WO2009126927A2 - Methods of generating insulin-producing cells - Google Patents

Abstract

A method of generating a cell population having insulin-producing cells includes obtaining a population of intrahepatic biliary epithelial cells; expanding the population of cells; culturing the expanded population of cells in a medium lacking insulin and serum; and contacting the population of cells with a differentiation condition for a time sufficient that at least a portion of the cells secrete insulin.

Description

METHODS OF GENERATING INSULIN-PRODUCING CELLS

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims the benefit of priority of U.S. Provisional Patent Application Serial No. 61/123,968, filed April 11, 2008, the entire contents of which are hereby incorporated by reference.

TECHNICAL FIELD

This disclosure relates to methods of generating insulin-producing cells, and more particularly to methods of generating insulin-producing cells from intrahepatic bile epithelial cells (IHBECs).

BACKGROUND In 2005, the CDC estimated that 20.8 million people in the U.S. had diabetes.

There is currently a need for additional sources of insulin-producing cells for transplantation to treat diabetic conditions.

SUMMARY

The disclosure is based, inter alia, on the surprising discovery that IHBECs can be transdifferentiated into insulin-producing cells by various methods. Accordingly, this application provides methods of insulin-producing cells from IHBECs. These insulin- producing cells can introduced into subjects for the treatment of diabetic conditions. Accordingly, in one aspect the disclosure features methods of generating cell populations that include insulin-producing cells. The methods include the steps of: (a) obtaining a population of intrahepatic biliary epithelial cells; (b) optionally, expanding the population of cells; (c) optionally, culturing the IHBECs or the expanded population of cells in a medium lacking insulin and serum; and (d) contacting the IHBECs or the expanded population of cells with one or more differentiation conditions for a time sufficient that at least a portion of the cells secrete insulin, thereby generating a cell population that includes insulin-producing cells. In some embodiments, the cell population that includes insulin-producing cells produces at least 0.1 ng/ml (e.g., at least

1, 10, 100, or 1000 ng/ml) insulin per 10⁵ cells over 48 hours in culture. In some embodiments, the cell population that includes insulin-producing cells secretes insulin in response to glucose or an insulin secretagogue.

In some embodiments, expanding the population of cells includes culturing the cells in a three dimensional scaffold (e.g., a collagen gel). In some embodiments, the population of cells is expanded in a growth medium that includes insulin and/or serum. In some embodiments, a differentiation condition includes introducing into the population of cells a nucleic acid that expresses a transcription factor (e.g., Pdx-1, NeuroD, Pdx 1/VP16, Ngn3, Nkxβ.l, or pancreas-specific transcription factor lα (Ptflα)). In some embodiments, the nucleic acid is contained within a vector, e.g., a viral vector (e.g., an adenovirus vector, an adeno-associated virus vector, or a retroviral vector). In other embodiments, a differentiation condition includes inhibiting Notch signaling in the cells, e.g., by contacting the population of cells with a gamma-secretase inhibitor. In further embodiments, the differentiation condition includes inhibiting HNF6 expression in the cells, e.g., by introducing to the population of cells a nucleic acid that inhibits HNF6 expression.

In other embodiments, a differentiation condition includes exposing the cells to one or more factors produced by pancreatic cells, e.g., fetal pancreatic cells. In some embodiments, a differentiation condition includes mixing the population of cells with a population of fetal pancreatic cells or contacting the population of cells with conditioned medium from fetal pancreatic cells.

In some embodiments, the methods further include isolating the portion of the cells that secrete insulin, e.g., by cell sorting or immunological methods. The cell population that includes insulin-producing cells or the isolated portion can be introduced into a subject, e.g., a human. In some embodiments, the initial population of intrahepatic biliary epithelial cells is obtained from the same subject to which the cell population that includes insulin-producing cells or the isolated portion is introduced.

In another aspect, the disclosure features a cell population comprising insulin- producing cells generated by any of the methods described herein.

In a further aspect, the disclosure features methods of generating an insulin- producing cell. The methods include providing an intrahepatic biliary epithelial cell or a cell derived from an intrahepatic biliary epithelial cell; and contacting the intrahepatic biliary epithelial cell or the cell derived from an intrahepatic biliary epithelial cell with one or more differentiation conditions for a time sufficient that the cell secretes insulin, thereby generating an insulin-producing cell. In some embodiments, the insulin- producing cell secretes insulin in response to glucose or an insulin secretagogue. In some embodiments, the differentiation condition includes introducing into the intrahepatic biliary epithelial cell or the cell derived from an intrahepatic biliary epithelial cell a nucleic acid that expresses a transcription factor (e.g., Pdx-1, NeuroD, Pdx 1/VP16, Ngn3, Nkxβ.l, or pancreas-specific transcription factor lα (Ptflα)). In some embodiments, the nucleic acid is contained within a vector, e.g., a viral vector (e.g., an adenovirus vector, an adeno-associated virus vector, or a retroviral vector). In other embodiments, a differentiation condition includes inhibiting Notch signaling in the intrahepatic biliary epithelial cell or the cell derived from an intrahepatic biliary epithelial cell, e.g., by contacting the intrahepatic biliary epithelial cell or the cell derived from an intrahepatic biliary epithelial cell with a gamma-secretase inhibitor. In further embodiments, a differentiation condition includes inhibiting HNF6 expression in the cells, e.g., by introducing to the intrahepatic biliary epithelial cell or the cell derived from an intrahepatic biliary epithelial cell a nucleic acid that inhibits HNF6 expression.

In other embodiments, a differentiation condition includes exposing the intrahepatic biliary epithelial cell or the cell derived from an intrahepatic biliary epithelial cell to one or more factors produced by pancreatic cells, e.g., fetal pancreatic cells. In some embodiments, a differentiation condition includes mixing the intrahepatic biliary epithelial cell or the cell derived from an intrahepatic biliary epithelial cell with a population of fetal pancreatic cells or contacting the intrahepatic biliary epithelial cell or the cell derived from an intrahepatic biliary epithelial cell with conditioned medium from fetal pancreatic cells.

In another aspect, the disclosure features an insulin-producing cell generated by any of the methods described herein.

In a further aspect, the disclosure features methods of treating a subject with an insulin deficiency (e.g., diabetes type I or type II) by generating a cell population that includes insulin-producing cells by a method described herein and introducing into the subject the cell population that includes insulin-producing cells or an isolated portion thereof.

In some embodiments, the disclosure includes methods of generating a cell population comprising insulin-producing cells where the methods include obtaining a intrahepatic biliary epithelial cell and culturing (e.g., expanding) the cell in a first medium comprising insulin and serum for a first time period, e.g., to produce a cell population. The resulting cell or cells can then be isolated from the first medium, or certain components of the first growth medium (e.g., insulin and/or serum) can be removed from the medium, before culturing (e.g., expanding) the cell or cells in a second medium lacking insulin and serum for a second time period to produce a population of intrahepatic biliary epithelial cells. This population of intrahepatic biliary epithelial cells can then be contacted with at least one (e.g., one or more) of a nucleic acid encoding a transcription factor (e.g., Pancreatic and duodenal homeobox-1 (Pdx-1), NeuroD, Pdx 1/VP 16 (which is also known in the art or is marketed as Etoposide, Etopophos^®, Vepesid^®), Neurogenin 3 (Ngn3), Nkxβ.l, and pancreas-specific transcription factor lα (Ptflα), an inhibitor of Notch signaling, an inhibitor of HNF 6 expression, a fetal pancreatic cell, or a solution contacted by cultured fetal pancreatic cells for a time sufficient that at least a portion of the cells secrete insulin to generating a cell population comprising insulin-producing cells. In some embodiments, at least 10%, e.g., at least 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, 98%, 99% or 100% of the cells in the cell population comprising insulin- producing cells are insulin producing cells.

Among the cell populations in the liver, the intrahepatic bile duct epithelium seems to have considerable promise as a new source of insulin-producing cells, β-cells have been reported in the common bile duct (CBD) (Hara et al., 2003, Am. J. Physiol. Endocrinol. Metab., 284:E177-183; Dutton et al., 2007, J. Cell ScL, 120:239-245). However, extrahepatic duct cells are less desirable as a source of insulin-producing cells because the removal of sufficient amounts of tissue can require difficult surgery with the potential for severe morbidity (Antolovic et al., 2007, J. Gastrointest. Surg., 11 :555-561). Another potential drawback is that only a small numbers of cells would be available. In contrast, this application provides an advantage, as the surgical removal of a modest amount of liver parenchyma can provide for safe isolation of large numbers of intrahepatic biliary epithelial cells (IHBECs) for transdifferentiation. A considerable amount of liver tissue can be surgically removed with less potential for complications because of the capacity of the liver for regeneration (Michalopoulos et al., 1997, Science, 276:60-66).

