Improved Differentiation Method

20210047618 ยท 2021-02-18

Assignee

Inventors

Cpc classification

International classification

Abstract

The invention relates to methods and media for differentiating cells, for example for obtaining enteroendocrine cells, and to uses of the cells and organoids obtained by said methods. The invention also relates to methods for modulating hormone expression in enteroendocrine cells and medical uses relating to such methods.

Claims

1. A method for differentiating progenitor cells, wherein said method comprises: culturing the cells in a differentiation medium comprising a basal medium and further comprising one or more EGFR pathway inhibitors, a Notch inhibitor and one or more Wnt inhibitors.

2. The method of claim 1, wherein: (a) the one or more EGFR pathway inhibitors: (i) is selected from: (1) an EGFR inhibitor, (2) an EGFR and ErbB2 inhibitor, (3) an inhibitor of the RAS-RAF-MAPK pathway, (4) an inhibitor of the PI3K/AKT pathway and (5) an inhibitor of the JAK/STAT pathway; and/or (ii) comprises: A. an EGFR inhibitor, such as Gefitinib; B. an EGFR and ErbB-2 inhibitor, such as Afatinib; and/or C. an inhibitor of the RAS-RAF-MAPK pathway, e.g. a MEK inhibitor, such as PD0325901, and/or an ERK inhibitor, such as SCH772984; (b) the Notch inhibitor is a gamma-secretase inhibitor, optionally DAPT or dibenzazepine (DBZ) or benzodiazepine (BZ) or LY-411575; and/or (c) the one or more Wnt inhibitors is selected from: (1) an inhibitor of Wnt secretion, (2) a competitive or non-competitive inhibitor of the interaction between Wnt or Rspondin and the Wnt receptor complex, (3) an inhibitor that promotes the degradation of components of the Wnt receptor complex, (4) an inhibitor of Dishevelled family proteins, (5) an activator that promotes destruction complex activity, (6) an inhibitor of the deoligomerisation of the destruction complex and/or (7) an inhibitor of -catenin target gene expression, optionally wherein the one or more Wnt inhibitors comprise an inhibitor of Wnt secretion, e.g. a Porc inhibitor selected from IWP 2, LGK974 and IWP 1.

3.-9. (canceled)

10. The method of claim 1, wherein the differentiation medium further comprises one or more components selected from the group consisting of: a p38 inhibitor, a TGF-beta inhibitor, gastrin, a glucocorticoid, a receptor tyrosine kinase ligand, a BMP pathway activator, a cAMP pathway activator, a Hedgehog activator, a Hedgehog inhibitor, a modulator of mTOR signalling, B27 and N2; or one or more components selected from the group consisting of: a p38 inhibitor, a TGF-beta inhibitor, gastrin, a glucocorticoid, a receptor tyrosine kinase ligand, a BMP inhibitor, a cAMP pathway activator, a Hedgehog activator, a Hedgehog inhibitor, a modulator of mTOR signalling, B27 and N2; or a BMP activator, such as BMP7, BMP4 or BMP2; or a BMP inhibitor, such as noggin, sclerostin, chordin, CTGF, follistatin, gremlin, tsg, sog, LDN193189 or dorsomorphin optionally, wherein the differentiation medium comprises less than 1 mM EGF.

11. A differentiation medium comprising a basal medium and further comprising one or more EGFR pathway inhibitors, a Notch inhibitor and one or more Wnt inhibitors.

12. A method for differentiating intestinal progenitor cells to obtain a population of intestinal cells enriched in enteroendocrine cells, wherein said method comprises: culturing the intestinal progenitor cells in a differentiation medium according to claim 11.

13. A method for culturing epithelial stem cells, wherein said method comprises: culturing the epithelial stem cells in the presence of an expansion medium for epithelial stem cells to obtain expanded epithelial stem cells; and subsequently culturing the one or more expanded cells in a differentiation medium according to claim 11.

14. The method of claim 1, wherein: a) the cells are cultured in contact with an extracellular matrix; b) the method further comprises obtaining and/or isolating a differentiated cell population or a differentiated organoid; c) the progenitor cells are epithelial cells, for example, from the intestine, stomach, pancreas, liver, prostate, lung, breast, ovary, salivary gland, hair follicle, skin, oesophagus, bladder, ear or thyroid; d) the progenitor cells are from the intestine, stomach, pancreas or lung; and/or e) the progenitor cells are mammalian progenitor cells, for example, human progenitor cells.

15.-18. (canceled)

19. A method for culturing intestine epithelial stem cells, preferably to obtain a differentiated intestine organoid, and wherein said method comprises: culturing one or more intestine epithelial stem cells in contact with an extracellular matrix in the presence of an expansion medium; preferably wherein the expansion medium comprises a basal medium, and further comprises: a receptor tyrosine kinase ligand, a BMP inhibitor and a Wnt agonist and, optionally, valproic acid and a GSK-3 inhibitor (e.g. CHIR99021); and subsequently culturing the one or more expanded intestine epithelial stem cells in contact with an extracellular matrix in the presence of a differentiation medium according to claim 11.

20. An organoid obtainable or obtained by a method of claim 1.

21. The organoid of claim 20, wherein the organoid is derived from the liver, pancreas, intestine, stomach, prostate, lung, breast, ovarian, salivary gland, hair follicle, skin, oesophagus, bladder, ear or thyroid, preferably from the intestine, stomach, pancreas or lung.

22. An organoid according to claim 20 or 21, wherein the organoid is derived from: a) a human, and in which at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90% or at least 99% of the cells express enteroendocrine cell markers; or b) a mouse, and in which at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90% or at least 99% of the cells express enteroendocrine cell markers; optionally wherein the enteroendocrine cell markers are selected from Chga, Chgb, Tac1, Tph1, Gip, Fabp5, Ghr1, Pyy, Nts, Neurod1, Sst, Sct, cholecystokinin, glucagon and/or pro glucagon.

23. (canceled)

24. (canceled)

25. A composition comprising an organoid according to claim 20, and a differentiation medium comprising a basal medium and further comprising one or more EGFR pathway inhibitors, a Notch inhibitor and one or more Wnt inhibitors.

26. (canceled)

27. A method of treating a disorder, condition or disease comprising transplanting the organoid according to claim 20, or a cell derived from said organoid into a patient.

28. (canceled)

29. A pharmaceutical formulation comprising one or more EGFR pathway inhibitors, a Notch inhibitor and one or more Wnt inhibitors.

30. A method for screening for a therapeutic or prophylactic pharmaceutical drug or cosmetic, wherein the method comprises: contacting a differentiated organoid according to claim 20 with a candidate molecule (or a library of candidate molecules), evaluating said organoid for any effects (e.g. any change in the cell, such as a reduction in or loss of proliferation, a morphological change and/or cell death) or a change in organoid (e.g. the organoid size or motility); identifying the candidate molecule that causes said effects as a potential drug or cosmetic; and optionally preparing said candidate molecule as pharmaceutical or cosmetic.

31. A method for inducing Lgr5+ stem cell quiescence, wherein said method comprises: treating the cells with one or more EGFR pathway inhibitors.

32. The method of claim 31, wherein the cells have continued Lgr5 expression, e.g. as assessed by FACS, and/or Wnt signaling, e.g. as assessed by pTOPFLASH and pFOPFLASH Tcf luciferase reporter constructs; optionally wherein: a) quiescence is indicated by loss of KI67 expression; and/or b) the cells are treated with one or more EGFR pathway inhibitors for at least one week.

33. (canceled)

34. (canceled)

35. A quiescent stem cell population obtained by the method of claim 31, wherein the cells express Lgr5 and Lef1 and do not express KI367 and M phase marker phospho-histone H3.

36. A method of obtaining a population of cells enriched in EECs, wherein the method comprises culturing a population of cells in a differentiation medium according to claim 11.

37. The method of claim 36, wherein: a) the population of cells is enriched in GLP1-secreting EECs, and wherein the differentiation medium comprises a BMP inhibitor; or b) the population of cells is enriched in secretin-secreting EECs, wherein the differentiation medium comprises a BMP pathway activator.

38. (canceled)

39. A method of treating or preventing diabetes mellitus or an associated disease or disorder, wherein the method comprises administering a therapeutically effective amount of a BMP inhibitor to a subject in need thereof.

