Combined Expression of Pancreatic Duodenal Homeobox 1 and Islet Factor 1 Induces Immature Enterocytes to Produce Insulin

Immature rat intestinal stem cells, IEC-6 (passage 11; American Type Culture Collection, Rockville, MD), a nontransformed line, were maintained in culture, as previously described (22). The IEC-6 cells were grown in T-150 flasks containing Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 5% dialyzed fetal bovine serum (FBS), 5.6 mmol/l d-glucose, 44.0 mmol/l NaHCO3, 0.1 g/l streptomycin, and 10⁵ units/l penicillin G. These cells were transfected with or without Pdx-1 and/or Isl-1 and then seeded onto a six-well 0.5-μm micropore filter (Falcon Cell Culture Insert; Becton Dickinson, Rutherford, NJ), and grown for 4 days in DMEM containing 5% FBS before manipulations. The cells were washed twice with PBS before being used for histologic and molecular biological analyses.

The complete coding sequence of the rat Isl-1 or mouse Pdx-1 cDNA was inserted separately into the plasmid vector, pcDNA3, to yield pcDNA3-Pdx-1 or pcDNA3-Isl-1, respectively. The two clones had differential antibiotic resistance to neomycin (G418) and hygromycin (Calbiochem-Novabiochem, La Jolla, CA), respectively. First, 3 μg of the pcDNA3-Pdx-1 (neomycin resistant vector) or empty vector (pcDNA3) were transfected into IEC-6 cells by electroporation at 250 V and 975 μF using the Gene Pulser (BioRad, Osaka City, Japan). Then, 50 of the neomycin resistant clones were isolated and labeled as YK cells. The number designation was used to identify individual clones. Clones YK-12, -14, and -15 were randomly cloned from isolates as representative of the Pdx-1−expressing cells. In this study, we tested all three clones to probe for representative features. The same procedure was used to establish Isl-1−expressing cells by transfection of 3 μg pcDNA3-Isl-1 into IEC-6 cells and then selecting for hygromycin resistance. Again, three randomly selected isolates—IS-1, IS-2, and IS-3—were chosen for the studies. Furthermore, Isl-1−YK cells that expressed both Pdx-1 and Isl-1 were established by transfecting 3 μg of pcDNA3-Isl-1 into YK-12, -14, and -15 cells that already expressed Pdx-1. These cells should then have had double antibiotic resistance and were selected in medium containing 0.6 mg/ml of G418 and 0.2 mg/ml of hygromycin and then screened for Pdx-1 and Isl-1 expression using Northern blot analysis. No difference in growth was observed between YK and IEC-6 cells containing empty vectors. The YK cells grown on micropore filter became trabecular in shape and showed strong positive staining of Pdx-1 in the nucleus (22). The effect of betacellulin was examined by the daily addition of 2 nmol/l of the IFG-1 for 10 days to cells seeded on micropore filters.

Total RNA was extracted from cells using a procedure previously described (24). The extracted product was used as a template in reactions containing RT (New England Biolabs) and the 3′ primer of interest to synthesize the first strand of the cDNA. RT-PCR analysis was done as previously described (24). The first-strand cDNA was used in subsequent PCR analyses, with the oligonucleotide primers listed below for each one of the genes of interest somatostatin: 5′) TCTCTGCTGCCTGGGACCT, 3′) GCCAAGAAGTACTTG GCCAGTTC insulin: 5′) TGCCCAGGCTTTTGTCAAACAGCACC, 3′) CTCCACCCAGCTCCA GTT PP: 5′) TGAACAGAGGGCTCAATACG, 3′) AGACAGAAGGGAGGCTACAAA TCC Nkx6.1: 5′) CCGGTATTCACACCACTTGGC, 3′) AGCGCAAAGGGGCTGCCT TTTC Isl-1: 5′) GCGGAGGATGGGCTTTTCTG, 3′) TTCTGCTTTTCGTTGAGCACAG neuroD: 5′) CTTGGCCAAGAACTACATCTGG, 3′) GGAGTAGGGATGCACCG GGAA Pax4: 5′) CTCACGGAGCCTTAAGGTATC, 3′) TGAATCTGGATACTGCCCACG Pax6: 5′) GTCTACCAGCCAATTCCACAGC, 3′) CAGCCATTTCTCTTTCTTT CCT

The identity of the PCR products was confirmed by size after agarose gel electrophoresis and nucleotide sequence analysis. Total RNA was extracted from neonatal rat intestine, neonatal rat pancreas, IEC-6 cells, Isl-1-transfected IEC-6 cells, YK cells, YK cells exposed to 2 nmol/l betacellulin, and Isl-1 transfected YK cells. A 10-μg aliquot of each sample of total RNA was used for Northern blot analysis. The transferred blots were hybridized with a ³²P-labeled fragment of Pdx-1 (772–1,134), Isl-1 (341–576), or insulin II (56–238) cDNA. The signal arising from actin mRNA served as the internal control to evaluate differences among various samples.