The current methods provide for the generation of large numbers of IHBECs for differentiation and maintenance of long-term (>24 hour) viability of these cell preparations. The capacity to generate large numbers of insulin-producing cells from IHBECs can allow for the production of sufficient numbers of cells for transplantation. If sufficient numbers of insulin-producing cells are generated, surplus cells can be frozen for later transplantation.

The term "treating" as used herein includes reducing or alleviating at least one adverse effect or symptom, e.g., absolute or relative insulin deficiency, fasting hyperglycemia, glycosuria, development of arteriosclerosis, microangiopathy, nephropathy, and neuropathy, of disorders characterized by insufficient insulin activity.

As used herein, a cell that is "derived from" an animal is a cell that was taken from the animal, or a cell that is a progeny cell of a progenitor cell that was taken from the animal, e.g., removed from the animal surgically or by some other method.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs. Although methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present disclosure, suitable methods and materials are described below. All publications, patent applications, patents, and other references mentioned herein are incorporated by reference in their entirety. In case of conflict, the present specification, including definitions, will control. In addition, the materials, methods, and examples are illustrative only and not intended to be limiting.

The details of one or more embodiments of the disclosure are set forth in the accompanying drawings and the description below. Other features, objects, and advantages of the disclosure will be apparent from the description and drawings, and from the claims. DESCRIPTION OF DRAWINGS

FIG. 1 is an outline of an exemplary protocol for inducing the differentiation of IHBECs into insulin-producing cells (Group 3) and appropriate controls (Group 1 and Group 2). FIG. 2 A is a representation of the macroscopic appearance of IHBECs with

CEFCM at days 0 and 14. Isolated IHBECs were suspended in rat tail collagen plus growth medium; the collagen gel containing the IHBECs gradually contracted. Scale bar, 10 mm.

FIG. 2B is a phase contrast light microscopy image of cultured IHBECs. Left, day 0 with CEFCM; right, after 5 days with CEFCM, IHBECs formed three-dimensional ductal cysts.

FIG. 2C is a phase contrast light microscopy image of cultured IHBECs at 14 day with CEFCM. Asterisk shows the same position as in panel A. The IHBECs had expanded to form ductal structures. Scale bar, 100 um. FIG. 2D is a line graph depicting the diameter of the collagen gel over time.

FIG. 2E is a line graph depicting the number of IHBECs over time for culture with or without CEFCM. Bars show mean ± SEM of five independent experiments. P<0.05.

FIGs. 2F-2I are RT-PCR gels showing gene expression profiles of IHBECs at days 0 and 17 of culture, and at day 26 following transduction with AdV-Pdx- 1/VP 16. Mouse liver (FIGs. 2F-2H) or islet (FIG. 21) was used as a control. The oligonucleotide primers and cycle number used for semiquantitative PCR are shown in Table 1. All primers were designed to cross intron(s). Mouse liver and islets were used as controls. Gene expression studies in all groups were repeated at least 3 times with similar results. FIG. 2F, hepatocyte markers (tryptophan oxygenase [TO], tyrosine aminotransferase [TAT], cytochrome P450 1A2 [CYP1A2]). FIG. 2G, BEC markers (Cytokeratin [CK] 7 and 19, γ-glutamyltranspeptidase [γ-GTP]. FIG. 2H, endoderm progenitor markers (HNFsI β, 3β, 4α, and 6). FIG. 21, islet markers (insulinl, glucagon, somatostatin, and pancreatic polypeptide).

FIGs. 2J-K are immunofluorescence micrographs depicting staining for CK 7 in IHBECs. IHBECs were positive for CK 7 (blue) at days 0 (FIG. 2J) and 17 (FIG. 2K). Propidium Iodide (PI, red) was used to show nuclei. Scale bar, 100 um. FIG. 3 A is a pair of phase contrast light microscopy images of IHBECs cultured for 17 days with (right) or without (left) CEFCM. At 17 days in culture, the CEFCM- cells displayed signs of senescence (black arrows), including flattening and multinucleation, and numerous cytoplasmic vacuoles (white arrow heads). FIG. 3B is a bar graph depicting the number of signs of senescence and cytoplasmic vacuoles per colony at day 17 for cells cultured with or without CEFCM. Bars show mean ± SEM of 25 independent experiments. *, P < 0.095.

FIGs. 4A-4C are sets of micrographs of IHBECs transduced with AdV- Pdx-1 (FIG. 4A), NeuroD (FIG. 4B), or Pdx-l/VP16 (FIG. 4C) at the indicated MOI with each AdV expressing GFP as a reporter. For each figure, the left panel is a phase contrast light microscopy image, the middle panel depicts GFP fluorescence, and the right panel depicts the two images merged together. Scale bar, 100 um.

FIGs. 5A-5D are RT-PCR gels showing gene expression profiles of IHBECs at day 14 of culture, Group 1 control cells (differentiation medium (DM only)), Group 2 control cells (AdV-GFP), and Group 3 cells transduced with AdV expressing Pdx-1,

NeuroD, or Pdx-l/VP16. Mouse islets, liver, and insulinoma cell line MIN6 were used as positive controls. The oligonucleotide primers and cycle number used for semiquantitative PCR are shown in Table 1. All primers were designed to cross intron(s). Gene expression studies in all groups were repeated at least 3 times with similar results. FIG. 5A, beta-cell related markers (insulin 1 and 2, glucagon, somatostatin, pancreatic polypeptide (PP), prohormone convertase (PC) 1 and 2, glucose transporter 2 (GLUT2)). FIG. 5B, transcription factors (Neurogenin 3 (Ngn3), NeuroD, Nkx2.2, Paired box 4 (Pax4), Nkxβ.l, Pax6, MafA, Pdx-1, and islet- 1 (IsIl)). FIG. 5C, exocrine cell related markers (pancreas-specific transcription factor lα (Ptflα), Amylase 2). FIG. 5D, early liver and pancreas markers (HNFl β and HNF6) and β-actin control.

FIG. 5E is a bar graph depicting quantitative Insulin 1 mRNA analysis for IHBECs (Group 1), AdV-GFP control cells (Group 2), and transduced cells (Group 3 (Pdx-1 :white column, NeuroD: dot column, Pdx-1 /VP16: black column)) assessed 7 days after transduction. There was no significant difference among the groups. Quantification of insulin mRNA levels was carried using the TaqMan™ real-time PCR system. Mouse islets were used as a positive control and calculated for relative insulin 1 mRNA quantification. Data are presented as mean ± SEM of three independent experiments.

FIGs. 6A-6D are sets of immunofluorescence micrographs of AdVs-P dx- 1/VP 16 transduced IHBECs at 7 days after transduction. FIG. 6A, triple staining for NeuroD (left, red), GFP (middle, green), and CK7 (right, blue). FIG. 6B, triple staining for

NeuroD, (left, red), GFP (middle, green), and DAPI nuclear stain (right, blue). FIG. 6C, triple staining for Pdx-1 (left, red), GFP (left, green), and CK7 (right, blue). FIG. 6D, triple staining for Pdx-1, (left, red), GFP (middle, green), and DAPI nuclear stain (right, blue). For FIGs. 6B and 6D, there was complete overlap between transcription factors and DAPI in the nuclei.

FIGs. 6E-6F are sets of immunofluorescence micrographs of AdVs-P dx- 1/VP 16 transduced IHBECs at 7 days after transduction showing C-peptide staining for the confirmation of insulin synthesis. The micrographs show triple staining for C-peptide (left, red), GFP (middle, green), and CK-7 (right, blue). Of 4000 cells counted, 2.8% in Pdx-l/VP16 transduced cultures stained strongly for C-peptide. Scale bars, 50 μm. FIGs. 6G-6H are transmission electron micrographs of AdVs- Pdx-l/VP16 transduced IHBECs 7 days after transduction. FIG. 6G shows secretory granules densely packed in the cytoplasm. Some granules show a clear halo surrounding a denser core, a morphology that is characteristic of insulin-containing granules. These cells remains apical microvilli which polarity of IHBECs (FIG. 6G, asterisks). FIG. 6H shows a higher magnification view of the secretory granules, with the crystalline formation of the granular core. Scale bars, 1 μm.

FIG. 7A is a bar graph depicting insulin release into the culture medium of IHBECs (Group 1), control cells (Group 2), transduced cells (Group 3, Pdx-1 : white column, NeuroD: dotted column, Pdx-l/VP16: black column) following a 48 hour incubation. $, P = 0.0342. *, P < 0.001. #, P = 0.0065.

