- Email: [email protected]
In vitro induction of adult hepatic progenitor cells into insulin-producing cells
BBRC Biochemical and Biophysical Research Communications 318 (2004) 625–630 www.elsevier.com/locate/ybbrc
In vitro induction of adult hepatic progenitor cells into insulin-producing cells Natsuki Nakajima-Nagata,a Tomonori Sakurai,b Toshihiro Mitaka,c Tomoya Katakai,a Eiji Yamato,d Jun-ichi Miyazaki,d Yasuhiko Tabata,e Manabu Sugai,a and Akira Shimizua,f,* a Center for Molecular Biology and Genetics, Kyoto University, 53 Kawahara-cho Shogoin, Sakyo-ku, Kyoto 606-8507, Japan Department of Organ Reconstruction, Institute for Frontier Medical Sciences, Kyoto University, 53 Kawahara-cho Shogoin, Sakyo-ku, Kyoto 606-8507, Japan c Department of Pathology, Cancer Research Institute, Sapporo Medical University School of Medicine, Chuo-ku, S1W17 Sapporo, Japan d Division of Stem Cell Regulation Research, G6, Osaka University Medical School, Suita 565-0871, Japan Department of Biomaterials, Institute for Frontier Medical Sciences, Kyoto University, 53 Kawahara-cho Shogoin, Sakyo-ku, Kyoto 606-8507, Japan f Translational Research Center, Kyoto University Hospital, 54 Kawahara-cho Shogoin, Sakyo-ku, Kyoto 606-8507, Japan b
e
Received 16 March 2004 Available online 24 April 2004
Abstract Organ-specific stem cells are the natural progenitors in tissue regeneration and possess plasticity to differentiate into specialized cells in adult tissues. Small hepatocytes (SHCs) identified in the adult liver are one such cell type. Here we show that SHCs, which are capable of self-renewal and differentiation into hepatocytes, can be induced to generate insulin-producing cells under appropriate culture conditions. These differentiated cells express pancreatic b cell differentiation-related transcripts and hepatocyte differentiation-related transcripts, as shown by reverse-transcription PCR/nested PCR. In addition, enforced expression of the homeodomain transcription factor Pdx1 in these cells contributes to enhancement of insulin release in response to insulin secretagogues. These results indicate that the SHCs described here have the ability to differentiate into insulin-producing cells, and further support the idea that engineering to generate insulin-secreting cells could provide a useful resource for future therapies for diabetes mellitus. Ó 2004 Elsevier Inc. All rights reserved. Keywords: Pancreatic b cell; Pancreatic duodenal homeobox-1; Small hepatocytes; Glucagon-like peptide-1; Reverse transcription
Cell-fate-determination studies in mice have shown that the cells constituting the ventral portion of the foregut in early developmental stages give rise to both the liver and the ventral rudiment of the pancreas at a later stage. A gain-of-function study demonstrated that ectopic expression of Pdx-1 resulted in a substantial increase in hepatic insulin content and an increase of the plasma insulin level [1]. These data also support the idea that it may be possible to induce insulin-producing cells from adult liver cells. In the present study, we focused on the potential of small hepatocytes (SHCs) to differentiate into insulin* Corresponding author. Fax: +81-75-751-4190. E-mail address: [email protected] (A. Shimizu).
0006-291X/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2004.04.059
producing cells and the actual role of Pdx-1 in insulin production and release using this cell population. SHCs have been reported to be hepatocyte progenitor cells, and are isolated from the adult liver within the fraction of non-parenchymal cells (NPCs). These cells can survive and grow for more than 5 months, and interact with hepatic NPCs to reconstruct three-dimensional hepatic organoid structures under appropriate culture conditions in vitro [2,3]. Here we describe an experimental strategy for developing insulin-producing cells that release insulin in response to some insulin secretagogues. These cells were produced by inducing SHCs to differentiate and forcing them to express Pdx-1. Our results provide evidence that SHCs can be induced to generate insulin-producing cells
626
N. Nakajima-Nagata et al. / Biochemical and Biophysical Research Communications 318 (2004) 625–630
and that Pdx-1 might contribute to the regulation of insulin release.
Experimental procedures Isolation of non-parenchymal cell fraction containing small hepatocytes. We isolated liver SHCs from male adult Sprague–Dawley rats (100–150 g body weight, Shimizu Animal, Kyoto, Japan) according to the modified liver perfusion method reported previously [2] (Step 1 in Fig. 1). In Step 2, 5 104 viable isolated cells/cm2 were plated on dishes and cultured in HGM medium, which consisted of DMEM (GibcoBRL) supplemented with 10% FBS, 20 mM Hepes, 25 mM NaHCO3 , 30 mg/L L -proline, 10 mg/L insulin–transferrin–sodium selenite (ITS) (Roche, Mannheim, Germany), 20 ng/mL EGF (Gibco-BRL), 0.004 g/L trace metals (ZnCl2 , ZnSO4 , CuSO4 , and MnSO4 ), 10 mM nicotinamide, 1 mM ascorbic acid 2-phosphate, 107 M dexamethasone (Sigma), and antibiotics. The medium was replaced with fresh medium 2 days after the cells were plated, and the cells were cultured for 3 more days. In Step 3, 1% dimethyl sulfoxide (DMSO; Wako Chemical, Tokyo, Japan) was added to the culture medium. In Step 4, the cells were cultured in DMEM/F12K (1:1) medium supplemented with 0.2% FBS, 20 ng/mL HGF (PeproTech EC, London, UK), 10 ng/mL EGF, 25 mM NaHCO3 , 30 mg/L L -proline, 10 mg/L insulin–transferrin–sodium selenite (ITS) (Roche), 10 mM nicotinamide, 107 M dexamethasone, and antibiotics. As a control, other SHCs continued to be cultured under the conditions of Step 3. Plasmid construction for adenoviral infection. We used an E1-defective adenovirus (Ad) vector designated AdAG-Pdx1-IRES-GFP containing an expression cassette that bicistronically expressed mouse Pdx-1 protein coding sequences and green fluorescent protein (GFP) in the adenovirus genome of pALC under the control of the modified chicken b-actin (AG) promoter. As a control vector, we also used adenovirus vector designed to contain the expression cassette of IRESGFP. Nearly 15–20% of the isolated SHCs were shown to express GFP after infection with one of these vectors at MOI 30. Total RNA isolation, synthesis of complementary DNA, and reversetranscription polymerase chain reaction. Total cellular RNA was extracted from SHCs by the single-step method using TRIZOL reagent (Invitrogen, Carlsbud, CA). The extracted total RNA was reversetranscribed into cDNA with Superscript-II Reverse Transcriptase (Invitrogen) using standard procedures. RT-PCR or nested PCR was performed using Taq polymerase (TAKARA, Shiga, Japan) to monitor the transcription of genes related to the genesis of insulin-producing cells. The primers used were complementary to the mRNA sequences of the genes of interest and are listed in Table 1. All the primers were designed to cross exon–exon boundaries in order to exclude the possibility that genomic DNA was amplified. The mouse Pdx-1 primers were designed to detect Pdx-1 protein-coding sequence transgene transcripts specifically. The rat Pdx-1 PCR primer pair could not detect the RTPCR products from mouse samples. The PCR amplification reaction was performed using different amounts of cDNA to ensure that a linear response of the RT-PCR amplification process was achieved. b-Actin expression was used to normalize for the amount of input template cDNA in order to analyze relative gene expression. Measurement of insulin content The cells were washed four times with PBS and then treated with acid–ethanol (0.1 N hydrochloric acid in absolute ethanol) at 4 °C overnight. The clear supernatants were used to assay the intracellular insulin content and the values obtained were normalized relative to the total protein content. The insulin content was measured by an enzymelinked immunosorbent assay (Insulin-ELISA kit, Shibayagi, Gunma, Japan). Total protein content was measured by the BCA method (Pierce, Rockford, IL).
