Luxburg U. 2007. A tutorial on spectral clustering. Stat Comput 17: 395-416

[0291] Wells JM, Esni F, Boivin GP, Aronow BJ, Stuart W, et al. 2007. Wnt / p-catenin signaling is required for development of the exocrine pancreas. Bmc Developmental Biology 7

[0292] Willert K, Brown JD, Danenberg E, Duncan AW, Weissman IL, et al. 2003. Wnt proteins are lipid- modified and can act as stem cell growth factors. Nature 423: 448-52 Zhao L, Chen Dowbaj AM, Sljukic A, Bratlie K, et al. 2022. Organoids. Nat Rev Methods Primers 2

[0293] Zhou Q, Law AC, Rajagopal J, Anderson WJ, Gray PA, Melton DA. 2007. A multipotent progenitor domain guides pancreatic organogenesis. Developmental Cell 13: 103-14

[0294] Zhu Z, Verma N, Gonzalez F, Shi ZD, Huangfu D. 2015. A CRISPR / Cas-Mediated Selection-free Knockin Strategy in Human Embryonic Stem Cells. Stem Cell Rep 4: 1103-11

Claims

CLAIMS1. A method for producing cells having an acinar cell phenotype comprising(a) exposing pancreatic progenitor cells derived from human pluripotent stem cells to Wnt pathway activator in the absence of fibroblast growth factor (FGF), or, if the pancreatic progenitor cells are cultured in a medium comprising FGF, in the presence of an FGF inhibitor, thereby differentiating said pancreatic progenitor cells into cells having an acinar cell phenotype.

2. The method of claim 1, wherein the Wnt pathway activator is a canonical Wnt pathway activator, more preferably an agonist of Frizzled or LRP5 and / or LRP6, even more preferably a Wnt activating the canonical Wnt pathway acting via GSK3, Norrin, the peptide p-catenin pathway agonist PG-008 or a GSK-inhibitor and is most preferably a GSK-inhibitor.

3. The method of claim 2, wherein the GSK3-inhibitor is CHIR99021 (6-[[2-[[4-(2,4-Dichlorophenyl)- 5-(5-methyl-lH-imidazol-2-yl)-2-pyrimidinyl]amino]ethyl]amino]-3-pyridinecarbonitrile) (CAS No.: 252917-06-9), TWS119 (CAS No.: 601514-19-6), is preferably TWS119 and CHIR99021, and is most preferably CHIR99021, wherein CHIR99021 is preferably present in step (a) at a concentration of about 0.6 pM to about 7.5 pM, preferably about 1 pM to about 5 pM, more preferably about 2 pM to about 4 pM and most preferably about 3 pM.

4. The method of any one of claims 1 to 3, wherein the FGF inhibitor is an FGF receptor inhibitor.

5. The method of claim 4, wherein the FGF receptor inhibitor is SU5402 (CAS No.: 215543-92-3) or Infigratinib (CAS No. 872511-34-7), wherein SU5402 is preferably at a concentration in step (a) of about 5 pM to about 15 pM, more preferably at about 8 pM to about 12 pM and most preferably at about 10 pM, or wherein Infigratinib is preferably at a concentration in step (a) of about 0.5 pM to about 100 nM.

6. The method of any one of claims 1 to 5, wherein step (a) is carried out for at least 5 days, preferably at least 6 days and most preferably at least 7 days, and / or up to 30 days, preferably up to 20 days and most preferably up to 10 days.

7. The method of any one of claims 1 to 6, wherein in step (a) the pancreatic progenitor cells are in or on an extracellular matrix.

8. The method of claim 7 , wherein the extracellular matrix (i) comprises one or more and preferably all of collagen (preferably collagen type IV), entactin, perlecan (preferably heparan sulfate proteoglycan), fibronectin and laminin; and / or (ii) is a reconstituted basement membrane derived from extracts of mammalian cells, preferably mouse cells, and most preferably Engelbreth-Holm-Swarm mouse tumor cells.

9. The method of claim 7 or 8 further comprising before step (a)(a') adding the pancreatic progenitor cells, preferably in the form of pancreatic progenitor organoids into the extracellular matrix.

10. The method of any one of claims 1 to 9, further comprising after step (a)(b) isolating the cells having an acinar cell phenotype, wherein the cells having an acinar cell phenotype are preferably in the form of organoids.

11. The method of any one of claims 1 to 10, further comprising before step (a) and if present after step (a')(a") culturing the pancreatic progenitor cells in a medium comprising (a) FGF, preferably FGF2, and / or (b) a Rock inhibitor, preferably Y27632 (Cas No. 146986-50-7) and / or (c) insulin, preferably in the form of B27 medium supplement.

12. The method of any one of claims 1 to 11, further comprising before step (a) and if present also before steps (a') and (a")(al) differentiating human pluripotent stem cells into pancreatic progenitor cells, and (a2) optionally expanding and / or maintaining the pancreatic progenitor cells, wherein pancreatic progenitor cells are preferably in the form of pancreatic progenitor organoids.

13. The method of any one of claims 1 to 12, wherein the cells having an acinar cell phenotype are characterized by the expression of AMY, CPA2, PRSSl / Trypsin, MISTI, CEL, GATA6, UEA1, REGIA and SPINK1 and / or the reduced expression of progenitor markers, preferably the progenitor markers PDX1 and / or NKX6-1 and / or the reduced expression of the ductal markers SOX9, CFTR, KRT19 and endocrine progenitor marker CHGA.

14. The method of any one of claims 1 to 13, wherein the cells having an acinar cell phenotype are characterized by zymogen granules and / or cellular organization into acini and / or digestive enzyme production, wherein the digestive enzymes are one or more of amylase, trypsin and carboxypeptidase.

15. A cell having an acinar cell phenotype obtained by the method according to any one of claims 1 to 14.

Images