Relative amounts of Pdx-1 and Isl-1 proteins in the cells were measured using Western blot analysis, according to a procedure previously described (24). Specifically, 5 μg cellular lysate were separated by standard SDS-PAGE and then transferred to polyvinylidine fluoride membrane. Specific antibodies against Pdx-1 (25) or Isl-1 (26) protein were used to probe the blot. The antibody-protein complex was detected using an enhanced chemiluminescence kit (Amersham Pharmacia Biotech).

The cells were stained as previously described (22). In brief, the cells were fixed for 2 h with 4% paraformaldehyde, and washed for 4 days with 0.1 mol/l PBS containing 0.3% triton X (PBST) then incubated with an anti-insulin antibody (26) diluted to 1:5,000 in PBST. Immunocytochemical staining was achieved using the avidin-biotin complex method and ABC-nickel reactions (22). For immunofluorescence double staining, the fixed cells were incubated with a mixture of antibodies against Pdx-1 (25) and insulin diluted to 1:5,000 in PBST or a mixture of antibodies against Isl-1 (26) and insulin diluted to 1:5,000 in PBST. The cells were then incubated with the species-specific secondary antibodies conjugated to either fluorescein isothiocyanate or Texas Red. The positive reaction was observed under fluorescence microscopy (Olympus IX70; Olympus, Osaka, Japan) and the images were taken with a charge-coupled device camera (Cool SNAP/HQ; Nippon Roper, Osaka, Japan). Specificity of the positive staining was examined by an immunocytochemical absorption study. Electron microscopic studies of secretory granules were accomplished by incubating cells for 1 h with 1% OsO4 in 0.1 mol/l PBS at 4^oC. The specimen was embedded in epoxy resin and cut into ultra-thin sections. These samples were stained with 2% uranyl acetate and Reynold’s solution and then observed under an electron microscope (H-7100; Hitachi, Tokyo, Japan).

Pdx-1−expressing YK-14 cells were plated onto six-well dishes (Corning Glass Works, Corning, NY) at a density of 1 × 10⁶ cells per well and allowed to grow for 6 days with DMEM containing 5% FCS and 2 nmol/l betacellulin. DMEM containing betacellulin was replaced each day. The cells were washed three times for 10 min each in DMEM with 0.1% BSA and 5.6 mmol/l glucose. After being washed, the cells were incubated with DMEM containing 5.6 or 22.0 mmol/l glucose and 0.1% BSA for 12 h. The medium was collected and analyzed for its insulin content with a radioimmunoassay (RIA) requiring a specific antibody against insulin (27). The protein content in the cell was analyzed by Bradford assay (BioRad Laboratories). Insulin secretion was normalized based on the total protein in each sample.

Male SD rats, age 8 weeks, were used in the implantation studies. Rats were rendered diabetic with a single intravenous injection of streptozotocin (STZ) (Sigma, Madrid, Spain) at a dosage of 50 mg/kg body wt. Blood glucose was measured between 1700 and 1800 in 8-h-fasted animals using the glucose oxidase method (Glutest-Pro; Sanwa, Kyoto, Japan). Then, 6 days after the STZ injection, five diabetic rats were implanted with Isl-YK-14 cells. Preparation of cells for implantation involved washing and resuspending cells in PBS supplemented with 5.6 mmol/l glucose at 5 × 10⁶ cells/ml. The animals were placed under general anesthesia using ketamine hydrochloride (95 mg/kg, i.p.) and injected with 5 × 10⁶ cells in the peritoneal cavity. The concentration of blood glucose and insulin were measured on a daily basis.

Results are given as means ± SD, unless otherwise stated. Scheffe’s multiple comparison test was used to determine the significance of any differences among more than two groups, and the unpaired Student’s t test was used to determine the significance of any differences between two groups.