Fig. 1. Generation of insulin-producing cells from SHCs. (A) General outline of the culture system and Pdx-1 gene transfection protocol for inducing the differentiation of SHCs. (B) The experimental schedule starting from the day SHCs were isolated (day 0). On day 9, adenovirus vector Pdx-1-IRES-GFP or IRES-GFP was transfected into the Group 1 cells. (C) Phase-contrast photographs of the cells on day 7. (D) Cells of Group 1 (+Pdx-1) were subjected to immunofluorescence staining for insulin (red) at the end of Step 4 (day 15) (left panel). The merged image of insulin and DAPI (blue) staining is shown in the right panel. DAPI was used for nuclear staining. (E) Group 2 cells were also stained for insulin (red) (left panel), and the merged image of insulin and DAPI (blue) staining is shown in the right panel. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this paper.)
N. Nakajima-Nagata et al. / Biochemical and Biophysical Research Communications 318 (2004) 625–630
627
Table 1 PCR primers used in this study Genes
Primer sequences
Accession No.
RAT INSULIN I
F 50 -ATGGCCCTGTGGATGCGCTT-30 R 50 -TAGTTGCAGTAGTTCTCCAGCT-30 NEST F50 -GGGAGCCCAAGCCTGCCCAGGC-30 NEST R50 -CAATGCCACGCTTCTGCCGGGC-30
J00747
RAT INSULIN II
F 50 -ACAGTCGGAAACCATCAGCAA-30 R 50 -GCTGGTGCAGCACTGATCCACG-30 NEST F 50 -TGCCCCTGCTGGCCCTGCTC-30
V01243
MOUSE PDX-1
50 -CCGGCTCTGACTGACCGCGTTAC-30 50 -TGCTCCAGTGATCCCAGCGAGC-30
AK020261
RAT PDX-1
F 50 -TACGCGGCCACACAGCTCTACAAGGAC R 50 -CCACTTCATGCGACGGTTTTGGAACCAGA-30 NEST F 50 -TCTCCTCCGGTTCTGCTGCGTATGCAC-30
NM022852
MOUSE NESTIN
F 50 -CTGCTGGAGGCTGAGAACT-30 R 50 -ATAGGTGGGATGGGAGTGC-30
M34384
RAT Isl 1
F 50 -GCCCGCTCTAAGGTGTACCACATC-30 R 50 -TCATGATGCTGCGTTTCTTGTCCTT-30 NEST F 50 -AAGGTGTACCACATCGAATGTTTCC-30
X53258
MOUSE PTF1
F 50 -GCAGCTGCGACAAGCCGCTAA-30 R 50 -CCTCTGGGGTCCACACTTTAGCTGTACGGA-30 NEST F 50 -GCGCCGCATGCAGTCCATCAACGA-30
AF298116
RAT NEUROD
F 50 -GCTTGGCCAAGAACTATATCTGGGC-30 R 50 -GTGAAAGATGGCATTAAGCTGGGC-30 NEST R 50 -CGAAAGACATAATATTGTCTATGGGG-30
NM019218
RAT PAX 6
F 50 -ACCAACGACAATATACCCAGTGTGTCATC-30 R 50 -GTGTTGCTGGCCTGTCTTCTCTGGTTCC-30 NEST R 50 -TGGCCTGTCTTCTCTGGTTCCTC-30
NM013001
RAT HNF1A
F 50 -CCCCATCTGAAGGTGCCAACCTCAAC-30 R 50 -CCGCTTGATCTTGCCTGGGTCACTC-30
NM012669
RAT HNF1B
F 50 -CCAGAAGCGAGCTGCCCTGTACAC-30 R 50 -AGGCTGCTAGCCACACTGTTGATGAC-30
NM013103
RAT HNF4A
F 50 -CCCCATCTGAAGGTGCCAACCTCAAC-30 R 50 -CCGCTTGATCTTGCCTGGGTCACTC-30
NM022180
RAT CK8
F 50 -CAACTTCCTCCGGCAGATCCAT-30 R 50 -ATCCTTCACGGCCAGTTCCCC-30
NW043527
RAT AFP
F 50 -GCGATGCGTTGGCTGCAATGAAGG-30 R 50 -GTTGTCAGCTTTGCAGCATGCTGGAAC-30
NM012493
RAT c-MET
F 50 -AGAGAACATTGGGGTGGCTGAAA-30 R 50 -ATAGAGTGGAAGCAAGCAGTCTC-30
U65007
MOUSE NGN3
F 50 -TACCTAGGGACTGCTCCGAAGCAGAAGTGG-30 R 50 -ACCAGTGCTCCCGGGAGCAGATAGGATG-30 NEST F 50 -GGCCCAAGAGCGAGTTGGCACTCAGCA-30 NEST R 50 -TCCAATGAGGCCGTGGGGCTCAGGTT-30
Y09167
Measurements of insulin secretion in response to insulin secretagogues. Insulin secretion was measured using static incubation in Krebs–Ringer bicarbonate buffer (KRBB) supplemented with 5 mM glucose and 0.2% FBS. The cells were plated in 6-well culture plates (35mm diameter; IWAKI, Tokyo, Japan), cultured, and washed four times with KRBB on day 15. They were then incubated in KRBB with high glucose (25 mM) and 10 nM glucagon-like peptide 1(GLP-1 [7–37]) (Sigma) or 45 mM KCl and 0.2 mM tolbutamide (Sigma), or low glucose (5 mM) in KRBB as a control. The concentration of the insulin secreted into the buffer solution was measured as described above.