40. A method of treating hyperchlorhydria or obesity, wherein the method comprises administering a therapeutically effective amount of a BMP activator to a subject in need thereof.

Description

DESCRIPTION OF THE DRAWINGS

[0597] FIG. 1. EGFR Inhibition induces cell cycle exit in intestinal organoids. (A) Experimental setup. Organoids were treated with either EGFR (or MEK/ERK inhibitors) or DMSO one week after plating in BME. Samples were collected 1 (d1), 2 (d2), 4 (d4) or 7 (d7) after the treatment. The procedure was repeated from the treatment step onwards for replating experiments. (B) Intestinal organoids after 4 days in control (ENR) or EGFR inhibition (EGFRi). Lgr5.sup.GFPiresCreER fluorescence increases following EGFRi treatment. RFP channel is used to display the background. Lower panels are brightfield images. (C) Analysis of the cell cycle of intestinal organoids. EdU was administered 1 h prior to the sacrifice. Control (ENR) organoids continuously incorporate EdU (upper panels) and express KI67 (lower panels), while EGFRi treated organoids exit the cell cycle over time. (D) Quantification of C. (E) HOECHST analysis of the DNA content of control or EGFRi treated organoids. Left-hand peak is EGFRi treated organoids and right-hand peak is control organoids in the ENR medium. Bars on the right show quantification of 3 independent experiments. The top bands of the bars represent cells in G.sub.2/M phase, the middle bands of the bars represent cells in S phase and the bottom bands of the bars represent cells in G.sub.0/G.sub.1 phase. (F) Analysis of the cell cycle of intestinal organoids following reintroduction of EGF in the culture medium. (G) Lgr5.sup.GFPiresCreER+ cells exit the cell cycle upon 4 days of EGFRi treatment. Phospho-histone H3 (pH3) staining is used to visualize M phase. The graph at the bottom shows the quantification. DAPI is used to visualize the nuclei. Scale bars=50 m.

[0598] FIG. 2. EGFR signalling induced cell cycle exit is mediated by MAPK signalling pathway. (A) Histological analysis of ERK phosphorylation following EGFR inhibition. A rapid loss in pERK is gradually over 24 h, but remains low over 48 hours (upper panels). This period coincides with the cell cycle exit of organoids (KI67 staining, lower panels). (B) Single inhibition of Mek or Erk as well as simultaneous inhibition of EGFR and ErbB-2 using Afatinib yields similar results to Gefitinib induced EGFR inhibition. EdU is added to the culture medium 1 hour before the sacrifice. Middle panels show endogenous GFP expression from the Lgr5GFPiresCreER allele. Lower panels are brightfield images. DAPI is used to visualize the nuclei. Scale bars=50

[0599] FIG. 3. Lineage tracing indicates that qLgr5+ cells are stem cells. (A) Experimental setup. Organoids were re-plated in BME either in ENR (control) or EGFRi medium 7 days after dissociation. 4 days after the treatment, recombination was induced with 4OH Tamoxifen (T) for 16 hours and EGF signalling was restored. Organoids were collected either 4 or 12 days after Tamoxifen induction, or re-plated in ENR for 2 rounds. (B) Recombined (YFP+) cells generated CHGA+ EECs, LYZ+ Paneth cells (left panels) or Tuft cells identified by their apical actin and acetylated tubulin dense bundles (right panels). Both active (upper panels) and quiescent (lower panels) Lgr5+ cells were traced. (C) Quantification of B. The bottom band in each bar is CBC, the band immediately above the bottom band is EEC, the next band above is Paneth, the next band above is Tuft and the top band in each bar is Rest. (D) Recombined (identified by X-Gal staining) active (upper left panel) and quiescent (lower left panel) Lgr5+ cells generate entire organoids displaying multipotency. Recombined Dclk1+ cells (right panels) did not expand in either condition. Scale bars=50 m.

[0600] FIG. 4. RNA sequencing identifies key molecular differences between qLgr5+ and aLgr5+ stem cells. (A) Hierarchical clustering of the whole transcript of sorted Lgr5+ cells using the Lgr5.sup.GFPDTR (Lgr5.sup.DTR), Lgr5.sup.GFPiresCreER (Lgr5.sup.GFP) and Tuft cells using the Dclk1.sup.GFPiresCreER (Dclk1) alleles from control (ENR) or EGFR inhibited (EGFRi) conditions. Control organoids were added as a reference. (B) Principal component analysis (PCA). (C) A volcano plot showing the comparison of active and quiescent Lgr5 cells. X axis shows adjusted p value (q value, in log 10) and y axis shows fold change (in log 2). Grey dots indicate genes with a false discovery rate (FDR) less than 0.01, black dots depict genes that are not significantly changed. (D) Heatmap displaying genes differentially regulated by EGFR inhibition in Lgr5+ cells. Colours indicate z valued of each row (gene). (E) Boxplots displaying normalized expression values of marker genes. (F) Heatmap displaying k-means clustering of Pearson correlation of the whole transcriptome of individual active and quiescent Lgr5+ cells.

[0601] FIG. 5. Derivation of a high purity EEC culture. (A) Marker analysis of EECs (CHGA) and Paneth cells (LYZ). Organoids were treated for 4 days with Notch inhibitor DAPT (D), inhibitior of Wnt secretion IWP-2 (I), Gefitinib (EGFRi) or a combination of these treatments. DMSO is used as a control. (B) Inhibition of Mek signalling (Meki) together with Wnt and Notch signalling pathways similarly increases EEC cell numbers. Dark-grey and light-grey V-shaped arrows point to areas of high SCT and high GIP expression, respectively, in the left half of the figure. Dark-grey and light-grey V-shaped arrows point to areas of high CCK and high SST expression, respectively, in the left half of the figure. Gastric inhibitory protein (GIP), Secretin (SCT), Somatostatin (SST) and cholecystokinin (CCK) positive cell numbers dramatically increase. (C) qPCR analysis of EEC related marker gene expression in organoids. EI: EGFRi and MI:Meki. Scale bars=50 Error bars indicate standard deviation.

[0602] FIG. 6. Single cell transcriptome profiling reveals heterogeneity among induced EECs. (A) Heatmap displaying k-means clustering of Pearson correlation of the whole transcriptome of individual live organoid cells from Meki and EGFRi experiments. Numbers indicates clusters. The colours code for Pearson correlation of the whole cellular transcriptome. (B) t-SNE map depicting individual cells and marker genes expressed by group of cells. (C) t-SNE maps displaying the log 2 transformed colour coded transcript counts of respective genes. (D) Heatmap displaying colour-coded transcript counts of respective genes in clusters 2, 5, 6 and 7.

[0603] FIG. 7. The role of niche signalling pathways in proliferation of Lgr5+ cells. (A) FACS plots showing the endogenous Lgr5.sup.GFPDTR fluorescence (upper panels, x axis) and KI67.sup.efluor660 immunostaining (lower panels, x axis). PL3 channel is used (y axis) to discriminate background. Gates shows positive cells with respect to wild type controls. (B) Brightfield (upper) and fluorescent (lower) images of control (ENR) and EGFR inhibited (EGFRi) organoids using the Lgr5.sup.GFPiresCreER and Rosa-TOP.sup.CFP alleles. RFP channel is used to discriminate background. (C) Marker gene analysis of Lgr5.sup.GFPiresCreER+ cells 4 days after EGFRi treatment. Light-grey and dark-grey V-shaped arrows point to areas of high CHGA and LYZ expression, respectively, in the left-hand image. In the middle and right-hand images, the light-grey V-shaped arrows point to areas of high Phalloidin expression. (D) qPCR analysis of marker gene expression in 4 days EGFRi treated organoids. (E) Relative number of marker protein positive cells following reintroduction of EGF signalling for 1 (d1), 3 (d3) or 5 (d5) days compared to respective controls (DMSO treated, same days in culture). (F) Cell cycle entry of organoids following EGF introduction after repeated cycles of EGF withdrawal (NR) or EGFRi treatment (2NR+EGFRi). (G) Quantification of F. Scale bars=50 Error bars indicate standard deviation. Key: ENR=EGF, Noggin and R-spondin1; -R=R-spondin1 withdrawal; -N=Noggin withdrawal; EGFRi=inhibition of EGFR signalling (using Gefitinib accompanied by withdrawal of EGF from the culture medium).