We have recently established a cell line, YK, derived from wild-type IEC-6 cells stably transfected with Pdx-1 cDNA (22). In culture, these cells form a trabecular network and have the capacity to differentiate into cells that express several hormones, including serotonin, somatostatin, gastrin, and cholecystokinin (22). Three separate clones—YK-12, -14, and -15 from the pool of YK cells—were chosen for analysis. The results showed expression of Pdx-1 mRNA and protein in YK-12, -14, and -15 isolates (Figs. 1A and B, lanes 3, 4, and 5, respectively), but not in wild-type IEC-6 or cells carrying empty vector. The expression of actin mRNA was detected in both YK and IEC-6 cells (Fig. 1A, lanes 1 and 2). Pdx-1 protein was detected only in YK cells, and not IEC-6 cells, as a 46-kDa band (Fig. 1B). The apparent molecular mass of Pdx-1 is 46 kDa and differs from its cDNA estimated value of 31 kDa likely because of translocation, as described in a recent report (28). These studies showed clearly that IEC-6 cells stably transfected with Pdx-1 cDNA express both the mRNA and the protein.

The process of differentiation leading to the formation of pancreatic cells requires specific transcription factors and extrinsic signals, as noted above. This knowledge prompted us to postulate that immature intestinal cells transfected with Pdx-1 might acquire the ability to express insulin after exposure to betacellulin. We tested this hypothesis using YK cells exposed to 2 nmol/l betacellulin. YK cells treated with betacellulin (BYK cells) were examined for the expression of hormones somatostatin, pp, insulin, and glucagon using RT-PCR (Fig. 2). When the three isolates BYK-12, -14, and -15 were examined, they revealed the same finding (only data from BYK-14 are shown). YK-14 cells (Fig. 2, lane 5) expressed pp and somatostatin, whereas the BYK-14 clone newly expressed insulin in addition to pp and somatostatin (lane 6) but not glucagon. The identity of the 182-bp RT-PCR product was presumed to be that of insulin cDNA, verified by nucleotide sequencing. In contrast, wild-type IEC-6 cells or those transfected with empty vector (Fig. 2, lanes 3 and 4, respectively) did not express any of the mRNA-encoding hormones of interest. In addition, treatment of IEC-6 cells with betacellulin did not cause them to produce insulin or alter expression of the other hormones (data not shown). As expected, rat pancreatic cells (Fig. 2, lane 2) expressed mRNAs for all four hormones.

To further demonstrate that BYK isolates express insulin, the cells were immunostained for the presence of the hormone. Immunoreactivity for insulin was observed in the cytoplasm of trabecular-shaped BYK-14 cells (Fig. 3B, indicated by arrows). In agreement with results of RT-PCR (Fig. 2), no insulin was found in untreated YK-14 cells (Fig. 3A). Next, we examined the ultrastructure of BYK-14 cells (Fig. 4B [left and right]) and YK-14 cells (Fig. 4A [left and right]) using electron microscopy. Results showed the presence of secretory granules in the cytoplasm of both cell types. Whereas YK-14 cells contained secretory granules of pleomorphic shapes (Fig. 4A, right), those of BYK-14 cells were round in shape with an electron-dense core (Fig. 4B, right). Findings similar to those in BYK-14 cells were also observed in BYK-12 and BYK-15 isolates (data not shown). The addition of these findings to those of RT-PCR demonstrated that Pdx-1−expressing YK isolates derived from IEC-6 cells exposed to betacellulin enabled these cells to express insulin.

Glucose-regulated expression of the insulin gene is a key feature of this gene in pancreatic cells. To test whether glucose induced the expression of insulin in BYK cells, the cells were incubated in medium containing either 5.6 or 22.0 mmol/l d-glucose and then assayed for their concentration of insulin using RIA. Results showed that the expression of insulin was not detected in YK-14 cells (Table 1). However, after exposure of the same cells to betacellulin, insulin (106.2 ± 6.6 nmol/l) was secreted into the cell culture medium. The concentration of insulin in the medium was the same in the presence of 5.6 or 22.0 mmol/l glucose (data not shown). These results demonstrated that exposure of the Pdx-1−expressing IEC-6 cells (YK cells) to betacellulin induces the cells to express and secrete insulin, although this response is not regulated by the glucose concentration in the medium.