Immunocytochemistry. Cells were incubated in the Step 4 medium described above supplemented with Golgestop solution (Pharmingen) plus the secretagogues described above for 4 h in a CO2 incubator. Immunocytochemistry was carried out using standard protocols in which the cells were sequentially incubated with mouse anti-preproinsulin antibody (Biogenesis, Poole, UK) diluted 1:1000 in PBS, biotin-labeled anti-mouse IgG antibody (Nichirei, Tokyo, Japan), and PE-conjugated streptavidin (Molecular Probes), and assessed by examining the cells with an Olympus AX80 microscope.
628
N. Nakajima-Nagata et al. / Biochemical and Biophysical Research Communications 318 (2004) 625–630
Statistical analysis. Statistical comparisons were made by analysis of variance and, when appropriate, using the unpaired Student’s t test, with P < 0:05. Experimental results were expressed as the means standard deviation of the mean. All experiments were repeated at least three times independently.
Results Generation of insulin-producing cells from SHCs We used an experimental strategy (Figs. 1A and B) whereby in Step 1 an SHC-rich fraction was separated from adult rat liver as previously described [2]. In the early steps of culturing (Steps 2 and 3), we detected nestin-positive SHCs (Figs. 1C and 2), which are progenitors of neurons and pancreatic endocrine cells [4,5]. This strategy was very similar to that used for generating insulin-producing cells from ES cells [6]. Next, we tried to determine the optimal conditions for the differentiation of insulin-producing cells from nestin-positive cells (Step 4). Among the conditions we examined, culturing the cells in the presence of HGF, EGF, nicotinamide, and dexamethasone in 0.2% FBS-containing medium effectively induced the production of insulin-producing cells. To further examine the function of Pdx-1 in insulin
Fig. 2. Gene expression profiles of cultured SHCs. We performed RTPCR to detect mouse PDX-1, nestin, Id2, HNF1a, HNF1b, HNF4a, AFP, CK8, c-Met, and b-actin mRNA transcripts, and RT-PCR followed by nested PCR to detect insulin 1, insulin 2, rat Pdx-1, neurogenin, NeuroD, Isl 1, Pax6, and PTF1 mRNA transcripts. The cells of Group 1 showed expression of these genes on day 0 (Step 1), day 7 (Step 3), day 9 (Step 4), and day 15 (Step 4) in culture. We assessed the effects of Pdx-1 overexpression in the cells of Group 1 on day 15 (shown by “+Pdx-1” and the control “)Pdx-1”). As an additional control, the cells of Group 2 were also assessed on day 15. Gene expression studies in all groups were repeated at least three times with similar results.
production, we employed an adenovirus-mediated approach for introducing Pdx-1 into nestin-positive SHCs. Mouse Pdx-1 (mPdx-1) cDNA was cloned into the Ad AG-IRES-GFP vector, which allowed Pdx-1 expression to be assessed by monitoring green fluorescent protein (GFP). The Group1 progenitor cells were infected with the viral vector on day 9 and analyzed for the expression of insulin on day 15 by immunocytochemistry (Figs. 1D and E). There was no obvious difference of GFP expression or cell morphology between the Pdx-1transduced and control populations (data not shown). Gene expression profiles of cultured SHCs To further analyze the molecular events that occurred in SHCs during the series of culturing steps, the gene expression profiles were determined by RT-PCR (Fig. 2). SHCs at Step 1 expressed albumin, a marker of mature hepatocytes, c-Met, expressed in mature hepatocytes and b cells, and neurogenin 3, a marker of islet progenitor cells. At the end of Step 3, albumin and neurogenin 3 expression disappeared and the expression of nestin was observed (Fig. 2, lanes 1 and 2). On day 15, insulin 1 expression was only detected in cultured cells in Group 1. This observation suggested the importance of Step 4 in the differentiation of SHCs into insulin-producing cells. Like insulin 1, a-fetoprotein (AFP), a marker of immature hepatocytes, was expressed only in Group 1. However, the importance of AFP in insulin 1 production is not clear. AFP expression seems to indicate the re-differentiation of hepatocytes under these culture conditions, because AFP expression is not observed in the pancreas (data not shown). We found that insulin 2 expression was completely dependent on the expression of Pdx-1. This observation is consistent with data reported previously [1]. Forced expression of Pdx-1 also induced the expression of endogenous Pdx-1 and NeuroD. These results indicate that Pdx-1 and/or NeuroD are required for insulin 2 expression. Islet1 and Pax6 were detected after 9 days of culturing under both conditions (Group 1 and Group 2). In contrast, no expression of PTF1, a marker of exocrine cells, was observed in any of the cells analyzed. These data are consistent with the idea that b cell differentiation occurred under these culture conditions. The expression levels of HNF1a, 1b, and 4a, whose mutations cause diabetes mellitus, did not vary among the cells examined. Insulin content in differentiated SHCs The intracellular insulin content in the differentiating SHCs was measured and normalized relative to the total protein content on days 0, 5, 9, 12, and 15 (Figs. 1 and 3). As a control, the cells from Group 2 were also assayed. Before day 9, only very low levels of insulin content were detected (data not shown). After 9 days of culturing, the
N. Nakajima-Nagata et al. / Biochemical and Biophysical Research Communications 318 (2004) 625–630
629
Insulin release by differentiated SHCs
Fig. 3. Intracellular insulin content in differentiating SHCs. The insulin levels in the cells (n ¼ 4) of Group 1 (consisting of subgroups of both Pdx-1-expressing cells (+Pdx-1: black columns) and non-Pdx-1-expressing cells ()Pdx-1: gray columns)), and control cells (Group 2: white columns) were assessed on day 9, day 12, and day 15. Three separate experiments were performed to determine the insulin level of each group. P > 0:05.
insulin level of the cells from Group 1 was significantly increased in comparison with the lower insulin level in control cells (Fig. 3). Pdx-1 had no effect on the insulin content, indicating that the major insulin source at this time was the insulin 1 gene (Fig. 3). Overall, the results of the insulin content analysis further support the idea that Step 4 is the most important condition for inducing insulin-producing cells.
To determine whether the ability of differentiated SHCs to secrete insulin was affected by Pdx-1 or not, we examined insulin release in response to several secretagogues (Fig. 4). According to the known mechanism by which glucose stimulates insulin release in vivo, transport of glucose into the cells results in adenosine triphosphate (ATP) production, which enhances Ca2þ influx and subsequent Ca2þ release from the intracellular store. The elevation of intracellular Ca2þ is coupled to multiple phosphorylation events modulated by the protein kinase A (PKA) and protein kinase C (PKC) cascades. These pathways are known to influence insulin release. The roles of these two pathways of insulin secretion were tested (Fig. 4). GLP 1, an activator of the cAMP–PKA pathway, stimulated insulin secretion from both Pdx-1-transduced cells (+Pdx-1) and control cells ()Pdx-1), but the induction was 2.5-fold higher in Pdx1-transduced cells (Fig. 4, left panels). Tolbutamide (sulfonylurea) and Kþ , inducers of Ca2þ influx, had the same effect on insulin secretion as seen in the left panels of Fig. 4 (central panels). In contrast, only a slight effect on insulin secretion was observed in the cultures incubated with a low glucose concentration for 30 min without agonist (Fig. 4, right panels). These observations support the idea of the functional similarity of these cells to b cells. In all the cultures examined, we observed about 2-fold higher insulin release (1.6- or 2.4fold higher in these stimulation tests) from +Pdx-1 cells than from )Pdx-1 cells in spite of almost equal insulin contents in these cells (see Fig. 3). These data indicate that the universal components of insulin release can be affected by Pdx-1 in in vitro differentiated SHCs.