[0604] FIG. 8. Cellular composition of organoids after EGFRi treatment. (A) Alcian blue, PAS and Mucin2 (MUC2) staining on paraffin sections of control (ENR) or EGFRi treated organoids. (B) 3D reconstruction of whole organoids stained for Paneth cell (LYZ) and EEC (CHGA) markers. Both cell types are more concentrated in EGFRi treated organoids (lower panel) compared to controls. Graph on the right provides the quantification. (C) Tuft cell numbers, visualized using the endogenous fluorescence of the Dclk1.sup.GFPiresCreER allele, increase after EGFRi treatment (right panel) compared to the controls (left panel). GFP fluorescence does not overlap with EEC (CHGA) or Paneth cell markers (LYZ). Scale bars=50 m.

[0605] FIG. 9. Gene ontology (GO) term and single cell analysis of Lgr5+ cell signatures. (A) GO term analysis (using Revigo) of genes downregulated after EGFRi treatment in Lgr5 cells. Cell cycle and division related terms as well as small molecule biosynthesis (all related to the cell cycle progression) are significantly higher in active compared to quiescent Lgr5+ cells. x-axis indicates the p-value (log 10), y axis shows the size of the GO term set. (B) t-SNE map displaying the distribution of the sequenced Lgr5+ cells. (C) t-SNE maps displaying log 2 transformed color coded transcript levels of Lyz and Dclk1 mRNAs that are used to identify Paneth and Tuft cells, respectively.

[0606] FIG. 10. Initial analysis of the single cell sequencing of induced EEC organoids. (A) t-SNE map displaying the distribution of clusters identified by RaceID. (B) t-SNE map comparing the distribution of cells derived from IDEGFRi and IDMeki experiments. (C) t-SNE maps displaying log 2 transformed colour coded transcript levels of the Apoa1 mRNA that is used to identify enterocytes.

[0607] Summary of intestinal progenitor cell fates. An intestinal progenitor cell can differentiate into a number of cell types, such as an enterocyte, goblet cell, Paneth cell or enteroendocrine cell. It was previously known that Notch activation, coupled with Wnt inhibition promoted enterocyte differentiation. It was also known that Notch inhibition coupled with Wnt activation could promote Paneth cell differentiation. In addition, it was known that Wnt and Notch inhibition could promote Goblet cell differentiation. However, there was previously a lack of understanding regarding how to enhance enteroendocrine cell differentiation.

[0608] FIG. 11. Composition of EEC culture. Treatment with IDMI differentiation medium resulted in differentiation mainly biased towards enterochromaffin differentiation. The absolute increase in EEC per organoid is mainly enterochromaffin cells. GLP1, CCK, NTS, STT and GIP producing cells are present to a lesser extent.

[0609] FIG. 12. qPCR results showing messenger RNA levels in different EEC differentiation protocols. Bar graphs show fold mRNA expression of various EEC markers relative to a control. Activation of the BMP pathway selectively enhances secretin messenger RNA at the expense of TAC1 (marker of Enterochromaffin cells). BMP4 was used to activate the BMP pathway and was present at a concentration of 10 g/ml. Basic condition: IDMI (IWP2, DAPT and a MEK inhibitor). MHY1485 is an mTOR activator that was tested at a concentration of 5 Vismogenib is a Hedgehog inhibitor (specifically a Smoothened inhibitor) that was tested at a concentration of 5 M.

[0610] FIG. 13. Staining for Secretin and GIP in different Enteroendocrine differentiation protocols. Activation of the BMP pathway greatly enhances the number of S cells.

[0611] FIG. 14. Staining for Serotonin in different Enteroendocrine differentiation protocols. Activation of the BMP pathway decreases the number of Enterochromaffin cells.

[0612] FIG. 15. qPCR results showing log 2-change in messenger levels of different EEC markers in EEC differentiation protocol compared to standard differentiation medium. EEC differentiation medium greatly enhances the number of EECs in the human organoids, but not of Paneth cells.

[0613] FIG. 16. CHGA expression in different Enteroendocrine differentiation protocols.

[0614] FIG. 17. Triple inhibition of Notch, Wnt and MEK generates cultures enriched in EECs. 30-80% CHGA+ cells per organoid were observed.

[0615] FIG. 18. Hormone expression in EEC culture. All different subtypes of EEC are present in the IDMI cultures and express a single hormone. Some cells are CHGA but positive for hormone.

[0616] FIG. 19. Hormone secretion by EECs in culture. 2-day culture medium was collected for the forskolin results. The cultures were then induced with forskolin for 1 hr and the medium was collected for the +forskolin results.

[0617] FIG. 20. Differential localization of Tac1, GLP1 and Secretin immunoreactive cells in the intestinal crypt and villus A-C) PYY-GLP1 double positive cells are located in the crypt, whereas L-cells in the villus lose expression of GLP1 D-E) Enterochromaffin cells in the crypt that express Serotonin, co-express Tac1. Tac1 expression in the villus is lost, where Enterochromaffin cells start co-expressing Secretin.

[0618] FIG. 21. BMP signaling is a driver of the villus hormone signature. A) Enteroendocrine cell differentiation (Inhibition of Wnt, Notch and MAPK) medium in which BMP signaling is inhibited by the presence of Noggin, generates a hormone signature reminiscent of the crypt: Secretin is absent, and Serotonin+ EECs always co-express Tac1. Exclusion of Noggin and the introduction of BMP4 to this differentiation cocktail, defined as EEC BMP high, greatly reduces GLP1 numbers, as well as Tac1. Single Serotonin+ Enteroendocrine cells, as well as Secretin+ cells, appear in this medium. B) Quantification of A) C) Brightfield image of small intestinal organoids in either BMP low or high EEC differentiation medium. GCG-Venus reporter was used to follow GLP1 positive cells. Although the different differentiation protocols generate morphologically indistinguishable organoids, GLP1 expression is greatly reduced in the background of BMP activation.

[0619] FIG. 22. Inhibition of BMP signalling in vivo causes an expansion of the GLP1+ compartment, and suppresses Secretin expression. A) 60 hr treatment through oral gavage of mice with LDN193189 causes an expansion of GLP1 numbers. GLP1 is widely expressed by L-cells in the villus, which do not always co-express PYY. B) Secretin is greatly reduced in the villus of BMPR inhibited mice compared to control treated.

EXAMPLES

[0620] Materials and Methods

[0621] Organoid Culture

[0622] The basic culture medium (advanced Dulbecco's modified Eagle's medium/F12 supplemented with penicillin/streptomycin, 10 mM HEPES, Glutamax, B27 [Life Technologies, Carlsbad, Calif.] and 1 mM N-acetylcysteine [Sigma]) was supplemented with 50 ng/ml murine recombinant epidermal growth factor (EGF; Peprotech, Hamburg, Germany), R-spondin1 (conditioned medium, 5% final volume), and Noggin (conditioned medium, 5% final volume), called ENR medium. Conditioned media were produced using HEK293T cells stably transfected with HA-mouse Rspol-Fc (gift from Calvin Kuo, Stanford University) or after transient transfection with mouse Noggin-Fc expression vector. Advanced Dulbecco's modified Eagle's medium/F12 supplemented with penicillin/streptomycin, and Glutamax was conditioned for 1 week. Cells were plated in BME (Trevigen). For inhibition of EGF signalling, cells were treated with Gefitinib (5 M; Santa Cruz Biotechnology) and EGF was withdrawn from the medium. Wnt secretion was inhibited with IWP-2 (1.5 M; Sigma Aldrich) and Notch with DAPT (1 mM, Sigma Aldrich). All treatments were performed on 5 day post passage organoids. For EGFR reactivation experiments, organoids were replated in fresh BME and ENR medium to make sure EGFR inhibitor is washed away. For induction of Cre-ERT activity, organoids were treated 0/N with 4-hydroxy tamoxifen (1 uM). All control organoids were treated with similar concentrations of the compound dissolvent, dimethyl sulfoxide (DMSO). During treatments, cells were imaged using an EVOS microscope (Electron Microscopy Sciences).

[0623] For the induction of enteroendocrine differentiation, cells were either cultured in ENR or ENR plus Valproic acid and CHIR99021 (Yin et al. (2014) Nature methods 11:106-112). After 5 days of culture, medium was removed and organoids were washed with PBS. The cocktail for EEC differentiation included: IWP2(1.5 M; Sigma Aldrich), DAPT (1 mM, Sigma Aldrich) and MEK inhibitor PD0325901 (5 M; Sigma Aldrich).