FIG. 7B is a dot graph depicting insulin release in response to insulin secretagogues after 2 hour incubations by IHBECs (Group 1), control cells (Group 2), transduced cells (Group 3, Pdx-1 : white column, NeuroD: dotted column, Pdx-l/VP16: black column). Insulin secretion was evaluated in serum-free DM or Krebs-Ringer bicarbonate buffer (KRBB) with 0.2% BSA supplemented with 5 mM glucose, 25 mM glucose, or 25 rnM glucose + 45 mM KCl, and quantitated by ELISA (detection limit of 0.04 ng/ml). Dotted arrows show undetectable insulin secretion.

FIGs. 8A-8B are scatter graphs depicting flow cytometry analysis of control and AdVs-Pdx-l/VP16 transduced IHBECs 7 days after transduction. FIG. 8 A, flow cytometry analysis of untransduced control groups (box 1, gated cells with low PI staining and GFP fluorescence). FIG. 8B, flow cytometry analysis of AdVs- Pdx-l/VP16 transduced IHBECs 7 days after transduction (box 2, gated live non-GFP-emitting cells; box 3, gated live, GFP-emitting cells effectively transduced with AdVs).

DETAILED DESCRIPTION This application described methods of generating insulin-producing cells from

IHBECs and methods of introducing those insulin-producing cells into subjects for the treatment of disorders, e.g., diabetes. The disclosure is based, inter alia, on the surprising discovery that IHBECs are capable of being transdifferentiated into insulin-producing cells by contacting the IHBECs or cells derived therefrom with various differentiation conditions.

Obtaining and Culturing IHBECs

IHBECs are the cells that comprise the epithelial cell lining of the intrahepatic biliary tree. IHBECs are distinct from epithelial cells of the extrahepatic ducts and the gall bladder. Methods of obtaining, isolating, and culturing IHBECs, including from human subjects, are known in the art. See, e.g., Joplin, 1994, Gut, 35: 875-878; Joplin et al., 1992, J. Clin. Invest., 90:1284-89; Auth et al., 2001, Hepatology, 33:519-529; and Ochiai et al., 2004, Pediatr. Surg. Int., 20:685-688. Exemplary methods of obtaining IHBECs include biopsy and perfusion of the liver with digestive enzymes, such as collagenase. IHBECs can be isolated from other liver cells by, e.g., differential density gradient centrifugation (Gall et al., 1985, Cell Biol. Int. Rep., 9:315-322; Sirica et al., 1990, Pathobiology, 58:44-64), centrifugal elutriation (Yaswen et al., 1984, Cancer Res., 44:324-331), fluorescence activated cell sorting (Doolittle et al., 1987, Hepatology, 7:696-703), and immunological selection methods (Joplin et al., 1989, In Vitro Cell Dev. Biol, 25:1189-92; Ishii et al., 1989, Gastroenterology, 97:1236-47).

To increase the numbers of isolated IHBECs, e.g., to obtain an enriched population of IHBECs, the cell or cells can be cultured in a liquid medium or on a support (e.g., a collagen gel support). In some embodiments, the medium can contain one or more of growth factors (e.g., EGF, HGF), insulin, and serum. See, e.g., Joplin et al., 1992, J. Clin. Invest., 90:1284-89; Ochiai et al., 2004, Pediatr. Surg. Int., 20:685-688.

Differentiation Conditions

In some embodiments, IHBECs or cells derived from IHBECs are exposed to one or more differentiation conditions such that at least a portion of the differentiated cells secrete insulin. The cells can be exposed to the one or more differentiation conditions simultaneously or consecutively. The cells can be exposed to the differentiation conditions either in vitro or in vivo. In some embodiments, the differentiation conditions include increasing the expression or activity of one or more differentiation-related transcription factors in the cells, inhibiting the expression or activity of a component of the Notch signaling pathway in the cells, inhibiting the expression or activity of HNF 6 in the cells, exposing the cells to factors secreted by pancreatic cells (e.g., fetal pancreatic cells), or transplantation of the cells into a subject. Typically, if the cells are exposed to the one or more differentiation conditions in vitro, the cells will be in a medium that does not contain insulin or serum.

In some embodiments, a differentiation condition includes increasing the expression or activity of one or more transcription factors related to pancreatic lineage. For example, the one or more transcription factors can be selected from: pancreatic and duodenal homeobox 1 (Pdx-1; GenBank Accession No. NP 000200), neurogenic differentiation 1 (NeuroD; GenBank Accession No. NP 002491), Pdx-1 carrying the VP16 transactivator domain (Pdx-1 /VP16; see Kaneto et al, 2005, Diabetes, 54:1009- 22), neurogenin 3 (Ngn3; GenBank Accession No. NP_066279), NK6 transcription factor related, locus 1 (Nkxβ.l; GenBank Accession No. NP 006159), or pancreas-specific transcription factor lα (Ptflα; GenBank Accession No. NP 835455). The one or more transcription factors can be expressed (e.g., exogenously expressed) within the cell by any means known in the art. To generate cells that express a transcription factor, the cells may be transfected, transformed, or transduced using any of a variety of techniques known in the art. Any number of transfection, transformation, and transduction protocols known to those in the art may be used, for example those outlined in Current Protocols in Molecular Biology, John Wiley & Sons, New York. N.Y., or in numerous kits available commercially (e.g., Invitrogen Life Technologies, Carlsbad, Calif). Such techniques may result in stable transformants or may be transient. One suitable transfection technique is electroporation, which may be performed on a variety of cell types, including mammalian cells, yeast cells and bacteria, using commercially available equipment. Optimal conditions for electroporation (including voltage, resistance and pulse length) are experimentally determined for the particular host cell type, and general guidelines for optimizing electroporation may be obtained from manufacturers.

Exemplary methods of expression include transduction with a virus that includes a nucleic acid that expresses the transcription factor, e.g., an adenovirus, an adeno- associated virus, or a retrovirus.

In some embodiments, a differentiation condition includes inhibiting the expression or activity of a component of the Notch signaling pathway (Leach, 2005, J. Clin. Gastroenterol, 39:S78-82) in the cells. For example, Notch activity can be inhibited by contacting the cells with an antibody that binds specifically to Notch or a Notch ligand. In exemplary methods, Notch signaling is inhibited by contacting the cells with an inhibitor of gamma-secretase. In other exemplary methods, the expression of a component of the Notch signaling pathway can be inhibited, e.g., by antisense or siRNA methods.

In some embodiments, a differentiation condition includes inhibiting the expression or activity of hepatocyte nuclear factor 6 (HNF6; human mRNA sequence GenBank Accession No. NM 004498). In exemplary methods, the expression of HNF 6 can be inhibited, e.g., by antisense or siRNA methods. In some embodiments, insulin-secreting cells can be isolated following differentiation. For example, the insulin-secreting cells can be isolated by cell sorting or immunological selection methods to obtain a population of cells in which at least 10%, e.g., at least 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, 98%, 99% or 100% of the cells in the population express and/or secrete insulin.

Transplantation Methods

Common methods of administering cells, e.g., insulin-secreting cells, to subjects, particularly human subjects, are described in detail herein. For example, cells can be administered to a subject by injection or implantation of the cells into target sites in the subject. In addition, the cells can be inserted into a delivery device which facilitates introduction by injection or implantation of the cells in the subjects. Such delivery devices include tubes, e.g., catheters for injecting cells and fluids in to the body of a recipient subject. In a preferred embodiment, the tubes additionally have a needle, e.g., a syringe, through which the cells of the disclosure can be introduced into the subject at a desired location. The pancreatic cells can be inserted into such a delivery device, e.g., a syringe, in different forms. For example, the cells can be suspended in a solution or embedded in a support matrix when contained in such a delivery device. As used herein the term "solution" includes a pharmaceutically acceptable carrier or diluent in which the cells of the disclosure remain viable. Pharmaceutically acceptable carriers and diluents include saline, aqueous buffer solutions, solvents and/or dispersion media. The use of such carriers and diluents is well known in the art. The solution is preferably sterile and fluid to the extent that easy syringability exists. Preferably, the solution is stable under the conditions of manufacture and storage and preserved against the contaminating action of microorganisms such as bacteria and fungi through the use of, for example, parabens, chlorobutanol, phenol, ascorbic acid, thimerosal and the like. Solutions of the disclosure can be prepared by incorporating the pancreatic cells described herein in a pharmaceutically acceptable carrier or diluent and as require other ingredients enumerated above, followed by filtered sterilization.