Fig. 4. Insulin release in response to insulin secretagogues. The effects of various agonists on insulin secretion from insulin-producing cells of Group 1 (either Pdx-1-expressing cells (+Pdx-1) or non-Pdx-1-expressing cells ()Pdx-1)), and from Group 2 cells were tested on day 15. Insulin release from these cells was observed in response to stimulation by 10 nM GLP-1 with high-dose glucose (25 mM) (left panels). The insulin release in response to a high physiological concentration (45 mM) of Kþ and 0.2 mM tolbutamide with low-dose glucose (5 mM) was also examined (central panels). As a control, insulin release in response to low-dose glucose (5 mM) in basal buffer (KRBB solution) was also examined (right panels). The integral amount of insulin release from 0 to 5 min (open columns), 0–10 min (gray columns), and 0–30 min (closed columns) is shown in the upper panels.
630
N. Nakajima-Nagata et al. / Biochemical and Biophysical Research Communications 318 (2004) 625–630
Discussion We report here that SHCs could efficiently generate insulin-producing cells having significantly elevated insulin content in their cytoplasm. The generation of insulin-secreting cells from adult stem cells is an important advance for use in future therapies for diabetes mellitus. Various secretagogues were used for functional analysis of these SHC-derived insulin-producing cells. Only Pdx1-transfected SHC-derived cells, but not control cells, released significantly increased amounts of insulin via mechanisms similar to those employed in vivo. Our new culture system also provides material for analyzing b cell differentiation and the properties of SHCs in the adult liver. As shown in Fig. 1, Step 4 is the critical step required for producing insulin1 from the cells derived from SHCs, but we could not identify any molecular event involved in regulating the transcription of insulin 1 during this step. Further investigations will be required to clarify the mechanism of insulin 1 transcription. In contrast, it is known that Pdx-1 is required for the transcription of insulin 2. Our observations are consistent with previous observations showing that adenovirus-mediated transfer of Pdx-1 into liver cells resulted in greater production of insulin 2 than insulin 1 [1]. Regarding the expression of Islet1, Pax6 in differentiating SHCs may contribute to the differential expression of insulin genes, because Pax 6-binding sites exist only in the regulatory regions of insulin1, not insulin 2. Previous reporter assays indicated that Pax6, HNF3b, and NeuroD are positive regulators of Pdx-1 expression [7–9]. However, we could not detect Pdx-1 expression in our culture system, although it is required for b cell differentiation and the secretion of insulin. This observation implies that the expression of Pax6 alone is not sufficient to induce Pdx1 gene expression in differentiating SHCs, because HNF3b and NeuroD are absent in our culture system. We have demonstrated here an important property of Pdx-1 in b cells, namely, forced expression of Pdx-1 in differentiated SHCs induces endogenous Pdx-1 and NeuroD expression in these cells. NeuroD is known to act as a positive regulator of Pdx-1 [8], and thus Pdx-1 and NeuroD will make a positive feedback loop in b cells of islets, and this may contribute to maintaining the properties of b cells in vivo. In summary, the SHCs described here have the ability to differentiate into hepatocytes and insulin-
producing cells, indicating their stem cell nature. Furthermore, the current study provides useful information for understanding b cell differentiation and Pdx-1 function in vivo and may contribute to the development of new therapeutic concepts for treating diabetes mellitus.
Acknowledgments We thank Kazutomo Inoue for his great encouragement, Shin-ichi Sugimoto for the critical advice about techniques, and Asuka Iwanaga for technical assistance. This work was supported in part by Research Fellowships of the Japan Society for the Promotion of Science for Young Scientists.
References [1] S. Ferber, A. Halkin, H. Cohen, I. Ber, Y. Einav, I. Goldberg, I. Barshackm, R. Seijffers, J. Kopolovic, N. Kaiser, A. Karasik, Pancreatic and duodenal homeobox gene 1 induces expression of insulin genes in liver and ameliorates streptozotocin-induced hyperglycemia, Nat. Med. 6 (2000) 568–572. [2] T. Mitaka, F. Sato, T. Mizuguchi, T. Yokono, Y. Mochizuki, Reconstruction of hepatic organoid by rat small hepatocytes and hepatic nonparenchymal cells, Hepatology 29 (1999) 111–125. [3] T. Mitaka, T. Mizuguchi, F. Sato, C. Mochizuki, Y. Mochizuki, Growth and maturation of small hepatocytes, J. Gastroenterol. Hepatol. 13 (1998) S70–77. [4] H. Zulewski, E.J. Abraham, M.J. Gerlach, P.B. Daniel, W. Moritz, B. Muller, M. Vallejo, M.K. Thomas, J.F. Habener, Multipotential nestin-positive stem cells isolated from adult pancreatic islets differentiate ex vivo into pancreatic endocrine, exocrine, and hepatic phenotypes, Diabetes 50 (2001) 521–533. [5] S.H. Lee, N. Lumelsky, L. Studer, J.M. Auerbach, R.D. McKay, Efficient generation of midbrain and hindbrain neurons from mouse embryonic stem cells, Nat. Biotechnol. 18 (2000) 675–679. [6] N. Lumelsky, O. Blondel, P. Laeng, I. Velasco, R. Ravin, R. McKay, Differentiation of embryonic stem cells to insulin-secreting structures similar to pancreatic islets, Science 292 (2001) 1389– 1394. [7] S.E. Samaras, M.A. Cissell, K. Gerrish, C.V. Wright, M. Gannon, R. Stein, Conserved sequences in a tissue-specific regulatory region of the pdx-1 gene mediate transcription in Pancreatic beta cells: role for hepatocyte nuclear factor 3 beta and Pax6, Mol. Cell. Biol. 22 (2002) 4702–4713. [8] S. Sharma, U.S. Jhala, T. Johnson, K. Ferreri, J. Leonard, M. Montminy, Hormonal regulation of an islet-specific enhancer in the pancreatic homeobox gene STF-1, Mol. Cell. Biol. 17 (1997) 2598– 2604. [9] C.S. Lee, N.J. Sund, M.Z. Vatamaniuk, F.M. Matschinsky, D.A. Stoffers, K.H. Kaestner, Foxa2 controls Pdx1 gene expression in pancreatic beta-cells in vivo, Diabetes 51 (2002) 2546–2551.