[0624] Immunostainings

[0625] Whole organoids were collected by gently dissolving the matrigel in icecold PBS, and subsequently fixed overnight at 4 C. in 4% paraformaldehyde. Next, organoids were permeabilized and blocked in PBS containing 0.5% Triton X-100 and 2% normal donkey serum (Jackson ImunoResearch) for 30 minutes at room temperature. Organoids were incubated for 2 hours at room temperature in blocking buffer containing primary antibodies. Primary antibodies used were rabbit anti-Lysozyme (1:500; DAKO), goat anti-Chromogranin A (1:500; Santa Cruz), mouse anti-Ki67 (1:250; BD Pharmingen), rabbit anti-phospho-Histone 3 (pH3 Ser10, 1:1000; Millipore), mouse anti-Cytokeratin 20 (1:1000; Dako), goat anti-Cholestocystokin (sc-21617, 1:100; Santa Cruz), rabbit anti-Neurotensin (sc-20806, 1:100; Santa Cruz), goat anti-Secretin (sc-26630, 1:100; Santa Cruz), goat anti-Somatostatin (sc-7819, 1:100; Santa Cruz), rabbit anti-Gastric Inhibitory Protein (ab22624-50, 1:500, Abcam), rabbit anti-Glucagon like peptide 1 (ab22625, 1; 500, Abcam) and mouse anti-acetylated tubulin (1:100; Santa Cruz). Organoids were incubated with the corresponding secondary antibodies Alexa488, 568 and 647 conjugated anti-rabbit, anti-goat and anti-mouse (1:1000; Molecular Probes), in blocking buffer containing DAPI (1; 1000, Invitrogen), or with Phalloidin Texas Red (1:1000; Life technologies). EdU incorporation was visualized using the Click-iT Assay Kit (Thermo Fisher), after 1 hr pre-incubation with EdU (10 uM). LacZ staining was performed as previously described (Barker et al. (2007) Nature 449(7165):1003-7). Sections were embedded in Vectashield (Vector Labs) and imaged using a Sp5 and Sp8 confocal microscope (Leica). Image analysis was performed using ImageJ software.

[0626] FACS Sorting

[0627] For FACS analysis of LGR5 and KI67 expression, Lgr5.sup.GFPDTR organoids were first dissociated into single cells through mechanical disruption, after 15 minutes of Trypsin treatment at 37 C. (TrypLE Express; Life Technologies, Carlsbad, Calif.). Single cells were fixed on ice using 4% paraformaldehyde for 30 minutes, and washed 3 times in PBS. Cells were permeabilized in PBS containing 0.5% Triton X-100 for 30 minutes, and were stained with an eFluor-660 conjugated rat anti-KI67 (1:1000; eBioscience) antibody for 30 minutes on ice. For cell cycle analysis, cells were stained in 1 ug/ml Hoechst 33342 (ThermoFisher). Subsequently, stained cells were analyzed on a BD FACS Calibur (BD Biosciences).

[0628] For expression analysis, organoids were dissociated and immediately sorted on a BD FACS Aria (BD Biosciences). Cells were sorted as single cells in Trizol in a 96 well plate, or as bulk in Trizol in eppendorf tubes.

[0629] RNA Isolation

[0630] For RNA-sequencing, cells were sorted into Trizol (Life Technologies) and total RNA was isolated according to the manufacturer's instructions, with the following alterations. RNA was precipitated overnight at 20 C., with 2 ug glycogen (Life Technologies). No additional RNA isolation step was used for cells sorted into 384-wells. For quantitative PCR analysis, RNA was isolated from organoids using the RNAeasy kit (QIAGEN) as instructed in the manufacturer's protocol.

[0631] Quantitative PCR

[0632] PCR analysis was performed using the SYBR-Green and Bio-Rad systems as described (Munoz et al. (2012) The EMBO Journal 31:3079-3091). PCR reactions were performed in triplicate with a standard curve for every primer. Changes in expression was calculated using CFX manager software (Bio-Rad). Primers were designed using the NCBI primer design tool.

TABLE-US-00006 Gene name Senseoligo Antisenseoligo Ki67 CCAGCTGCCTGTAGTGTCAA TCTTGAGGCTCGCCTTGATG Ccnb2 GCCAAGAGCCATGTGACTATC CAGAGCTGGTACTTTGGTGTTC Lgr5 ACCCGCCAGTCTCCTACATC GCATCTAGGCGCAGGGATTG Atoh1 GCTGTGCAAGCTGAAGGG TCTTGTCGTTGTTGAAGG ChgA CAGCTCGTCCACTCTTTCCG CCTCTCGTCTCCTTGGAGGG Lyz GGAATGGATGGCTACCGTGG CATGCCACCCATGCTCGAAT Gob5 ACTAAGGTGGCCTACCTCCAA GGAGGTGACAGTCAAGGTGAGA Alpi AGGATCCATCTGTCCTTTGG ACGTTGTATGTCTTGGACAG Sct GACCCCAAGACACTCAGACG TTTTCTGTGTCCTGCTCGCT Glu CTTCCCAGAAGAAGTCGCCA GTGACTGGCACGAGATGTTG Cck GAAGAGCGGCGTATGTCTGT CCAGAAGGAGCTTTGCGGA Sst GACCTGCGACTAGACTGACC CCAGTTCCTGTTTCCCGGTG Gip AACTGTTGGCTAGGGGACAC TGATGAAAGTCCCCTCTGCG Nts TGCTGACCATCTTCCAGCTC GAATGTAGGGCCTTCTGGGT

[0633] Single Cell and Bulk Sequencing

[0634] RNA samples were prepared using a modified version of the CEL-seq protocol as described previously (Grun et al. (2015) Nature 525:251-255; Hashimshony et al. (2012) Cell reports 2:666-673). ERCC spike-in was added to the Trizol solution. RNA pellets were dissolved in primer mix and incubated for 2 minutes at 70 C. Cells sorted into 384-well were directly lysed at 65 C. for 5 minutes. cDNA libraries were sequenced on an Illumina NextSeq500 using 75-bp paired-end sequencing.

[0635] Analysis of RNA sequencing data. Paired-end reads were quantified as described before (Grun et al. (2015) Nature 525:251-255) with the following exceptions. Reads that did not align or aligned to multiple locations were discarded. For analysis of the bulk sequencing, UMIs were ignored; instead read counts for each transcript were determined by the number of reads that uniquely mapped to that transcript. This count was divided by the total number of reads that mapped to all transcripts and multiplied by one million to generate the reads-per-million (RPM) count. RPM was used in preference of RPKM because CEL-seq only allows 3 end sequencing. Differential gene expression was evaluated using the DESeq package in R platform (Anders and Huber (2010) Genome biology 11:R106). Cut-offs used were an adjusted p-value <0.1 and FDR <0.1 and at least 2-fold difference to the compared population. To prevent samples with no reads disabling ratiometric analysis, all 0 reads were converted into 0.1 reads prior to ratio calculation and log 2 conversion. Gene ontology analysis was performed using the Revigo (Supek et al. (2011) PloS one 6:e21800) and Gorilla (Eden et al. (2009) BMC bioinformatics 10:48) softwares. Single cell sequencing data was analyzed as described previously (Grun et al. (2015) Nature 525:251-255).

Example 1

[0636] To understand how Lgr5+ stem cells are kept in cycle, we manipulated key signalling pathways active in the crypt niche. Flow cytometry analysis of Lgr5.sup.GFPDTR organoids with antibodies against KI67, a marker of cycling cells in all cell cycle phases, confirmed that the majority of the Lgr5+ cells cycle (FIG. 7). We inhibited Wnt signalling using two independent methods; i) IWP-2 treatment inhibits Wnt3 secretion by Paneth cells, and ii) withdrawal of R-spondin1 from the culture medium results in loss of Frizzled receptors from the cell surface. R-spondin1 withdrawal immediately causes loss of Lgr5.sup.GFP expression (FIG. 7A). IWP treatment poses a milder Wnt inhibition that depends on dilution of ligands through proliferation. Lgr5.sup.GFP expression was gradually down-regulated while stem cells differentiated. Yet, the remaining Lgr5.sup.GFP cells maintained KI67 expression (63.52.8% versus 94.42.1% in control). Withdrawal of the BMP inhibitor Noggin or addition of the Notch inhibitor DAPT both induced a rapid decrease in GFP+ cell numbers, but did not affect proliferation of the remaining Lgr5 cells (82.31.4% in Noggin withdrawal, 45.110% in DAPT) (FIG. 7). Next, we inhibited EGFR signalling using Gefitinib accompanied by withdrawal of EGF from the culture medium (EGFRi treatment, FIG. 7). While Lgr5.sup.GFP expression persisted, the Lgr5 cells eventually lost KI67 expression (13.11.0%), indicating an exit from the cell cycle.