Support matrices in which cells, e.g., insulin-secreting cells, can be incorporated or embedded include matrices that are recipient-compatible and that degrade into products that are not harmful to the recipient. Natural and/or synthetic biodegradable matrices are examples of such matrices. Support matrices include plasma clots; collagen matrices; basement membrane matrices (e.g., Matrigel™ basement membrane matrix) (Bonner-Weir et al, 2000, Proc. Natl. Acad. Sci. USA, 9:7999-8004); thermoreversible gelation polymer (TGP) matrices; alginate (Duvivier-Kali et al., 2001, Diabetes, 50:1698- 1705); and synthetic polymers such as polyanhydrides, polyorthoesters, and polylactic acid. Other example of synthetic polymers and methods of incorporating or embedding cells into the matrices are known in the art. See, e.g., U.S. Pat. No. 4,298,002 and U.S. Pat. No. 5,308,701. The matrices provide support and/or protection (e.g., protection from immunological responses) for the cells, e.g., insulin-secreting cells, in vivo. Other methods of encapsulation of cells to provide protection from immunological responses are known in the art, such as implantation of cells in implantable hollow fibers. Such fibers can be pre-spun and subsequently loaded with the cells (Aebischer et al. U.S. Pat. No. 4,892,538; Hoffman et al., 1990 Expt. NeurobioL, 110: 39-44; Jaeger et al., 1990, Prog. Brain Res., 82: 41-46; and Aebischer et al, 1991, J. Biomech. Eng., 113:178-183), or can be co-extruded with a polymer that acts to form a polymeric coat about the cells (Lim, U.S. Pat. No. 4,391,909; Sefton, U.S. Pat. No. 4,353,888; Sugamori et al., 1989, Trans. Am. Artif. Intern. Organs, 35:791-799; Sefton et al., 1987, Biotehnol. Bioeng., 29:1135-1143; and Aebischer et al., 1991, Biomaterials 12:50-55). Rezania et al., U.S. Pat. App. Pub. No. 2004/0197374, describes implantable pouches that can be seeded with insulin-producing cells for transplantation. The terms "introduction," "administration," and "transplantation" are used interchangeably herein and refer to delivery of cells to a subject by a method or route which delivers the cells to a desire location.

Common methods of administering cells, e.g., insulin-secreting cells, to subjects, particularly human subjects, include implantation of cells in a pouch of omentun (Yonda et al., 1989, Diabetes, 38: 213-216); intraperitoneal injection of the cells (Wahoff et al., 1994, Transplant. Proc, 26:804); implantation of the cells under the kidney capsule of the subject (see, e.g., Liu et al., 1991, Diabetes, 40:858-866; Korgren et al., 1998, Transplantation, 43: 509-514; Simeonovic et al., 1982, Aust. J. Exp. Biol. Med. ScL, 60:383); subcutaneous injection; injection into the epididymal fat pad; and intravenous injection of the cells into, for example, the portal vein (Braesch et al., 1992, Transplant. Proc, 24: 679-680; Groth et al., 1992, Transplant. Proc, 24:972-973). To facilitate transplantation of the cells, the cells can be embedded in a support matrix. Cells can be administered in a pharmaceutically acceptable carrier or diluent as described herein.

Disorders

The term "diabetes" is a general term to describe diabetic disorders as they are recognized in the art, e.g., Diabetes Mellitus. Diabetes Mellitus is characterized by an inability to regulate blood glucose levels. The two most prevalent types of diabetes are known as Type I and Type II diabetes. The term also encompasses the myriad secondary disorders caused by diabetes, both acute and chronic, e.g., diabetic complications, e.g., hypoglycemia and hyperglycemia, retinopathy, angiopathy, neuropathy, and nephropathy. Examples of diabetes include insulin dependent diabetes mellitus and non-insulin dependent diabetes. Insulin dependent diabetes mellitus (Type 1 diabetes) is an autoimmune disease, where insulitis leads to the destruction of pancreatic beta-cells. At the time of clinical onset of type 1 diabetes mellitus, significant number of insulin producing beta cells are destroyed and only 15% to 40% are still capable of insulin production (McCulloch et al., 1991, Diabetes, 40:673-679). Beta-cell failure results in a life long dependence on daily insulin injections and exposure to the acute and late complication of the disease. Type 2 diabetes mellitus is a metabolic disease of impaired glucose homeostasis characterized by hyperglycemia, or high blood sugar, as a result of defective insulin action which manifests as insulin resistance, defective insulin secretion, or both. A patient with Type 2 diabetes mellitus has abnormal carbohydrate, lipid, and protein metabolism associated with insulin resistance and/or impaired insulin secretion. The disease leads to pancreatic beta cell destruction and eventually absolute insulin deficiency. Without insulin, high glucose levels remain in the blood. The long term effects of high blood glucose include blindness, renal failure, and poor blood circulation to these areas, which can lead to foot and ankle amputations. Early detection is critical in preventing patients from reaching this severity. The majority of patients with diabetes have the non-insulin dependent form of diabetes, currently referred to as Type 2 diabetes mellitus.

Exemplary models of Type I diabetes include: Komeda diabetes-prone rat (Yokoi, 2005, Exp. Anim., 54: 111-115); streptozocin (STZ) treated animals; BBDP rat (Hillebrands et al, 2006, J. Immunol, 177:7820-32); NOD mouse (Aoki et al, 2005, Autoimmun. Rev., 4:373-379); the Gόttingen minipig (Larsen et al., 2004, ILAR J., 45:303-313); and several non-human primate models (Gaur, 2004, ILAR J., 45:324-333).

Exemplary models of Type II diabetes include: a transgenic mouse having defective Nkx-2.2 or Nkx-6.1; (U.S. Pat. No. 6,127,598); Zucker Diabetic Fatty fa/fa (ZDF) rat. (U.S. Pat. No. 6,569,832); Rhesus monkeys, which spontaneously develop obesity and subsequently frequently progress to overt type 2 diabetes (Hotta et al., 2001, Diabetes, 50: 1126-33); and a transgenic mouse with a dominant-negative IGF-I receptor (KR-IGF-IR) having Type 2 diabetes-like insulin resistance.

EXAMPLES Experimental results are expressed as the mean ± standard error of the mean

(mean ± SEM). Student's t test, ANOVA and Fisher's protected least significant difference test were used, and a p value < 0.05 was considered significant. All experiments were repeated at least three times. Example 1. Generation of insulin-producing cells from IHBECs

We have developed a five step-protocol for inducing the differentiation of IHBECs to insulin-producing cells (Fig. 1). In Step 1, the IHBEC-rich fraction is isolated from adult mouse liver with 1 x 10⁵ viable cells/mouse being recovered. Adult 8 week male C57/BL6 mice were obtained from Taconic Farms (Germantown, NY). All animals were bred and maintained under pathogen- free conditions in accordance with housing and husbandry guidelines. To isolate IHBECs, a two-step liver perfusion was performed (Nagaya et al., 2006, Hepatology, 43:1053-62). After removal of hepatocytes, the remnant tissues were collected and minced, transferred into a flask, and then treated with 0.02% soybean trypsin inhibitor (GIBCO, Invitrogen Corporation, California, CA) and 0.04% collagenase solution (collagenase D, Roche, Indianapolis, IN) for 7 minutes. The digested tissues were suspended in Dulbecco's modified Eagle's medium (DMEM, GIBCO) with 10% FBS and 100 mg/L penicillin and 100 mg/L streptomycin, and centrifuged at 150 x g for 1.5 minutes. The pellet was resuspended in medium, filtered sequentially through 250, 100, and 40 μm nylon mesh. Small cell aggregates on the 40 μm mesh were retrieved. 1 x 10⁵ viable cells were typically recovered.