In vitro induction of adult hepatic progenitor cells into insulin-producing cells Natsuki Nakajima-Nagata,a Tomonori Sakurai,b Toshihiro Mitaka,c Tomoya Katakai,a Eiji Yamato,d Jun-ichi Miyazaki,d Yasuhiko Tabata,e Manabu Sugai,a and Akira Shimizua,f,* a Center for Molecular Biology and Genetics, Kyoto University, 53 Kawahara-cho Shogoin, Sakyo-ku, Kyoto 606-8507, Japan Department of Organ Reconstruction, Institute for Frontier Medical Sciences, Kyoto University, 53 Kawahara-cho Shogoin, Sakyo-ku, Kyoto 606-8507, Japan c Department of Pathology, Cancer Research Institute, Sapporo Medical University School of Medicine, Chuo-ku, S1W17 Sapporo, Japan d Division of Stem Cell Regulation Research, G6, Osaka University Medical School, Suita 565-0871, Japan Department of Biomaterials, Institute for Frontier Medical Sciences, Kyoto University, 53 Kawahara-cho Shogoin, Sakyo-ku, Kyoto 606-8507, Japan f Translational Research Center, Kyoto University Hospital, 54 Kawahara-cho Shogoin, Sakyo-ku, Kyoto 606-8507, Japan b
e
Received 16 March 2004 Available online 24 April 2004
Abstract Organ-specific stem cells are the natural progenitors in tissue regeneration and possess plasticity to differentiate into specialized cells in adult tissues. Small hepatocytes (SHCs) identified in the adult liver are one such cell type. Here we show that SHCs, which are capable of self-renewal and differentiation into hepatocytes, can be induced to generate insulin-producing cells under appropriate culture conditions. These differentiated cells express pancreatic b cell differentiation-related transcripts and hepatocyte differentiation-related transcripts, as shown by reverse-transcription PCR/nested PCR. In addition, enforced expression of the homeodomain transcription factor Pdx1 in these cells contributes to enhancement of insulin release in response to insulin secretagogues. These results indicate that the SHCs described here have the ability to differentiate into insulin-producing cells, and further support the idea that engineering to generate insulin-secreting cells could provide a useful resource for future therapies for diabetes mellitus. Ó 2004 Elsevier Inc. All rights reserved. Keywords: Pancreatic b cell; Pancreatic duodenal homeobox-1; Small hepatocytes; Glucagon-like peptide-1; Reverse transcription
Cell-fate-determination studies in mice have shown that the cells constituting the ventral portion of the foregut in early developmental stages give rise to both the liver and the ventral rudiment of the pancreas at a later stage. A gain-of-function study demonstrated that ectopic expression of Pdx-1 resulted in a substantial increase in hepatic insulin content and an increase of the plasma insulin level [1]. These data also support the idea that it may be possible to induce insulin-producing cells from adult liver cells. In the present study, we focused on the potential of small hepatocytes (SHCs) to differentiate into insulin* Corresponding author. Fax: +81-75-751-4190. E-mail address: [email protected] (A. Shimizu).
0006-291X/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2004.04.059
producing cells and the actual role of Pdx-1 in insulin production and release using this cell population. SHCs have been reported to be hepatocyte progenitor cells, and are isolated from the adult liver within the fraction of non-parenchymal cells (NPCs). These cells can survive and grow for more than 5 months, and interact with hepatic NPCs to reconstruct three-dimensional hepatic organoid structures under appropriate culture conditions in vitro [2,3]. Here we describe an experimental strategy for developing insulin-producing cells that release insulin in response to some insulin secretagogues. These cells were produced by inducing SHCs to differentiate and forcing them to express Pdx-1. Our results provide evidence that SHCs can be induced to generate insulin-producing cells
626
N. Nakajima-Nagata et al. / Biochemical and Biophysical Research Communications 318 (2004) 625–630
and that Pdx-1 might contribute to the regulation of insulin release.
Experimental procedures Isolation of non-parenchymal cell fraction containing small hepatocytes. We isolated liver SHCs from male adult Sprague–Dawley rats (100–150 g body weight, Shimizu Animal, Kyoto, Japan) according to the modified liver perfusion method reported previously [2] (Step 1 in Fig. 1). In Step 2, 5 104 viable isolated cells/cm2 were plated on dishes and cultured in HGM medium, which consisted of DMEM (GibcoBRL) supplemented with 10% FBS, 20 mM Hepes, 25 mM NaHCO3 , 30 mg/L L -proline, 10 mg/L insulin–transferrin–sodium selenite (ITS) (Roche, Mannheim, Germany), 20 ng/mL EGF (Gibco-BRL), 0.004 g/L trace metals (ZnCl2 , ZnSO4 , CuSO4 , and MnSO4 ), 10 mM nicotinamide, 1 mM ascorbic acid 2-phosphate, 107 M dexamethasone (Sigma), and antibiotics. The medium was replaced with fresh medium 2 days after the cells were plated, and the cells were cultured for 3 more days. In Step 3, 1% dimethyl sulfoxide (DMSO; Wako Chemical, Tokyo, Japan) was added to the culture medium. In Step 4, the cells were cultured in DMEM/F12K (1:1) medium supplemented with 0.2% FBS, 20 ng/mL HGF (PeproTech EC, London, UK), 10 ng/mL EGF, 25 mM NaHCO3 , 30 mg/L L -proline, 10 mg/L insulin–transferrin–sodium selenite (ITS) (Roche), 10 mM nicotinamide, 107 M dexamethasone, and antibiotics. As a control, other SHCs continued to be cultured under the conditions of Step 3. Plasmid construction for adenoviral infection. We used an E1-defective adenovirus (Ad) vector designated AdAG-Pdx1-IRES-GFP containing an expression cassette that bicistronically expressed mouse Pdx-1 protein coding sequences and green fluorescent protein (GFP) in the adenovirus genome of pALC under the control of the modified chicken b-actin (AG) promoter. As a control vector, we also used adenovirus vector designed to contain the expression cassette of IRESGFP. Nearly 15–20% of the isolated SHCs were shown to express GFP after infection with one of these vectors at MOI 30. Total RNA isolation, synthesis of complementary DNA, and reversetranscription polymerase chain reaction. Total cellular RNA was extracted from SHCs by the single-step method using TRIZOL reagent (Invitrogen, Carlsbud, CA). The extracted total RNA was reversetranscribed into cDNA with Superscript-II Reverse Transcriptase (Invitrogen) using standard procedures. RT-PCR or nested PCR was performed using Taq polymerase (TAKARA, Shiga, Japan) to monitor the transcription of genes related to the genesis of insulin-producing cells. The primers used were complementary to the mRNA sequences of the genes of interest and are listed in Table 1. All the primers were designed to cross exon–exon boundaries in order to exclude the possibility that genomic DNA was amplified. The mouse Pdx-1 primers were designed to detect Pdx-1 protein-coding sequence transgene transcripts specifically. The rat Pdx-1 PCR primer pair could not detect the RTPCR products from mouse samples. The PCR amplification reaction was performed using different amounts of cDNA to ensure that a linear response of the RT-PCR amplification process was achieved. b-Actin expression was used to normalize for the amount of input template cDNA in order to analyze relative gene expression. Measurement of insulin content The cells were washed four times with PBS and then treated with acid–ethanol (0.1 N hydrochloric acid in absolute ethanol) at 4 °C overnight. The clear supernatants were used to assay the intracellular insulin content and the values obtained were normalized relative to the total protein content. The insulin content was measured by an enzymelinked immunosorbent assay (Insulin-ELISA kit, Shibayagi, Gunma, Japan). Total protein content was measured by the BCA method (Pierce, Rockford, IL).