[0637] We further analysed the early events associated with EGFR inhibition (FIG. 1A). Despite extensive apoptosis of the villus compartments of the organoid, buds resembling crypt structures containing Lgr5+(Lgr5.sup.GFPiresCreER) cells persisted for up to a week (FIG. 1B). We corroborated these results using the Lgr5.sup.GFPDTR allele (Tian et al. (2011) Nature 478:255-259) confirming that CBCs survive in the absence of EGF signalling (FIG. 7). After 4 days of EGFRi treatment, Lgr5+ cells composed 44.40.8% (13.66.5% in control) of the organoids (FIG. 7). The TCF.sup.CFP (Rosa.sup.TCF-CFP) Wnt reporter allele (Serup et al. (2012) Disease models & mechanisms 5:956-966) confirmed that Wnt signals remained high in the non-proliferative Lgr5 cells (FIG. 7). The KI67 protein persisted for the first 24 h but was lost from 48 h onwards (FIGS. 1C and 1D). Using a short pulse of EdU as a measure of cells replicating their DNA, we found that EGFRi lead to rapid halt in DNA replication as early as 24 h, which persisted for at least a week (FIGS. 1C and 1D). Consistent with exit from the S phase and eventually from the cell cycle, labelling the DNA content of EGFRi treated organoids using Hoechst DNA staining confirmed that all cells were in G.sub.0/G.sub.1 phase (FIG. 1E). 4 days after EGFRi treatment, reconstitution of EGF signalling induced rapid cell cycle entry within 24 h (KI67.sup.+) and progression to the S phase within 48 h (EdU.sup.+) (FIGS. 1F and 7). Even after a second cycle of EGFRi treatment, quiescent Lgr5 cells re-entered the cell cycle (FIG. 7).

[0638] We then further refined the analysis of the non-dividing, EGFRi induced Lgr5+ cells. They lacked the cell cycle marker KI67 and the M phase marker pH3 and did not incorporate EdU, excluding that rare dividing cells persisted during EGFRi treatment (FIG. 1G). Lack of Lysozyme (LYZ) and ChromagraninA (CHGA) implied that the Lgr5+ cells were not differentiated Paneth cells or enteroendocrine cells (EECs), respectively (FIG. 7). Tuft cells (intestinal M-cells) are rare mechanosensory cells also involved in response to parasitic invasion (Howitt et al. (2016) Science 351:1329-1333). They can be distinguished by their typical apical actin bundles, marked by acetylated tubulin and Phalloidin (Hofer and Drenckhahn (1996) Histochemistry and cell biology 105:405-412). The vast majority of the Lgr5+ cells did not display Tuft cell morphology, andsimilarly-Tuft cells were mostly Lgr5(FIG. 7). Quantitative PCR analysis confirmed that the Lgr5 cells did not differentiate into enterocytes, Paneth, EEC, goblet or Tuft cells (FIG. 7).

[0639] Muc2 (Mucin 2) as well as PAS and Alcian blue staining to visualize mucous structures revealed a significant reduction in the number of goblet cells after EGFRi treatment (Figure S2). The majority of the dividing TA progenitors generate mature enterocytes. While we found an increase in the number of LYZ+ Paneth cells and CHGA+ enteroendocrine cells per bud 4 days after EGFRi treatment, the total number (per organoid) of either cell type was not significantly changed (FIG. 8). The total number of Tuft cells increased upon EGFRi treatment (Figure S2). We corroborated these results using the Dclk1 GFPiresCreER allele (Nakanishi et al. (2013), confirming that EGFRi treatment increased the absolute number of Dclk1+ Tuft cells 3.2-fold (11.36.6 in ENR, 35.88.8 in EGFRi) (FIG. 9). In brief, while EGFR inhibition affects the cell type composition in organoids, it drives Lgr5+ cells into quiescence without inducing differentiation into one of the intestinal lineages.

[0640] MAPK signalling is a major downstream target of EGFR signalling pathway and regulates cell cycle progression. MAPK kinase (Mek) phosphorylates MAPK (Erk) to induce its nuclear localization and activation. EGFRi treatment reduced ERK phosphorylation as early as after 1 h (FIG. 2A). However, we observed a gradual recovery in phospho-ERK (pERK) levels within 48 h, despite continuing quiescence (FIG. 2A). We asked whether Mek/Erk signalling is essential for cell cycle progression of intestinal stem cells using small inhibitors of either Mek (PD0325901; Meki) or Erk (SCH772984; Erki). Both inhibitors induced quiescence of Lgr5+ cells implying that the ERK pathway downstream of EGFR regulates proliferation of Lgr5+ cells (FIG. 2B). The use of Afatinib, which inhibits both EGFR and ErbB2, yielded similar results (FIG. 2B). We concluded that EGFR inhibition through MAPK signalling is sufficient to induce a reversible quiescent state in intestinal organoid stem cells.

[0641] To test whether quiescent Lgr5 cells retain stem cell potential, we used Lgr5.sup.GFPiresCreERRosa26.sup.YFP/LacZ and Lgr5.sup.GFPiresCreERRosa26.sup.tdTomato organoids to follow the fate of Lgr5+ cells (FIG. 3A). CreER induction using 4-OH Tamoxifen (Tmx) led to rapid recombination that could be visualized by YFP (or tdTomato) fluorescence or LacZ expression (FIGS. 3B and 3C). Labelled, dividing Lgr5 cells generated quiescent Lgr5 cells upon EGFRi treatment, indicating that the latter indeed derived from active Lgr5 cells (FIG. 3B). Moreover, the fraction of labelled Paneth cells and EECs increased upon EGFRi treatment (FIGS. 3B and 3C). In non-EGFRi treated controls, recombined cells generated entire organoids upon passage, indicating efficient labelling of stem cells (FIG. 3D). Upon passage and reactivation of EGF signalling, the progeny of recombined quiescent Lgr5 cells persisted over several passages and generated organoids (FIG. 3D). These results indicated that quiescent Lgr5 cells induced upon EGFRi treatment behave as genuine stem cells.

[0642] Tuft cells are non-dividing and might display stem cell-like properties by contributing to tissue regeneration upon injury and tumour growth (Nakanishi et al. (2013) Nature Genetics 45:98-103). We followed the fate of Tuft cells using the organoids derived from the Dclk1.sup.GFPiresCreERRosa.sup.YFP/LacZ mouse model. Both in normal and EGFRi treated organoids, labelled cells remained as single cells over time (FIG. 3C). When passaged in the presence of EGF, labelled cells were lost and did not contribute to organoid generation (FIG. 3C) arguing against Dclk1.sup.+ Tuft cells having stem cell potential in an organoid setting.