Isolated IHBECs were cultured with collagen embedded floating culture method (CEFCM). Viable isolated cells were suspended in DMEM/F12 (GIBCO) on ice containing 0.3 mg/mL collagen (Collagen type I rat tail, Becton Dickenson (BD), Franklin Lakes, NJ). Then 1.0 x 10⁴ cells/ml/well were plated in a 12-well dishes. The collagen containing the aggregates was allowed to solidify and was then incubated at 37 ⁰C for 2 hours. The cells with CEFCM were then cultured in growth medium (GM), which contained DMEM/F12 supplemented with 5% FBS, 5% NuSerum™ IV Serum (BD), 0.5 μg/mL insulin-transferrin-sodium selenite (ITS) (GIBCO), 10 mmol/L nicotinamide, 1 mmol/L ascorbic acid 2-phosphate, 10^"7 M dexamethasone (Sigma-

Aldrich Corp., St. Louis, MO), 25 ng/ml EGF (Fitzgerald Industries International, Inc., Concord, MA), 10 ng/ml HGF (R&D, systems, Minneapolis, MN), 4 μg/ ml forskolin (Calbiochem, San Diego, CA), and 3.4 μg/ ml 3,3,5-triiodo-L-thyromine (Calbiochem). GM was changed every 2 days. Fig. 2A left and 2B left show macroscopic and phase contrast microscopic views at days 0 and 1 , respectively. After about 5 days with

CEFCM, IHBECs formed three-dimensional ductal cysts (Fig. 2B, right) and rapidly expanded their number about 15-fold within 2 weeks (Fig. 2E). The collagen gel containing the expanding cells gradually shrank over 14 days (Fig. 2D). In cultures without CEFCM (CEFCM^"), IHBECs initially expanded but then started to die at around 8 days (Fig. 2E). Initial experiments with IHBECs were performed to determine their behavior in tissue culture and their potential for transdifferentiation. RT-PCR and immunofluorescent staining were carried out with hepatocyte, biliary epithelial cell (BEC), and islet markers. Total RNA was isolated from cultured cells using RNeasy™ Mini Kits (Qiagen, Valencia, CA) according to manufacturer-suggested protocols, and reverse transcription- polymerase chain reaction (RT-PCR) assays were conducted. The extracted total RNA was reverse-transcribed into cDNA with Omniscript RT Kit (Qiagen) using standard procedures. The cDNA products were then diluted to a concentration corresponding to 20 ng/μl. RT-PCR was performed using PCR Master Mix (ABgene, New York, NY) to monitor the transcription of genes related to the genesis of insulin-producing cells. The primers were complementary to the mRNA sequences of the genes of interest and are listed in Table 1. The primers were designed to cross exon-exon boundaries to exclude the possibility that genomic DNA was amplified. The exception was MafA, which lacks an intron. All primers had a calculated Tm of 60 ⁰C. For quantitative real-time PCR, ABI 7300 real-time PCR system (Applied Biosystems) was used according to the manufacturer's instructions for 36 cycles. The probe and primer sets of mouse insulin 1 (assay identification no. Mm01259683_gl) and β-Actin (part no. 4352341E) were purchased from Applied Biosystems; they are specific for mouse and do not bind to cDNA of mouse kidney, or liver. Twenty ng cDNA was applied to each well, and levels of mRNA were determined as the average of triplet aliquots. Levels of insulin mRNA expression were normalized to those of the internal control β-Actin. Data were compared with results from mouse liver run in parallel. Cells at day 0 expressed BEC markers (Fig. 2G), but neither hepatocyte (Fig. 2F) nor islet (Fig. 21) markers. In addition, IHBECs expressed a number of hepatocyte nuclear factor (HNF) family genes (HNF lβ, HNF3β, HNF4α, HNF6; Fig. 2H) considered to be endoderm progenitor markers (Wilson et al, 2003, Mech. Dev., 120:65-80). Table 1. Sequences of PCR primers

Immunostaining for CK7 and CKl 9 was performed essentially as previously described (Nagaya et al, 2006, Hepatology, 43:1053-62). Briefly, cells on dishes were fixed in cold absolute ethanol after three rinses with phosphate-buffered saline (PBS). Primary antibodies included mouse anti-cytokeratin 7 (CK7; Dako Cytomation) and rabbit anti-cytokeratin 19 (CKl 9), rabbit anti-NeuroD (1 :200; Cell Signaling Technology, Inc., Danvers, MA). Alexa™488-conjugated and Alexa™594-conjugated antibodies (Molecular Probes, Eugene, OR) were used as secondary antibodies. Cell nuclei were counter-stained with PI (VECTASHIELD™ Mounting Medium with PI; Vector Laboratories, Ltd., Peterborough, England). Cultures processed with secondary antibodies only were used as negative controls. Liver and pancreas sections were stained as positive controls. Day 0 and day 17 cells stained positive for both for CK7 (Figs. 2J- 2K) and CKl 9. Dispersed IHBECs at day 14 were seeded onto collagen-coated dishes and cultured for 3 days (Step 3). The cells were washed with Ca²⁺-free HEPES-buffered saline, and a 0.1% collagenase solution was added for 20 minutes to dissolve the collagen gel and release the aggregated IHBECs fragments. The cells were then counted with trypan blue on a hemocytometer. After counting, the dispersed IHBECs were allowed to settle, then resuspended in GM and plated onto collagen-coated, 35-mm plastic dishes (BD) or 12 well dishes (BD) and maintained at 37 ⁰C in a humidified 5% CO₂ incubator. The IHBECs formed colonies and expanded about 1.4-fold during these 3 days.

Contrast light microscopy of cultured IHBECs at day 17 without CEFCM displayed signs of senescence, including enlargement, flattening (Fig. 3 A, black arrows), and increased number of cytoplasmic vacuoles (Fig. 3 A, white arrows). The IHBECs subjected to CEFCM manipulation had fewer senescent cells or cytoplasmic vacuoles (Fig. 3B).

To measure signs of senescence in the cultures, at least 5 colonies were randomly selected, at least 1 mm apart, for each 5 dishes. Twenty five colonies were checked in each group (with or without CEFCM). Aggregates that contained senescent cells or increased numbers of cytoplasmic vacuoles were counted as positive. The monolayers of homogenous appearing cells maintained their BEC character even at day 17 (Figs. 2F-2I) and proliferated efficiently in culture. The purity of IHBECS at day 17 was confirmed by immunostaining and averaged >90%.

For Step 4, the media was changed to serum- and insulin-free differentiation medium (DM), which consisted of DMEM/F12 (GIBCO) supplemented with 10 mM nicotinamide (Sigma), 1 mM ascorbic acid 2-phosphate (Sigma), 10 mM MEM Vitamin Solution (GIBCO), 10 mM MEM Non-Essential Amino Acids Solution, 10 mM GlutaMAX™-I Supplement(GIBCO), 4 mg/mL transferrin (invirogen, California, CA), 10 ng/ml betacellulin (R&D systems), 10 ng/ml IGF-I (R&D systems), and 50ng/ml exendin-4 (Sigma), and the IHBECs were cultured for an additional 2 days.

After two days, the IHBECs were randomly divided into 3 groups. The cells were maintained in untreated condition (Group 1), transduced with adenovirus (AdV) expressing green fluorescent protein (GFP) (Group 2), or transduced with AdV expressing Pdx-1, NeuroD, and Pdx-l/VP16 (Group 3). The media were changed every 2 days. After 7 additional days (Step 5) in DM Control (Group 1), IHBECs started to express some endocrine progenitor genes (Ngn 3, NeuroD, Nkxβ.l and Pdx-1) but lacked insulin mRNA (Fig. 5A). Therefore, the IHBECs were then transduced with AdVs expressing Pdx-1, Neuro D, and Pdx-l/VP16.

Example 2. Transduction of IHBECs with AdVs- Pdx-1. NeuroD. or Pdx-l/VP16

Recombinant AdVs expressing Pdx-1, NeuroD, and Pdx-1 /VP 16 driven by a cytomegalovirus (CMV) promoter were prepared with the AdEasy ™ Adenoviral Vector System (Stratagene, La Jolla, CA) as previously described (Kaneto et al, 2005, Diabetes, 54:1009-22; Yatoh et al., 2007, Diabetes Metab. Res. Rev, 23:239-249). These AdVs all carried the reporter GFP. The control adenovirus (AdVs-GFP) was prepared in the same manner. The integration of each gene into the adenovirus was done by transfection into the adenovirus packing 293 cell line according to the manufacturer's instructions. Adenovirus titers were further increased up to 1 x 10¹⁰ plaque forming units (PFU)/ml with Vivapure AdenoPACK™ 100 purification kits (Vivascience, Edgewood, NY). For transduction, the medium was changed to serum- free DMEM medium containing purified recombinant AdV-Pdx-1, -Neuro D, or Pdx-l/VP16 and incubated for 2 hours at 37 ⁰C. To test which multiplicity of infection (MOI) was most effective for transduction, IHBECs were infected at MOIs of 1, 10, 25, 50, 100, or 200 for 2 hours at 37 ⁰C. Two days after transduction, GFP⁺ cells were counted by microscope (Olympus, Tokyo, Japan) to determine the transduction efficiency. Transduction efficiency was describes as percentage of GFP⁺ cells in all cells. Seven days after transduction, the cells were harvested and evaluated.