Fig. 1. Generation of insulin-producing cells from SHCs. (A) General outline of the culture system and Pdx-1 gene transfection protocol for inducing the differentiation of SHCs. (B) The experimental schedule starting from the day SHCs were isolated (day 0). On day 9, adenovirus vector Pdx-1-IRES-GFP or IRES-GFP was transfected into the Group 1 cells. (C) Phase-contrast photographs of the cells on day 7. (D) Cells of Group 1 (+Pdx-1) were subjected to immunofluorescence staining for insulin (red) at the end of Step 4 (day 15) (left panel). The merged image of insulin and DAPI (blue) staining is shown in the right panel. DAPI was used for nuclear staining. (E) Group 2 cells were also stained for insulin (red) (left panel), and the merged image of insulin and DAPI (blue) staining is shown in the right panel. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this paper.)
N. Nakajima-Nagata et al. / Biochemical and Biophysical Research Communications 318 (2004) 625–630
627
Table 1 PCR primers used in this study Genes
Primer sequences
Accession No.
RAT INSULIN I
F 50 -ATGGCCCTGTGGATGCGCTT-30 R 50 -TAGTTGCAGTAGTTCTCCAGCT-30 NEST F50 -GGGAGCCCAAGCCTGCCCAGGC-30 NEST R50 -CAATGCCACGCTTCTGCCGGGC-30
J00747
RAT INSULIN II
F 50 -ACAGTCGGAAACCATCAGCAA-30 R 50 -GCTGGTGCAGCACTGATCCACG-30 NEST F 50 -TGCCCCTGCTGGCCCTGCTC-30
V01243
MOUSE PDX-1
50 -CCGGCTCTGACTGACCGCGTTAC-30 50 -TGCTCCAGTGATCCCAGCGAGC-30
AK020261
RAT PDX-1
F 50 -TACGCGGCCACACAGCTCTACAAGGAC R 50 -CCACTTCATGCGACGGTTTTGGAACCAGA-30 NEST F 50 -TCTCCTCCGGTTCTGCTGCGTATGCAC-30
NM022852
MOUSE NESTIN
F 50 -CTGCTGGAGGCTGAGAACT-30 R 50 -ATAGGTGGGATGGGAGTGC-30
M34384
RAT Isl 1
F 50 -GCCCGCTCTAAGGTGTACCACATC-30 R 50 -TCATGATGCTGCGTTTCTTGTCCTT-30 NEST F 50 -AAGGTGTACCACATCGAATGTTTCC-30
X53258
MOUSE PTF1
F 50 -GCAGCTGCGACAAGCCGCTAA-30 R 50 -CCTCTGGGGTCCACACTTTAGCTGTACGGA-30 NEST F 50 -GCGCCGCATGCAGTCCATCAACGA-30
AF298116
RAT NEUROD
F 50 -GCTTGGCCAAGAACTATATCTGGGC-30 R 50 -GTGAAAGATGGCATTAAGCTGGGC-30 NEST R 50 -CGAAAGACATAATATTGTCTATGGGG-30
NM019218
RAT PAX 6
F 50 -ACCAACGACAATATACCCAGTGTGTCATC-30 R 50 -GTGTTGCTGGCCTGTCTTCTCTGGTTCC-30 NEST R 50 -TGGCCTGTCTTCTCTGGTTCCTC-30
NM013001
RAT HNF1A
F 50 -CCCCATCTGAAGGTGCCAACCTCAAC-30 R 50 -CCGCTTGATCTTGCCTGGGTCACTC-30
NM012669
RAT HNF1B
F 50 -CCAGAAGCGAGCTGCCCTGTACAC-30 R 50 -AGGCTGCTAGCCACACTGTTGATGAC-30
NM013103
RAT HNF4A
F 50 -CCCCATCTGAAGGTGCCAACCTCAAC-30 R 50 -CCGCTTGATCTTGCCTGGGTCACTC-30
NM022180
RAT CK8
F 50 -CAACTTCCTCCGGCAGATCCAT-30 R 50 -ATCCTTCACGGCCAGTTCCCC-30
NW043527
RAT AFP
F 50 -GCGATGCGTTGGCTGCAATGAAGG-30 R 50 -GTTGTCAGCTTTGCAGCATGCTGGAAC-30
NM012493
RAT c-MET
F 50 -AGAGAACATTGGGGTGGCTGAAA-30 R 50 -ATAGAGTGGAAGCAAGCAGTCTC-30
U65007
MOUSE NGN3
F 50 -TACCTAGGGACTGCTCCGAAGCAGAAGTGG-30 R 50 -ACCAGTGCTCCCGGGAGCAGATAGGATG-30 NEST F 50 -GGCCCAAGAGCGAGTTGGCACTCAGCA-30 NEST R 50 -TCCAATGAGGCCGTGGGGCTCAGGTT-30
Y09167
Measurements of insulin secretion in response to insulin secretagogues. Insulin secretion was measured using static incubation in Krebs–Ringer bicarbonate buffer (KRBB) supplemented with 5 mM glucose and 0.2% FBS. The cells were plated in 6-well culture plates (35mm diameter; IWAKI, Tokyo, Japan), cultured, and washed four times with KRBB on day 15. They were then incubated in KRBB with high glucose (25 mM) and 10 nM glucagon-like peptide 1(GLP-1 [7–37]) (Sigma) or 45 mM KCl and 0.2 mM tolbutamide (Sigma), or low glucose (5 mM) in KRBB as a control. The concentration of the insulin secreted into the buffer solution was measured as described above.
Immunocytochemistry. Cells were incubated in the Step 4 medium described above supplemented with Golgestop solution (Pharmingen) plus the secretagogues described above for 4 h in a CO2 incubator. Immunocytochemistry was carried out using standard protocols in which the cells were sequentially incubated with mouse anti-preproinsulin antibody (Biogenesis, Poole, UK) diluted 1:1000 in PBS, biotin-labeled anti-mouse IgG antibody (Nichirei, Tokyo, Japan), and PE-conjugated streptavidin (Molecular Probes), and assessed by examining the cells with an Olympus AX80 microscope.