[0643] To better understand the molecular characteristics of quiescent Lgr5 cells, we performed RNA sequencing on FACS-isolated active (DMSO control) and quiescent (EGFRi treatment, d4) Lgr5+ stem cells. We included both Lgr5.sup.GFPiresCreER (n=2) and Lgr5.sup.GFPDTR (n=2) alleles in our study. We also included sorted Dclk1.sup.GFP Tuft cells and bulk organoids for comparison. Hierarchical clustering and principal component analysis (PCA) revealed that Lgr5+ cells were more similar to each other than to whole organoids or to Tuft cells (FIGS. 4A and 4B). Consistent with differences in their cell cycle, active and quiescent Lgr5+ stem cells clustered separately (FIGS. 4A and 4B). Differential gene expression analysis between active and quiescent Lgr5 cells revealed 533 differentially regulated genes, 290 of which were enriched in quiescent Lgr5 cells (FDR <0.01 FIG. 5C). Transcriptional effectors of the ERK pathway (Etv4 [7.7, p-adj<0.001] and Etv5 [7.7, p-adj<0.001]) were down-regulated in quiescent Lgr5+ stem cells confirming efficient Erk inhibition (FIGS. 4C and 4D). Similarly, several cell cycle associated genes, such as Ccnb1 (2.1, p-adj<0.005) and Ccnb2 (1.9, p-adj<0.05) were decreased, consistent with the G0 arrest (FIGS. 4C and S9). GO analysis of genes downregulated upon EGFRi treatment confirmed a clear loss of cell cycle-associated genes (Figure S9). One of the defining transcripts in quiescent Lgr5 cells is Lef1 (not detected in active Lgr5, p-adj<0.001), a component of the Wnt signalling pathway. Consistent with our reporter expression, we observed a significant increase in some of the well-known Wnt target genes, including Rnf43 (2.3, p-adj<0.005) and Lgr5 (2, padj<0.05) (FIG. 4D). Quantitative PCR analysis in independent experiments confirmed these results (FIG. 9). We also noticed a strong increase of members of the AP-1 family of transcription factors (Junb, Fos, Fosb) in quiescent Lgr5 cells (FIG. 4D). Consistent with the histological analysis, Paneth cell, enterocyte and Goblet cell specific gene expression remained unchanged. Chga, expressed by EECs and their precursors, was 7.3-fold higher in quiescent compared to active Lgr5+ stem cells although high variation at low expression levels impeded further conclusions (p-val=0.019 and padj=0.28). Similarly, while Dclk1 (6, p-adj<0.05) and some other Tuft cell markers increased upon EGFRi treatment, their levels were significantly lower in quiescent Lgr5 cells than in Tuft cells (FIG. 4D). Global changes in gene expression might be either shared by all quiescent Lgr5 stem cells, or be the result of changes in a specific subpopulation. We have recently shown that individual active Lgr5 cells are rather homogenous (Grun et al. (2015) Nature 525:251-255). We performed single cell sequencing analysis of a total of 192 FACS purified single active or quiescent Lgr5 cells from control or EGFRi treated organoids. Using RaceID (Grun et al. (2015) Nature 525:251-255), we identified a single prominent population (cluster 1) containing cells from both control and EGFRi treated cells (FIG. 4F) containing both active and quiescent Lgr5+ cells (FIG. 9). Minor clusters 2 (four cells) and 4 (one cell) expressed Paneth (Lyz1) and Tuft (Dclk1) cell related genes, while cluster 3 (one cell) resembled cluster 1 (FIG. 9). Thus, the quiescent Lgr5 population is homogenous and despite some significant changes in gene expression, closely resembles quiescent Lgr5 cells.

Example 2

[0644] Quiescence combined with increased Chga and Lgr5 expression (as described in Example 1) is reminiscent of the label-retaining secretory precursors described by Winton and colleagues (Buczacki et al. (2013) Nature 495:65-69). These cells efficiently differentiate into EECs, suggesting that cell cycle exit of Lgr5 cells may facilitate generation of EECs. Hormones expressed by EECs regulate a wide variety of physiological responses, like gastric emptying, release of pancreatic enzymes, mood and glucose tolerance. They are also used to define subtypes (see introduction). G protein-coupled taste receptors have been identified as regulators of hormone secretion (Janssen and Depoortere (2013) TEM 24:92-100). EECs can have direct luminal contact and sense the content with microvilli. Other EECs, the so-called closed-type cells, do not have a luminal lining and need other sources to be stimulated (Janssen and Depoortere (2013) TEM 24:92-100). The length of their basal process also varies and may form synaptic contacts with enteric neurons to connect to the nervous system.

[0645] To establish a protocol for EEC differentiation, we used the quiescent Lgr5 cells as a starting point and modulated the Notch and Wnt signalling pathways, which are involved in secretory differentiation. Inhibition of Notch signalling by DAPT treatment (D) enhanced secretory differentiation, leading to a large increase in the number of LYZ+ Paneth cells (FIG. 5A). Inhibition of Wnt secretion using IWP-2 (I) in combination with DAPT abolished Paneth cell differentiation and induced EECs and Goblet cells (FIG. 5A). EGFRi treatment inhibited Goblet cell differentiation, but spared Paneth cells and EECs (FIG. 5A). Combined inhibition of EGFR/WNT/Notch pathways (IDEGFRi) resulted in a massive increase in EECs, while inhibiting enterocyte, Goblet and Paneth cell differentiation (FIG. 5A). Similarly, inhibition of Mek together with Wnt and Notch signalling pathways (IDMeki) increased CHGA+ cell numbers. The different EEC subtypes are rare in normal intestinal organoid cultures (FIG. 5B). IDMeki treatment resulted in a robust increase in the number of these EEC cell types. We used qPCR to corroborate these results. Expression of pan-EEC marker Chga was 25-fold higher in IDEGFRi and over 100-fold higher in IDMeki treated organoids (FIG. 5C). Coincidentally, expression of Sst (55), Gip (14), Sct (5), Cholecystokinin (15) and Glucagon (Gcg/Proglucagon, 4) mRNA were upregulated upon IDEGFRi treatment, following a similar trend to IDMeki (FIG. 5C). Nts was the sole hormone analyzed that was expressed at control levels. Thus, our protocol successfully generated high numbers of EECs and various subtypes expressing a panel of hormones regulating mammalian physiology (Egerod et al. (2012) Endocrinology 153:5782-5795).

[0646] To elucidate the cellular composition of EEC induced organoids and the extent of heterogeneity in hormonal expression of individual ECCs, we performed single-cell RNA sequencing analysis. We sorted live single cells (without additional markers) from IDEGFRi and IDMeki treated organoids. Among the 289 cells that passed our filtering we identified a cluster of 94 cells as enterocytes enriched in Aldob (4.9, p-adj<0.001), Apoa1 (12.6, p-adj<0.001) and Alpi (5.6, p-adj<0.001, FIG. 10). These were excluded from further analysis to better characterize remaining populations. Cells derived from both IDEGFRi or IDMeki treated organoids were distributed similarly in t-SNE space and were analyzed together (Figure S4).

[0647] Using RaceID, we identified 12 distinct clusters of cells (FIG. 6A). k-means clustering of the Pearson correlation of cellular transcriptomes revealed a clear separation between clusters as well as possible heterogeneity within clusters (e.g. 7 and 8, FIG. 6A). Differential gene expression analysis revealed signature genes for each cluster, which we used to classify cell types (Table 51).

[0648] The most prominent clusters were 3 (53 cells) and 4 (35 cells) that expressed pan-EEC markers Chga and Chgb (FIG. 6C). Chga and Reg4 expression formed a gradient, both being higher in Cluster 4. Hormonal production in these Chgb high clusters was best defined by Tac1 and Tph1 expression, both markers of Enterochromaffin cells. Tac1 encodes for the hormone Substance P while Tph1 encodes for the rate-limiting enzyme in Serotonin synthesis (Egerod et al. (2012) Endocrinology 153:5782-5795; Grun et al. (2015) Nature 525:251-255). Both may act as neurotransmitters exciting the connected enteric neurons (Latorre et al. (2016) Neurogastroenterology and motility: the official journal of the European Gastrointestinal Motility Society 28(5):620-630).

[0649] The other clusters had relatively low levels of Chga and Chgb transcripts but included cells expressing hormones (FIG. 6C). Cluster 2 (21 cells) was marked by Gip expression (74) that is expressed by K-cells. Fabp5 was also highly enriched in this cluster (12.6), consistent with its role in Gip secretion (Shibue et al. (2015) American journal of physiology, endocrinology and metabolism 308:E583-591). Cluster 5 (9 cells) members expressed very high levels of Sst (182) identifying them as D-cells (FIG. 6C). Ghrelin (Ghr1) expression was distributed to more than one cluster, but was highest in cluster 6 (19, 3 cells). We also noticed that Islet1 (Isl-1, 9.7) was coexpressed with Ghr1 in these cells. Islet1 plays an important role in cell fate specification and its loss leads to impaired glucose homeostasis (Terry et al. (2014) American journal of physiology, gastrointestinal and liver physiology 307:G979-991). Cells in cluster 7 (18 cells) all highly expressed Cck (55.7), and subgroups in this cluster were also rich in other hormones such as Gcg (28.2), Ghr1 (5.3) and Pyy (11.4). Consistently, I-cells are reported to co-express Cck with other hormones at varying levels (Egerod et al. (2012) Endocrinology 153:5782-5795).