The 50 MOI recombinant adenovirus transductions for Pdx-1, NeuroD, or 100 MOI for Pdx-l/VP16 resulted in efficient expression of the transgenes under the control of CMV promoter having 30-50% transduction efficiency with high cell survival (Figs. 4A-4C).

Example 3. Gene Expression Profiles of IHBEC-Derived Insulin-Producing Cells To analyze the molecular events occurring in IHBECs during the series of culture steps, gene expression profiles of transcription factors and pancreas-related genes at day 26 of culture were determined by RT-PCR as described above (Figs. 5A-5D). Whereas IHBECs at Step 1 only expressed biliary epithelial cell markers (Fig. 2G) and early liver and pancreas markers (Fig. 2H), At the end of Step 4, Group 1 (medium control) cells expressed the pancreatic progenitor markers Ngn3, NeuroD, Nkxβ.l, Pdx-1 (Fig. 5B), and pancreas-specific transcription factor lα (Ptflα) (Fig. 5C). These genes were not observed in IHBECs at day 14 in GM culture, therefore, these characteristics were obtained by the IHBECs in the DM culture. Transduced expression of Pdx-1, NeuroD or Pdx-l/VP16 (Group 3) led to expression of not only insulin but also GLUT2 and prohormone convertase 1 and 2 (Fig. 5A). Pdx-1 or Pdx-1 /VP 16 transduced cells faintly expressed amylase 2 (Fig. 5C). Expression of Ptflα was found in all groups of the cells (Fig. 5C).

Insulin 1 mRNA expression in all groups was assessed on day 26 by quantitative RT-PCR (Figs. 5 A, 5E). Pdx-1 /VP 16 was the most effective inducer of insulin expression in IHBECs, however, compared with primary mouse islets, the expression of insulin mRNA was low (Adv-Pdx-1, 6.18 ± 0.45 x 1(T⁵; Adv-NeuroD, 5.81 ± 0.54 x 10^~5, Adv-Pdx-l/VP16, 3.86 ± 1.71 x 10 ⁵ compared to adult mouse islets). Insulin 1 mRNA was not detected in IHBECs cultured in DM medium control (Group 1), or Adv-GFP control cells (Group 2) (Fig. 5E).

Example 4. Immunostaining of IHBECs After Transduction with AdVs- Pdx- 1 /VP 16

Pdx-1 /VP 16 transduction led to the appearance of endocrine precursor and islet related markers. Briefly, cells on dishes were fixed in cold absolute ethanol after three rinses with phosphate-buffered saline (PBS). Primary antibodies included mouse anti- cytokeratin 7 (CK7; Dako Cytomation), rabbit anti-cytokeratin 19 (CKl 9), rabbit anti- NeuroD (1 :200; Cell Signaling Technology, Inc., Danvers, MA), rabbit anti-Pdx-1 (1 :2000; IPF-IC antibody) and rabbit anti-C-peptide (1 :400; Cell Signaling). Alexa™488-conjugated and Alexa™594-conjugated antibodies (Molecular Probes, Eugene, OR) were used as secondary antibodies. Cell nuclei were counter-stained with PI (VECTASHIELD ™ Mounting Medium with PI; Vector Laboratories, Ltd., Peterborough, England) or DAPI (VECTASHIELD ™ Mounting Medium with DAPI). Immunofluorescence analysis showed that cells stained for NeuroD⁺ in the nuclei co-expressed both GFP and CK7 in the cytoplasm (Fig. 6A). There was complete overlap between NeuroD and DAPI in the nuclei (Fig. 6B), although a relatively small number of GFP^" cells were NeuroD⁺ (data not shown). Cells with Pdx-1 staining in the nucleus also co-expressed GFP and CK7 (Fig. 6C). As expected, because of the Pdxl/VP16 transduction the number of Pdx-1⁺ cells was larger than that of NeuroD⁺ cells (Fig. 6C). C-peptide staining confirmed the synthesis of insulin by IHBECs-derived insulin- producing cells. Pdxl/VP16 transduced cells were positive for C-peptide and CK7, indicating they maintained characteristics of BECs. Of 4000 cells counted in Pdx-1 /VP16 transduced cultures, 30.1% and 2.8% were GFP⁺ and C-peptide⁺, respectively (Figs. 6E, F). Moreover, some cells transduced with AdVs-P dx- 1/VP 16 contained secretory granules observable by electron microscopy. These cells retained apical microvilli, characteristic of the polarity of IHBECs (Fig. 6G, asterisks). Under higher magnification, several cells were observed to contain granules with a clear halo surrounding a dense core, a morphology characteristic of insulin-containing granules (Fig. 6H).

Example 5. Insulin Release by IHBEC-Derived Insulin-Producing Cells

Insulin release into the culture medium was measured 7 days after transduction (Fig. 7A). Insulin levels in culture supernatants was measured using the Ultra Sensitive Insulin ELISA (Crystal Chem, Inc, IL) (detection limit of 0.04 ng/ml). The fold stimulation was calculated for each culture by dividing the insulin concentration in the stimulation supernatant by the insulin concentration in the basal supernatant. After 48 hours in culture, AdVs-Pdx-1, NeuroD, and Pdx-l/VP16 transduced IHBECs all released insulin. Pdx-1 /VP 16 transduced cells had significantly higher insulin secretion than other transduced cells (Pdx-1, 1.1 ± 0.1; NeuroD, 8.0 ± 1.4; Pdx-l/VP16, 17.3 ± 3.1 ng / 1.2 x 10⁵ cells) (Fig. 7A). The large amount of insulin released over 48 hours is surprising considering that only a small percentage of cells (about 3%) were more highly differentiated. Cells of control groups 1 and 2 did not detectably release insulin (Fig. 7A). The release of insulin into the culture medium in response to various stimuli was monitored for group 3 cells (Fig. 7B). The cells were subjected to static incubation in Krebs-Ringer bicarbonate buffer (KRBB; 133 mM NaCl, 4.69 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 25 mM HEPES, 2.52 mM CaCl₂^H₂O, 5 mM NaHCO₃) supplemented with 5 mM glucose and 0.2% BSA. The cells were plated in 12-well culture plates (BD) cells and washed four times with KRBB. They were then incubated in KRBB with low glucose (5 mM), high glucose (25 mM), or high glucose with 45 mM KCl. The concentration of the insulin secreted into the buffer solution was measured as described above. IHBECs (Group 1, n=6), control cells (Group 2, n = 6), transduced cells (Group 3, Pdx-1 : n = 8, NeuroD: n = 6, Pdx-1 /VP16: n =8 ) were assessed by sequential experiments. No basal insulin release could be detected for any of the transduced cells. After exposure to 25 mM glucose, the cells transduced with Pdx-1 or NeuroD had no detectable insulin release, but those transduced with Pdx-l/VP16 transduced cells released 0.07 ± 0.03 ng/1.2 x 10⁵ cells. These data suggest that these transduced cells are responsive to glucose.

Example 6. Characterization of Transduced IHBECs by Single Cell PCR

From the previous results, it was clear that IHBECs transduced with Pdx-1, NeuroD, or Pdx-l/VP16 produce insulin. Single-cell gene expression is measured in these cells by RT-PCR. The samples were analyzed on an ARIA cell sorter with the

Summit software (BD). The cells were dispersed to mostly single cells, and propidium iodide (PI, Sigma) was used for exclusion of dead cells. Both PI GFP⁺ and PI GFP populations were sorted to retrieve living cells from Pdx-l/VP16 transduced cells. Both GFP⁺ and GFP^" single cells were retrieved into 96 well dishes. Group 1 cells were also dispersed into single cells and served as controls.