628
N. Nakajima-Nagata et al. / Biochemical and Biophysical Research Communications 318 (2004) 625–630
Statistical analysis. Statistical comparisons were made by analysis of variance and, when appropriate, using the unpaired Student’s t test, with P < 0:05. Experimental results were expressed as the means standard deviation of the mean. All experiments were repeated at least three times independently.
Results Generation of insulin-producing cells from SHCs We used an experimental strategy (Figs. 1A and B) whereby in Step 1 an SHC-rich fraction was separated from adult rat liver as previously described [2]. In the early steps of culturing (Steps 2 and 3), we detected nestin-positive SHCs (Figs. 1C and 2), which are progenitors of neurons and pancreatic endocrine cells [4,5]. This strategy was very similar to that used for generating insulin-producing cells from ES cells [6]. Next, we tried to determine the optimal conditions for the differentiation of insulin-producing cells from nestin-positive cells (Step 4). Among the conditions we examined, culturing the cells in the presence of HGF, EGF, nicotinamide, and dexamethasone in 0.2% FBS-containing medium effectively induced the production of insulin-producing cells. To further examine the function of Pdx-1 in insulin
Fig. 2. Gene expression profiles of cultured SHCs. We performed RTPCR to detect mouse PDX-1, nestin, Id2, HNF1a, HNF1b, HNF4a, AFP, CK8, c-Met, and b-actin mRNA transcripts, and RT-PCR followed by nested PCR to detect insulin 1, insulin 2, rat Pdx-1, neurogenin, NeuroD, Isl 1, Pax6, and PTF1 mRNA transcripts. The cells of Group 1 showed expression of these genes on day 0 (Step 1), day 7 (Step 3), day 9 (Step 4), and day 15 (Step 4) in culture. We assessed the effects of Pdx-1 overexpression in the cells of Group 1 on day 15 (shown by “+Pdx-1” and the control “)Pdx-1”). As an additional control, the cells of Group 2 were also assessed on day 15. Gene expression studies in all groups were repeated at least three times with similar results.
production, we employed an adenovirus-mediated approach for introducing Pdx-1 into nestin-positive SHCs. Mouse Pdx-1 (mPdx-1) cDNA was cloned into the Ad AG-IRES-GFP vector, which allowed Pdx-1 expression to be assessed by monitoring green fluorescent protein (GFP). The Group1 progenitor cells were infected with the viral vector on day 9 and analyzed for the expression of insulin on day 15 by immunocytochemistry (Figs. 1D and E). There was no obvious difference of GFP expression or cell morphology between the Pdx-1transduced and control populations (data not shown). Gene expression profiles of cultured SHCs To further analyze the molecular events that occurred in SHCs during the series of culturing steps, the gene expression profiles were determined by RT-PCR (Fig. 2). SHCs at Step 1 expressed albumin, a marker of mature hepatocytes, c-Met, expressed in mature hepatocytes and b cells, and neurogenin 3, a marker of islet progenitor cells. At the end of Step 3, albumin and neurogenin 3 expression disappeared and the expression of nestin was observed (Fig. 2, lanes 1 and 2). On day 15, insulin 1 expression was only detected in cultured cells in Group 1. This observation suggested the importance of Step 4 in the differentiation of SHCs into insulin-producing cells. Like insulin 1, a-fetoprotein (AFP), a marker of immature hepatocytes, was expressed only in Group 1. However, the importance of AFP in insulin 1 production is not clear. AFP expression seems to indicate the re-differentiation of hepatocytes under these culture conditions, because AFP expression is not observed in the pancreas (data not shown). We found that insulin 2 expression was completely dependent on the expression of Pdx-1. This observation is consistent with data reported previously [1]. Forced expression of Pdx-1 also induced the expression of endogenous Pdx-1 and NeuroD. These results indicate that Pdx-1 and/or NeuroD are required for insulin 2 expression. Islet1 and Pax6 were detected after 9 days of culturing under both conditions (Group 1 and Group 2). In contrast, no expression of PTF1, a marker of exocrine cells, was observed in any of the cells analyzed. These data are consistent with the idea that b cell differentiation occurred under these culture conditions. The expression levels of HNF1a, 1b, and 4a, whose mutations cause diabetes mellitus, did not vary among the cells examined. Insulin content in differentiated SHCs The intracellular insulin content in the differentiating SHCs was measured and normalized relative to the total protein content on days 0, 5, 9, 12, and 15 (Figs. 1 and 3). As a control, the cells from Group 2 were also assayed. Before day 9, only very low levels of insulin content were detected (data not shown). After 9 days of culturing, the
N. Nakajima-Nagata et al. / Biochemical and Biophysical Research Communications 318 (2004) 625–630
629
Insulin release by differentiated SHCs
Fig. 3. Intracellular insulin content in differentiating SHCs. The insulin levels in the cells (n ¼ 4) of Group 1 (consisting of subgroups of both Pdx-1-expressing cells (+Pdx-1: black columns) and non-Pdx-1-expressing cells ()Pdx-1: gray columns)), and control cells (Group 2: white columns) were assessed on day 9, day 12, and day 15. Three separate experiments were performed to determine the insulin level of each group. P > 0:05.
insulin level of the cells from Group 1 was significantly increased in comparison with the lower insulin level in control cells (Fig. 3). Pdx-1 had no effect on the insulin content, indicating that the major insulin source at this time was the insulin 1 gene (Fig. 3). Overall, the results of the insulin content analysis further support the idea that Step 4 is the most important condition for inducing insulin-producing cells.
To determine whether the ability of differentiated SHCs to secrete insulin was affected by Pdx-1 or not, we examined insulin release in response to several secretagogues (Fig. 4). According to the known mechanism by which glucose stimulates insulin release in vivo, transport of glucose into the cells results in adenosine triphosphate (ATP) production, which enhances Ca2þ influx and subsequent Ca2þ release from the intracellular store. The elevation of intracellular Ca2þ is coupled to multiple phosphorylation events modulated by the protein kinase A (PKA) and protein kinase C (PKC) cascades. These pathways are known to influence insulin release. The roles of these two pathways of insulin secretion were tested (Fig. 4). GLP 1, an activator of the cAMP–PKA pathway, stimulated insulin secretion from both Pdx-1-transduced cells (+Pdx-1) and control cells ()Pdx-1), but the induction was 2.5-fold higher in Pdx1-transduced cells (Fig. 4, left panels). Tolbutamide (sulfonylurea) and Kþ , inducers of Ca2þ influx, had the same effect on insulin secretion as seen in the left panels of Fig. 4 (central panels). In contrast, only a slight effect on insulin secretion was observed in the cultures incubated with a low glucose concentration for 30 min without agonist (Fig. 4, right panels). These observations support the idea of the functional similarity of these cells to b cells. In all the cultures examined, we observed about 2-fold higher insulin release (1.6- or 2.4fold higher in these stimulation tests) from +Pdx-1 cells than from )Pdx-1 cells in spite of almost equal insulin contents in these cells (see Fig. 3). These data indicate that the universal components of insulin release can be affected by Pdx-1 in in vitro differentiated SHCs.