[0650] One of the early inducers of EEC differentiation is Neurogenin-3 (Neurog3), which is followed by Neurod1. Neurog3 (5.2) expression was highest in cluster 9 (6 cells) and in some cells of cluster 3 that were most similar to cluster 9. Virtually all EEC clusters expressed Neurod1. Given the temporal expression of these transcription factors, we suggest that cluster 9 represents EEC progenitors, which then through Neurod1 generate a panel of EECs. Cluster 1 (18 cells) was enriched in Goblet and Paneth cell related genes, such as Agr2 (33), Muc2 (26), Ttf3 (23) and Defa24 (28). Even after filtering, some remaining enterocyte-like cells expressing Aldob and Mt1/2 were visible (Cluster 8, 7 cells). These clusters may have been generated prior to the EEC induction, as we do not see an increase in their numbers following EGFR or Mek inhibiton. Dclk1 and Trpm5 expression identified cluster 10 (15 cells) as Tuft cells (FIGS. 6B and 6C). In total, 145/289 cells (50% of all cells) analyzed were EECs or their progenitors, confirming high efficiency of our induction protocol.

[0651] Since multiple hormones can be co-expressed in the same cell, we further looked into the heterogeneity of hormone expression at a single cell level (FIG. 6D). We focused on cluster 2, 5, 6 and 7 where multiple hormones were expressed. 4 major groups based on Gip, Sst, Cck and Ghrelin expression were visible. Cck+ cells are further split into Gcg+Ghr1+, Gcg+Ghr1, Gcg-Ghr1+ and Gcg-Ghr1 clusters, with some also co-expressing low levels of Sst and/or Gip. Transcriptomes of Sst+ cells were more homogenous, coexpressed low levels of Gip and Cck while one cell co-expressed Ghr1 only. We have previously reported partial overlap between Cck+ and Tac1+ cells. Consistently, some of the Tac1+ cells in clusters 3 and 4 expressed low levels of Cck (FIGS. 6B and 6C).

[0652] Overall, our single cell analysis indicates that our protocol enriched for EECs to 50% of the culture based on marker gene expression. Moreover, we generated a panel of EEC subtypes including some rare cells with complex hormonal expression profiles.

Example 3

[0653] A murine EEC differentiation medium containing inhibitors of Wnt, Notch and MAPK signalling was used to generate organoids. These organoids contained a mixture of all EEC subtypes. Enterochromaffin cells that produce serotonin were the most abundant cell type in these cultures. The mechanism by which different EEC subtypes are generated was investigated. This was with a view to expand the uses of the culture and to understand development of these cells. In a first screen of signalling pathways that might control EEC subtype specification, modulations of the BMP, Hedgehog and mTOR pathways were found to affect the relative ratios of the EEC subtypes. Secretin-producing S cells seem to be to be particularly rare during Wnt, Notch and MAPK inhibition. S cells normally reside in the proximal duodenum, the same location as organoids in this study were isolated from. Activation of the BMP pathway by withdrawing Noggin from the EEC differentiation medium, and adding BMP4 (10 g/ml) to the culture medium changed the relative abundance of EEC subtypes. A drastic increase in the number of S-cells (on messenger, 400-fold over control), as well as secretin levels per cell (based on IHC), was observed after activation of the BMP pathway. This increase in S-cells seems to go at the expense of the number of enterochromaffin cells, suggesting potential overlapping developmental pathways. See further discussion of the role of BMP signaling in Example 5.

Example 4

[0654] Human small intestine (SI) organoids were cultured first in expansion medium (as described in Sato et al. (2011) Gastroenterology 141(5):1762-72) to increase stem cell number, and were then replated in a differentiation medium (expansion medium lacking Wnt conditioned medium, TGF-inhibitor, p38 inhibitor and Nicotinamide) to direct differentiation. Replating involved disruption of Matrigel (without dissociation of organoids) in cold medium, washing of the organoids with basal culture medium (PBS can also be used) and subsequent plating in fresh Matrigel (without dissociation of organoids). Organoids were then cultured in the differentiation medium for 1 day. On the next day, the EEC specific enteroendocrine differentiation protocol. This involved the addition of an EEC differentiation medium for 5 days, which was the same as standard differentiation medium (expansion medium lacking Wnt conditioned medium, TGF-inhibitor, p38 inhibitor and Nicotinamide) with the addition of 1.5 M IWP2, 10 mM DAPT, 100 nM MEKi PD0325901. Much lower levels of MAPK inhibitors were required for the differentiation of these human cells to an EEC fate in this experiment than were required for the differentiation of the murine cells to an EEC fate in Example 3 (100-500 nM versus 1-5 M in the murine system), possibly due to the absence of Paneth cells that produce EGF in the human system.

DISCUSSION

[0655] Here we identify EGF signalling as an indispensible driver of Lgr5 stem cell proliferation in organoids. Under conditions where Wnt signalling is untouched but EGF signalling is blocked, actively dividing Lgr5 stem cells convert into quiescent Lgr5 cells that retain expression of various Wnt target genes. This cellular state can be maintained for up to a week. Yet, the simple restoration of EGF signalling converts the quiescent cells back into their normal active stem cell state. Differential expression analysis revealed the loss of well-known proliferation-inducing transcription factors such as the Ets-like factors Etv4 and -5, suggesting that they play a key role in stem cell division.

[0656] Thus, chemical inhibition of EGFR signalling arrested Lgr5+ stem cells without affecting their stem cell potential. We believe that high-level Wnt signalling maintains this stem cell potential during induced quiescence. In organoids as well as in crypts, Lgr5+ cells are always the direct neighbors of the Wnt3-secreting Paneth cells (Sato et al. (2011) Nature 469:415-418). In this setting, Wnt3 does not diffuse over distances, but is loaded directly onto the Lgr5 stem cells (Farin et al. (2016) Nature 530:340-343). The quiescent Lgr5 stem cells remain juxtaposed to the Paneth cells in EGFRi treated organoids and are thus exposed to high local Wnt signals.

[0657] Indeed, 3 independent Wnt target gene alleles as well as gene expression analyses confirmed an increase in Wnt signalling upon EGFR inhibition. In sum, our results show that maintenance of stem cell fate requires Wnt but not EGF, whereas stem cell proliferation depends on the combination of Wnt and EGF.

[0658] Previous studies have identified quiescent cells close to the zone of differentiation at the +4 position with stem cell potential (Sangiorgi and Capecchi (2008) Nature Genetics 40:915-920; Takeda et al. (2011) Science 334:1420-1424; Yan et al. (2012) Proceedings of the National Academy of Sciences of the United States of America 109:466-471). We have reported the existence of Dll1+ precursors that exclusively generate secretory cells at this position (van Es et al. (2012) Nature cell biology 14:1099-1104). Using a histone label retention assay, Doug Winton's group identified a similar population with secretory differentiation potential. These label retaining cells share a signature with CBCs including the expression of Lgr5, but express good levels of some of the secretory lineage genes, such as Chga (Buczacki et al. (2013) Nature 495:65-69). Taken together, these secretory precursors represent transient states, yet can de-differentiate into stem cells when the need arises and can thus be considered facultative stem cells (Buczacki et al. (2013) Nature 495:65-69; van Es et al. (2012) Nature cell biology 14:1099-1104). A similar situation exists for the abundant enterocyte precursors in the crypt (Tetteh et al. (2016) Cell stem cell 18:203-213).

[0659] To our surprise, we noticed a slight bias of quiescent Lgr5 cells towards expression of EEC markers, such as Chga, which made them reminiscent of the Lgr5+ label-retaining cells identified by Doug Winton. EECs represent fewer than 1% of the cell of the intestinal epithelium, yet together form the largest endocrine organ in the human body in terms of cell number (Latorre et al. (2016) Neurogastroenterology and motility: the official journal of the European Gastrointestinal Motility Society 28(5):620-630). EECs have been proposed to act as sensors of luminal content such as nutrients and microorganisms, and regulate physiological responses like glucose sensitivity, gastric emptying, mood and appetite through the secretion of hormones (Janssen and Depoortere (2013) TEM 24:92-100). The functional study of EECs has been hampered by a lack of robust in vitro systems to generate these cells in large amounts. We have previously reported the directed differentiation of stem cells into secretory cells by Notch inhibition (van Es et al. (2005) Nature 435:959-963). Paneth cell formation requires active Wnt signalling combined with Notch inhibition (van Es et al. (2012) Nature cell biology 14:1099-1104; van Es et al. (2005) Nature 435:959-963), and combined inhibition of Wnt and Notch signalling generates mainly Goblet cells (van Es et al. (2005) Nature 435:959-963; Yin et al. (2014) Nature Methods 11:106-112). By combining inhibition of EGFR pathway with Wnt and Notch blockage, EECs of a diversity of types are efficiently formed in primary cell culture. EGFR signalling has been shown to be important for the production of Goblet cells (Heuberger et al. (2014) Proceedings of the National Academy of Sciences of the United States of America 111:3472-3477). We believe that simultaneous inhibition of enterocyte, Paneth and Goblet cell fate by inhibiting Notch, Wnt and EGFR signalling respectively, is the key to the generation of EECs.