To detect expression of genes of interest from single cells, nested PCR was performed. Multiplex single-cell RT-PCR analysis was performed according to a previously described method with some modifications (Miyamoto et al., 2002, Dev. Cell, 3:137-147). Briefly, reverse-transcription was performed followed by a first round of PCR. Sorted single cells were deposited into 96-well U-bottom plates (BD) with 7.5 μl lysis and RT buffer containing gene-specific reverse primers for CK 19, Neuro D, Pdx-1, Insulin 1, Insulin 2, Glucagon, Somatostatin, PP, and HPRT. The lysis buffer contained the following: [Ix First Strand Buffer (Invitrogen), 10 rnM DTT (Invitrogen), 1 mM dNTPs (New England BioLabs, Ipswich, MA), 0.5% Triton™ X-IOO (Sigma), 0.1% bovine serum albumin, 10 U/μl M-MLV Reverse Transcriptase (Invitrogen), 0.1 U/μl RNase inhibitor (Invitrogen), 0.4 pM reverse primers]. Retrieved cells were lysed by pipetting several times in the plate, and then cell lysates were transferred to a 96-well optical Reaction Plate with Barcode (Applied Biosystem, Foster City, CA). After incubation at 37 ⁰C for 90 minutes, the samples were incubated for generation of cDNA by reverse transcription and at 94 ⁰C for 30 seconds to inactivate the enzyme. The first- round PCR was carried out in the same plate by addition of premixed PCR buffer containing the gene-specific forward primers [Ix GeneAmp® PCR Gold Buffer (Applied Biosystem), 2.5 mM MgCl₂, AmpliTaq™ Gold 0.1 U/μl, 0.1 pM forward primers (Table I)] . The total volume of each of the 1st PCR reactions was 20 μl, and PCR was performed with the following parameters: 36 cycles of 30 seconds at 94 ⁰C, 90 seconds at 60 ⁰C, 90 seconds at 72 ⁰C. For nested PCR, the 2nd-round PCR was performed as follows: 1 μl of the first-round PCR reactions was plated into new PCR plates, which contained mixed buffer [Ix GeneAmp® PCR Gold Buffer (Applied Biosystem), 2.5mM MgCl₂, AmpliTaq™ Gold (Applied Biosystem) 0.1 U/μl, and 0.25 pM forward primers (Table I)] . For the 2nd-round PCR, each gene was analyzed separately using nested gene-specific primers (Table 1) with the following parameters: 35 cycles of 36 seconds at 94 ⁰C, 90 seconds at 60 ⁰C, 90 seconds at 72 ⁰C. Aliquots of 2nd-round PCR products were subjected to gel electrophoresis. Two hundred pg of total RNA isolated from mouse islets was used as a positive control.

Twenty three percent Of GFP⁺ cells expressed both the insulin 1 and 2 genes. Importantly, these insulin 1 and 2 expressing cells expressed other marker genes including islet genes (glucagon, somatostatin, and PP) and CKl 9 (Table 2). In Groups 1 and 2, no cells exhibited insulin gene expression (Tables 3-4). Greater than 94% of Pdx-l/VP16 transduced cells expressed CK19, and all cells that expressed insulin genes also expressed CKl 9.

Table 3. GFP cells from Pdx-l/VP16 transduction cells (Group 2)

Table 4. IHBECs in DM for 7 days (Group 1)

Example 7. Transdifferentiation by Inhibition of Notch Signaling To enhance transdifferentiation, the Notch signaling pathway is inhibited in cells with and without transduction using adenoviral (AdV) vectors expressing key transcription factors.

The notch ligands (jagged 1 and 2, delta-like 1-3) of cells interact with their receptors Notch 1-3 on adjacent cells. This interaction leads to the cleavage of Notch by a γ-secretase activity, such that the notch intracellular domain (NICD) is released and translocates to the nucleus where it regulates transcription. Several γ-secretase inhibitors have been used to block Notch signaling: γ-secretase inhibitor II (Calbiochem, 50 μM) in DMSO was effective in epithelial cells when added to cells for 6 hours (Chojnacki et al, 2003, J. Neurosci., 23:1730-41), and γ-secretase inhibitor X (Calbiochem, 4 uM) was effective over 60 hours. DAPT (N-[N-(3,5-Difluorophenacetyl-L-alanyl)]-S- phenylglycine t-butyl ester; 10 or 20 μM, Calbiochem) (Niimi et al., 2007, J. Cell Biol., 176:695-707) was used from before induction of differentiation and throughout the process with mesenchymal stem cells (Vujovic et al., 2007, Cell Prolif, 40:185-195). Since inhibiting γ-secretase activity can affect processing of other molecules, anti-Notch antibody is also used for neutralization.

Adult liver expresses all of the necessary components for active Notch signaling, but in the pancreas, the Notch signaling may be quiescent (Jensen et al., 2000, Nat. Genet., 24:36-44; Miyamoto et al., 2003, Cancer Cell, 3:565-576). Environmentally- induced alterations in Notch signaling in IHBECs in vitro can lead to accelerated pancreatic endocrine differentiation. Either γ-secretase inhibitor II or DAPT are added to the cell culture at Step 4 of transdifferentiation (Fig. 1). The efficacy of the inhibition is determined in cells at 9 days in DM by real-time PCR for mRNA of the Notch family member hesl (a key downstream mediator of Notch signaling as a measure of effectiveness) and for insulin 1 mRNA. Western blotting is carried out for Hesl and an ELISA assay is used to measure insulin content in transdifferentiated cells. Further, SOX9 expression is determined by antibody staining.

Example 8. Transdifferentiation by Inhibition of HNF 6 Expression To enhance transdifferentiation, expression of the HNF6 gene is knocked down with and without transduction using adenoviral (AdV) vectors expressing key transcription factors.

HNF6 is an early transcription factor controlling the initiation of BEC differentiation (Clotman et al, 2002, Development, 129:1819-28). Knockdown of HNF6 with siRNA enhances the transdifferentiation of IHBECs. That transdifferentiation of

IHBECs is enhanced by short-term down-regulation of HNF6 expression, because the knockdown enhances dedifferentiation.

To suppress the expression of HNF6, HNF6 siRNAs are used to dedifferentiate

IHBECs into cells that can then become insulin producing-cells. Both FITC and Cy ™ -3 fluorophore-labeled Luciferase (Luc) siRNA GL2 duplexes (Dharmacon Inc.) are used for the assessment of siRNA delivery into IHBECs. The sense and antisense strands of the 2'-ACE RNAs are: 5'-[FLUOROPHORE]-CGUACGCGGAAUACUUCGAdTdT-3'

(SEQ ID NO: J (sense), 5'-CGTACGCGGAATACTTCGA-S' (SEQ ID NO: J

(antisense). The mouse HNF6-siRNA is obtained (one cut domain, family member 1; NM_008262A; Abgene, Rochester, NY) as an annealed, predesigned siRNA. Samples are divided into 3 groups: (1) untreated controls, (2) HNF6-siRNA transfected, and

(3) Luciferase-siRNA (Luc-siRNA) transfected IHBECs as a nonspecific siRNA controls.

For in vitro delivery, the siRNA is packaged using Lipofectamine™ 2000 (Invitrogen) at a DNA-to-liposome (μg/μL) ratio of 1 :3, following the manufacturer's recommendations in a delivery volume of 100 μL. Dose titrations are optimized. Samples are transfected and incubated at 37 ⁰C, 5% CO₂ for 48 hours. The Cy3-labeled Luc-siRNA is used for fluorescent imaging of siRNA- transfected IHBECs in vitro, luciferase assay, and as a nonspecific siRNA control for quantitative RT-PCR. The FITC-labeled Luc-siRNA is used for FACS analysis of single- cell IHBECs suspensions to determine transfection efficiency and duration of the efficacy.

Example 9. Combination of Notch Inhibitor and HNF6 siRNA

A combination of inhibition of Notch signaling pathway by γ-secretase inhibitor or DAPT and knockdown of HNF 6 by siRNA manipulations is used for transdifferentiation with and without using adenoviral (AdV) vectors expressing key transcription factors. The efficacy of the combination is determined in cells at 9 days in DM by real-time PCR for insulin 1 mRNA and an ELISA assay to measure insulin content.

Example 10. Use of Adeno-Associated Viruses (AAV) Vectors

AAV, a non-pathogenic parvovirus, is another attractive vector for transduction of transcription factors. Human trials with AAVs are in progress (Williams, 2007, MoI. Ther., 15:2053-54). AAV serotype 8 vectors (which have low immunogenicity and high transduction efficiency, long-term and stable expression of the delivered transgene, and persistence in the liver, with transduction efficiencies comparable to that of AdVs) (Nakai et al., 2005, J. Virol., 79: 214-224) are used for transduction of TFs similar to AdVs in the methods described above. AAV-NeuroD and AAV- Pdxl/VP16 are produced, purified using cesium chloride density centrifugation, and titered by quantitative dot blot analysis. AAV-I, 2, 5, 6, and others are also produced and tested.

Example 11. Enhanced Differentiation to Insulin-Producing Cells in vivo

Enhanced differentiation to insulin-producing cells is fostered by the in vivo environment. The transplanted cells are characterized for function, structure, mass and their capacity to reverse diabetes. Differentiated or undifferentiated insulin-secreting IHBECs are transplanted into STZ-diabetic ICR-scid mouse recipients to determine their capacity for further maturation. ICR/scid mice are given STZ (180-220 mg/Kg BW - freshly dissolved in citrate buffer, pH 4.5) with a single intra-peritoneal injection. Once non-transplanted animals reach blood glucose levels greater than 350 mg/dL for at least 10 days, the IHBECs are transplanted.