Fig. 4. Insulin release in response to insulin secretagogues. The effects of various agonists on insulin secretion from insulin-producing cells of Group 1 (either Pdx-1-expressing cells (+Pdx-1) or non-Pdx-1-expressing cells ()Pdx-1)), and from Group 2 cells were tested on day 15. Insulin release from these cells was observed in response to stimulation by 10 nM GLP-1 with high-dose glucose (25 mM) (left panels). The insulin release in response to a high physiological concentration (45 mM) of Kþ and 0.2 mM tolbutamide with low-dose glucose (5 mM) was also examined (central panels). As a control, insulin release in response to low-dose glucose (5 mM) in basal buffer (KRBB solution) was also examined (right panels). The integral amount of insulin release from 0 to 5 min (open columns), 0–10 min (gray columns), and 0–30 min (closed columns) is shown in the upper panels.
630
N. Nakajima-Nagata et al. / Biochemical and Biophysical Research Communications 318 (2004) 625–630
Discussion We report here that SHCs could efficiently generate insulin-producing cells having significantly elevated insulin content in their cytoplasm. The generation of insulin-secreting cells from adult stem cells is an important advance for use in future therapies for diabetes mellitus. Various secretagogues were used for functional analysis of these SHC-derived insulin-producing cells. Only Pdx1-transfected SHC-derived cells, but not control cells, released significantly increased amounts of insulin via mechanisms similar to those employed in vivo. Our new culture system also provides material for analyzing b cell differentiation and the properties of SHCs in the adult liver. As shown in Fig. 1, Step 4 is the critical step required for producing insulin1 from the cells derived from SHCs, but we could not identify any molecular event involved in regulating the transcription of insulin 1 during this step. Further investigations will be required to clarify the mechanism of insulin 1 transcription. In contrast, it is known that Pdx-1 is required for the transcription of insulin 2. Our observations are consistent with previous observations showing that adenovirus-mediated transfer of Pdx-1 into liver cells resulted in greater production of insulin 2 than insulin 1 [1]. Regarding the expression of Islet1, Pax6 in differentiating SHCs may contribute to the differential expression of insulin genes, because Pax 6-binding sites exist only in the regulatory regions of insulin1, not insulin 2. Previous reporter assays indicated that Pax6, HNF3b, and NeuroD are positive regulators of Pdx-1 expression [7–9]. However, we could not detect Pdx-1 expression in our culture system, although it is required for b cell differentiation and the secretion of insulin. This observation implies that the expression of Pax6 alone is not sufficient to induce Pdx1 gene expression in differentiating SHCs, because HNF3b and NeuroD are absent in our culture system. We have demonstrated here an important property of Pdx-1 in b cells, namely, forced expression of Pdx-1 in differentiated SHCs induces endogenous Pdx-1 and NeuroD expression in these cells. NeuroD is known to act as a positive regulator of Pdx-1 [8], and thus Pdx-1 and NeuroD will make a positive feedback loop in b cells of islets, and this may contribute to maintaining the properties of b cells in vivo. In summary, the SHCs described here have the ability to differentiate into hepatocytes and insulin-
producing cells, indicating their stem cell nature. Furthermore, the current study provides useful information for understanding b cell differentiation and Pdx-1 function in vivo and may contribute to the development of new therapeutic concepts for treating diabetes mellitus.
Acknowledgments We thank Kazutomo Inoue for his great encouragement, Shin-ichi Sugimoto for the critical advice about techniques, and Asuka Iwanaga for technical assistance. This work was supported in part by Research Fellowships of the Japan Society for the Promotion of Science for Young Scientists.
References [1] S. Ferber, A. Halkin, H. Cohen, I. Ber, Y. Einav, I. Goldberg, I. Barshackm, R. Seijffers, J. Kopolovic, N. Kaiser, A. Karasik, Pancreatic and duodenal homeobox gene 1 induces expression of insulin genes in liver and ameliorates streptozotocin-induced hyperglycemia, Nat. Med. 6 (2000) 568–572. [2] T. Mitaka, F. Sato, T. Mizuguchi, T. Yokono, Y. Mochizuki, Reconstruction of hepatic organoid by rat small hepatocytes and hepatic nonparenchymal cells, Hepatology 29 (1999) 111–125. [3] T. Mitaka, T. Mizuguchi, F. Sato, C. Mochizuki, Y. Mochizuki, Growth and maturation of small hepatocytes, J. Gastroenterol. Hepatol. 13 (1998) S70–77. [4] H. Zulewski, E.J. Abraham, M.J. Gerlach, P.B. Daniel, W. Moritz, B. Muller, M. Vallejo, M.K. Thomas, J.F. Habener, Multipotential nestin-positive stem cells isolated from adult pancreatic islets differentiate ex vivo into pancreatic endocrine, exocrine, and hepatic phenotypes, Diabetes 50 (2001) 521–533. [5] S.H. Lee, N. Lumelsky, L. Studer, J.M. Auerbach, R.D. McKay, Efficient generation of midbrain and hindbrain neurons from mouse embryonic stem cells, Nat. Biotechnol. 18 (2000) 675–679. [6] N. Lumelsky, O. Blondel, P. Laeng, I. Velasco, R. Ravin, R. McKay, Differentiation of embryonic stem cells to insulin-secreting structures similar to pancreatic islets, Science 292 (2001) 1389– 1394. [7] S.E. Samaras, M.A. Cissell, K. Gerrish, C.V. Wright, M. Gannon, R. Stein, Conserved sequences in a tissue-specific regulatory region of the pdx-1 gene mediate transcription in Pancreatic beta cells: role for hepatocyte nuclear factor 3 beta and Pax6, Mol. Cell. Biol. 22 (2002) 4702–4713. [8] S. Sharma, U.S. Jhala, T. Johnson, K. Ferreri, J. Leonard, M. Montminy, Hormonal regulation of an islet-specific enhancer in the pancreatic homeobox gene STF-1, Mol. Cell. Biol. 17 (1997) 2598– 2604. [9] C.S. Lee, N.J. Sund, M.Z. Vatamaniuk, F.M. Matschinsky, D.A. Stoffers, K.H. Kaestner, Foxa2 controls Pdx1 gene expression in pancreatic beta-cells in vivo, Diabetes 51 (2002) 2546–2551.