[0660] Up to 20 subtypes of EECs can be distinguished based on the production of specific peptides. These cells exist at different frequencies in the proximal to distal GI tract. Certain EECs have been found to express several functionally related hormones both at transcript and protein level (Egerod et al. (2012) Endocrinology 153:5782-5795; Grun et al. (2015) Nature 525:251-255), although some of this might be an intermediate stage of maturation. We demonstrate the simultaneous generation of several principle subtypes of EECs, by single-cell transcriptome profiling. Confirming our previous work (Grun et al. (2015) Nature 525:251-255), we detect Chga-high andlow populations of EECs. The former is enriched in Tac1/Tph1 expressing EECs with some cells lowly expressing Cck. The latter is very diverse in hormonal expression. Cells with overall similar transcriptomes profiles cluster together, even though they can be subdivided into several classes based on hormonal expression. The wide distribution of hormonal expression levels suggests that separation between EEC subtypes is not clear-cut. It is tempting to speculate that their expression is at least in part stochastic and might even be temporally dynamic.

Example 5

[0661] Enteroendocrine Cells Alter Hormone Expression During Crypt to Villus Migration

[0662] Multiple Enteroendocrine cell hormones are differentially enriched in the intestinal crypt and villus. L-cells co-express GLP1 and PYY in the crypt, but mostly lack GLP1 expression in the villus (FIG. 20A-C). Enterochromaffin cells produce Serotonin along the whole crypt-villus axis, but selectively co-express Tac1 in the crypt and Secretin in the villus (FIG. 20D-E). This suggests that EECs can switch hormone expression during migration from the crypt to villus, and that there is a potential lineage relationship between crypt and villus EECs. However, an alternative explanation for this observation is that cells immunoreactive for Secretin or PYY in the villus are derived from an unrelated progenitor in the crypt, negative for Serotonin and Tac1 or GLP1. To show that such a transdifferentiation event can occur during the lifetime of EECs, we analyzed intestines from Tac1-Cre/Rosa26tdTomato mice. This reporter faithfully labels cells in the crypt that co-express Tac1 and Serotonin. Furthermore, Serotonin+ cells and the majority of Secretin+ cells in the villus are traced, while being negative for Tac1. Other hormones, including CCK, are not derived from a

[0663] Tac1/Serotonin+ progenitor cell. These data suggest that individual EECs can express different sets of hormones in the crypt and villus. The majority of Secretin-producing cells is part of the Enterochromaffin lineage. The capacity of EECs to switch hormone expression implies that there might be fewer unrelated differentiation pathways of EECs than previously anticipated, but that rather alterations in the types of hormones produced by a limited number of cells create a plethora of EEC states.

[0664] BMP Signaling Induces a Villus Hormone Signature in Enteroendocrine Cells

[0665] We next asked whether hormonal switches that occur from crypt to villus migration are part of a default maturation process in EECs, or that this is simply reflecting different niche signals where EECs are exposed to. Multiple morphogens are known to exist from crypt to villus, including ascending levels of BMP and Hedgehog, and descending levels of Wnt signals. The differentiation protocol for Enteroendocrine cells in the murine intestinal organoid system described in earlier examples is based primarily on the triple inhibition of the Wnt, MAPK and Notch signaling pathway. We used this differentiation medium and added stem cell factor Noggin to the differentiation cocktail, thus creating a BMP low environment. Strikingly, we observed that all Enterochromaffin cells in this culture always co-express Serotonin and Tac1, while Secretin expression was lacking. This culture thus reflects an EEC hormone signature reminiscent of the crypt state, implying that niche signals are dominant over temporally driven EEC maturation.

[0666] To address whether niche signals indeed might be orchestrating EEC hormone signatures, we used our previously defined EEC differentiation system as a starting point and modulated signaling pathway on the background of this cocktail. Vismogenib was added to inhibit Hedgehog signaling, Noggin was withdrawn to activate BMPR1a/BMPR2 signaling and ALK5/4/7 inhibitor A83 or TGF-beta1 added to modulate different TGF-beta family receptor pathways. Although Wnt is already limited in the EEC differentiation cocktail by Porcupine inhibitor IWP2, we removed R-spondin on top to get a more pronounced and rapid inhibition of Wnt. We used Sct and Gcg transcript levels as a proxy of the villus and crypt hormone signatures respectively, while ChgA was included as transcript that has constant expression from crypt to villus. Inhibition of Hedghehog signaling does not modulate any of the hormones assessed, consistent with its predominant role in the intestinal mesenchyme. A83 and TGF-beta1 both inhibited expression of all EEC hormones. R-spondin removal similarly impaired EEC differentiation of all hormones assessed. This is potentially related due to a rapid differentiation of stem cells into the absorptive lineage when Wnt signals are inhibited rather than choosing a secretory fate when losing Notch signals. Finally, withdrawal of Noggin during EEC differentiation induced expression of Sct while repressing Gcg transcription, consistent with a switch from a crypt to villus expression profile. We added exogenous BMP4 to this EEC differentiation cocktail to further amplify BMP signaling, and named this combination EEC BMPhigh compared to the EEC BMPlow medium as we have previously defined. Stimulating organoids with this EEC BMPhigh differentiation medium induces the production of cells immunoreactive for Secretin, as well as Enterochromaffin cells expressing Serotonin without Tac1. The GLP1+ cell numbers were greatly diminished in this regime (FIG. 21A-C), without obvious morphological changes in the organoids. Organoids established from different intestinal regions maintain their regional identity in terms of hormone signatures, with enrichment of GIP in the proximal SI and Pyy, Nts and Gcg in the distal SI. Other hormones or EEC markers that show homogenous distribution from crypt to villus, including ChgA, Tph1, CCK and duodenal GIP, are only mildly affected by different BMP levels. We do observe an upregulation of Pyy and Nts which are expressed highest in the villus. The effects of BMP are summarised in the table below. Collectively, these data point to a role of BMP signals in the induction of hormones that are associated with EECs in the villus.

TABLE-US-00007 EEC BMPhigh (BMP activator) EEC BMPlow (BMP inhibitor) Villus-like EEC characteristics Crypt-like EEC characteristics High secretin expression Low secretin expression Low Tac1 and Gcg expression High Tac1 and Gcg expression Low numbers of GLP1+ cells High numbers of GLP1+ cells

Example 6

[0667] LDN193189 (Selleckchem Catalog. No.S2618) at 35 mg/kg was administered by oral gavage to mice in a 60 hour treatment. The LDN193189 was dissolved in citric acid at pH 3.1 and the oral gavage was given twice daily. This resulted in an expansion of GLP1 expressing cells in the villus of the intestine compared to control mice. Secretin expression was greatly reduced in cells in the villus of the intestine compared to control mice (see FIG. 22). The dose used is at the high end and it is anticipated that lower doses would also be effective. It is hypothesised that these results can be extrapolated to humans. LDN193189 is a selective BMP signaling inhibitor that inhibits the transcriptional activity of the BMP type I receptors ALK2 and ALK3 with IC50 of 5 nM and 30 nM in C2C12 cells, respectively, exhibits 200-fold selectivity for BMP versus TGF-. Other inhibitors of BMP signalling, such as those described herein, are also envisaged to have similar effects in vivo. It is concluded that BMP inhibitors (and conversely BMP activators) have potential for use in therapy where modulation of GLP1 and/or secretin in vivo is desirable. Examples include, but are not limited to, treatment of diabetes, obesity, hypochlorhydia or hyperchlorhydria.

[0668] Structure of LDN193189:

##STR00044##