Aliquots of IHBECs are aspirated into a 200 μl pipette tip connected to a 1 ml Hamilton syringe and allowed to sediment. The tip is then connected to PE-50 polyethylene tubing, into which the IHBECs are injected. The tubing is then folded and centrifuged for 1.5 minutes at 1200 rpm, and the tip again connected to the Hamilton syringe. With the mouse under light anesthesia, the left kidney is exposed via a flank incision. A capsulotomy is performed in the lower kidney pole, the tubing inserted beneath the capsule and gently advanced to the upper pole. IHBECs are gently injected and the capsulotomy cauterized with a disposable low-temperature cautery device.

Initially, the cells are transplanted to the renal subcapsular space, and cells are also transplanted with a basement membrane matrix (e.g., Matrigel™ basement membrane matrix) (Bonner-Weir et al, 2000, Proc. Natl. Acad. Sci. USA, 9:7999-8004) or a thermoreversible gelation polymer (TGP) matrix subcutaneously or into the epididymal fat pad. TGP has been used in clinical drug-response testing in vitro (Isogai et al., 2007, Surgery, 142:741-748; Matsuoka et al., 2006, Surgery 140:387-395). In vitro aggregates of differentiated cells keep their 3D structure in TGP as a gel and are transplanted as such. The cells remain in the gel in vivo, which facilitates retrieval. Then for analysis of the retrieved cells, the gel is liquefied at a temperature below 20 ⁰C.

The number of the cells used for transplants is 5,000,000 for non-transduced cells and 50,000 for transduced cells, respectively. After transplantation, body weight and glucose levels are determined once a week. Blood is obtained from the snipped tail and glucose is measured with a portable glucose meter (Precision QID). Grafts are initially evaluated at 3 weeks, and also evaluated at later time points.

Initial evaluation is performed with C-peptide immunostaining. Graft cell composition, insulin-producing cell mass, insulin content, transmission EM for examination of granules, proinsulin/insulin ratio, and glucose-stimulated insulin secretion (GSIS) with IP glucose tolerance tests (IPGTT) are performed. Cell composition is determined by immunostaining of sections; at the three time points, grafts will be assessed for insulin- producing cell mass by C-peptide staining, non-insulin-producing-cell mass using a cocktail of antibodies against glucagon, somatostatin and PP, and ductal phenotype with CK-7 and/or -19 staining. Quantitation is performed with point counting, as described previously (Montana et al, 1993, J. Clin. Invest., 91 :780-787). AU of these measurements are compared with measurements of normal mouse islets in vivo and in vitro.

Example 12. Transplantation of IHBECs Inhibited for Notch signaling and HNF6

IHBECs are manipulated to inhibit the Notch signaling pathway (as described in Example 7) and knockdown HNF6 (as described in Example 8) in vitro. 50,000 to 5,000,000 cells are then transplanted into mice as described in Example 11. Blood glucose levels and weights of the mice are monitored.

Example 13. Transplantation of IHBECs with Multiple Manipulations

Insulin-producing cells for transplantation are generated using a combination of AAVs-NeuroD or Pdx- 1 /VP 16 with inhibition of the Notch signaling pathway and/or knockdown of HNF 6 gene. The cells are transplanted following transdifferentiation with or without purification of insulin-producing cells by FACS using the anti-β-cell antibody IC2 (Brogren et al., 1986, Diabetologia, 29:330-333; Aaen et al., 1990, Diabetes, 39:697- 701; Moore et al., 2001, Diabetes, 50:2231-36). Transduced cells, which produce insulin, are retrieved on an ARIA cell sorter with the Summit software (BD). The cells are dispersed, and Propidium Iodide (PI) is used for exclusion of dead cells. Both PI IC2⁺ and PI^~IC2^~ populations are gated to retrieve living cells from the transduced cell population for transplantation.

Example 14. Transplantation of IHBECs Encapsulated in Alginate Microbeads

IHBECs and IHBEC-derived cells are also placed into alginate capsules and transplanted into the peritoneal cavity. For encapsulation of IHBECs, simple alginate microcapsules (900-1,100 mm in diameter) with no poly lysine coating are used, which have been used previously for successful transplants in mice (Duvivier-Kali et al., 2001, Diabetes, 50: 1698-1705). IHBEC derived cells are resuspended in 1.2 ml of 1.5% (w/v) ultrapurified sodium alginate, and the droplets are made with an electrostatic droplet generator. The gel beads are formed by crosslinking with BaCl₂ (10 mM). Encapsulated IHBECs are suspended in 5 ml of sterile phosphate-buffered saline and transplanted into the peritoneal cavity of diabetic mice through a small ventral incision using a 50 ml pipette. Mice are monitored with twice weekly measurement of non-fasting blood glucose and body weight.

Example 15. Mixing of IHBECs with Fetal Pancreatic Cells

Differentiation of IHBEC derived cells can be enhanced by growth and differentiation signals from embryonic pancreas. IHBEC-derived cells obtained from mouse insulin promoter GFP mice (MIP-GFP) are mixed with pieces of mouse pancreas from stage El 5. The GFP allows for identification of insulin-producing cells derived from the IHBECs. Transdifferentiation is monitored in vitro and following transplantation into STZ-diabetic ICR-scid mice.

OTHER EMBODIMENTS A number of embodiments of the disclosure have been described. Nevertheless, it will be understood that various modifications may be made without departing from the spirit and scope of the disclosure. Accordingly, other embodiments are within the scope of the following claims.

Claims

WHAT IS CLAIMED IS:

1. A method of generating a cell population comprising insulin-producing cells, the method comprising:

(a) obtaining a intrahepatic biliary epithelial cell;

(b) culturing the cell in a first medium comprising insulin and serum for a first time period, and then culturing the cells in a second medium lacking insulin and serum for a second time period, thereby producing an population of intrahepatic biliary epithelial cells; and

(c) contacting the population of intrahepatic biliary epithelial cells with at least one of a nucleic acid encoding a transcription factor, an inhibitor of Notch signaling, an inhibitor of hepatocyte nuclear factor 6 (HNF6) expression, a fetal pancreatic cell, or a solution contacted by cultured fetal pancreatic cells for a time sufficient that at least a portion of the cells secrete insulin, thereby generating a cell population comprising insulin-producing cells.

2. The method of claim 1, wherein steps (b) or (c), or (b) and (c) comprise culturing the cell in a three-dimensional scaffold.

3. The method of claim 2, wherein the three-dimensional scaffold is a collagen gel.

4. The method of claim 1 , wherein the transcription factor is selected from the group consisting of Pancreatic and duodenal homeobox-1 (Pdx-1), neurogenic differentiation 1 (NeuroD), Pdx-1 carrying the VP 16 transactivator domain (Pdx-l/VP16), Neurogenin3 (Ngn3), NK6 transcription factor related, locus 1 (Nkxβ.l), and pancreas-specific transcription factor lα (Ptflα).

5. The method of claim 4, wherein the transcription factor is Pdx-1, NeuroD, or Pdx-l/VP16.

6. The method of claim 1, wherein the nucleic acid is contained within a viral vector.

7. The method of claim 6, wherein the viral vector is an adenovirus or adeno- associated virus vector.

8. The method of claim 1, wherein the inhibitor of Notch signaling is a gamma- secretase inhibitor.

9. The method of claim 1, wherein the inhibitor of HNF 6 expression is a nucleic acid that inhibits HNF6 expression.

10. The method of claim 1, wherein the insulin producing cell secretes at least 0.1 ng/ml insulin per 10⁵ cells over 48 hours in culture.

11. The method of claim 1 , wherein the insulin producing cell secretes insulin in response to glucose.

12. The method of claim 1, further comprising introducing the insulin producing cell into a subject.

13. The method of claim 12, wherein the subject is a human.

14. The method of claim 12, wherein the intrahepatic biliary epithelial cell is obtained from the subject.

15. The method of claim 1, wherein the enriched population of intrahepatic biliary epithelial cells is contacted with an inhibitor of Notch signaling and an inhibitor of FINF6 expression.

16. An enriched cell population comprising an insulin producing cell generated by the methods of any of claims 1-14.

17. The enriched cell population of claim 16, wherein at least 60% of the cells in the population are insulin producing cells.

18. An insulin-producing cell generated by the method of any of claims 1-14.

19. A method of treating a subject with or at risk for developing diabetes, the method comprising selecting a subject who has or is at risk for developing diabetes and transplanting a cell generated by the method of any of claims 1-14 into the subject.