We previously showed that IEC-6 cells express a variety of transcription factors, including hepatocyte nuclear factor (HNF)-1α, -3α, -3β, and -4α and NeuroD/Beta2 (22). We then postulated that in the presence of betacellulin, the expression of added transcription factors that enable BYK cells to produce insulin is probably different than in untreated Pdx-1−expressing cells. To examine this possibility, we used RT-PCR to probe for the expression of several transcription factors known to be important for pancreatic β-cell differentiation. The results showed that two Pdx-1 expressing cells, isolates YK-12 and -14 (Fig. 5, lanes 4 and 5, respectively), contained mRNAs encoding NeuroD/Beta2, Pax6, and Nkx6.1, but not mRNAs for Isl-1 and Pax4. After exposure of the YK-12, -14, and -15 cells to betacellulin (Fig. 5, lanes 6–8), we detected Isl-1 expression with the disappearance of Pax6 expression. As expected, rat pancreas expressed all factors (lane 1), but the wild-type IEC-6 cells transfected with empty vector (lane 2) or those exposed to betacellulin (lane 3) expressed none of the preceding factors except NeuroD/Beta2. These findings demonstrate that betacellulin induces the expression of Isl-1 in YK cells and that this event appears to be important in enabling the expression of insulin in these cells.

If the expression of Isl-1 in YK cells is critical for the expression of insulin, then overexpression of Isl-1 in Pdx-1−expressing IEC-6 cells might mimic the effects of betacellulin treatment. Therefore, we stably transfected YK-12, -14, and -15 clones with Isl-1 cDNA and then tested the isolates for the presence of Isl-1 expression using Northern and Western blot analysis. The results (Fig. 6) showed that neither wild-type IEC-6 nor YK-12 cells expressed Isl-1 mRNA (Fig. 6A, lanes 1 and 2) or protein (Fig. 6B, lanes 1 and 2). However, in cells transfected with Isl-1 cDNA, both Isl-1 mRNA and protein were observed (Fig. 6B, lanes 5–7). As expected, IEC-6 cells transfected with Isl-1 cDNA also expressed both Isl-1 mRNA and protein (lanes 3 and 4).

Next we tested for the presence of insulin mRNA in YK cells transfected with Isl-1 (Isl-YK cells) and BYK cells using Northern blot analysis. The results showed that, as expected, insulin mRNA was detected in the pancreatic cells (Fig. 6C, lane 1), but not in the rat intestinal cells (Fig. 6C, lane 2). Similarly, the mRNA was not present in wild-type IEC-6 cells (Fig. 6C, lane 3), IEC-6 cells transfected with Isl-1 (Fig. 6C, lane 4), or YK-14 cells (Fig. 6C, lane 5). However, insulin mRNA was present in BYK-14 and -15 cells (Fig. 6C, lanes 6 and 7) and Isl-YK-12, -14, and -15 cells (Fig. 6C, lanes 8–10). Although not shown here, the abundance of insulin mRNA in the cells and protein concentration in the medium did not change in BYK cells after exposure to 5.6 or 22.0 mmol/l d-glucose. These findings show that equipping the Pdx-1−expressing YK cells with the ability to make Isl-1 enables the cells to produce insulin mRNA without the need for treatment with betacellulin.

Next, we searched for co-localization of Pdx-1, Isl-1, and insulin expression in Isl-YK-14 cells using immunofluorescence double staining. In Isl-YK-14 cells, positive staining for insulin was localized to the cytoplasm (Fig. 7B and E). Consistent with the transcriptional role of Pdx-1, positive staining for this protein was found mainly in the nucleus but also minimally in cytoplasm (Fig. 7A). Insulin and Pdx-1 staining co-localized to Isl-YK-14 cells (Fig. 7C). Similarly, positive staining for Isl-1 appeared in the nucleus of Isl-YK-14 cells (Fig. 7D), and this staining co-localized with that for insulin in Isl-YK-14 cells (Fig. 7F). The same studies using Isl-YK-12 and -15 cells showed the identical findings (results not shown). These observations demonstrated that the expression of Pdx-1, Isl-1, and insulin co-localize in Isl-YK cells. As expected, positive staining of Isl-1 was also observed in the nucleus of IEC-6 cells transfected with Isl-1 cDNA alone (Fig. 7G), but the staining for insulin (Fig. 7H) was not present in these cells.

To determine whether Isl-YK-14 cells could produce insulin in vivo, we implanted these cells into the peritoneal cavity of diabetic rats. Rats were rendered diabetic, on day 0, with a single injection of STZ, and the appearance of diabetes was monitored over the ensuing 5 days. On day 6 of the study, the animals were injected with Isl-YK-14 cells and then monitored for glucose levels. In animals injected with YK-14 cells (Fig. 8A), the mean glucose continued to climb from 19.3 up to 21.8 mmol/l spanning days 7–9 of the study. In contrast, the rats implanted with the insulin-producing Isl-YK-14 cells showed a rapid decrease in blood glucose levels from 19.4 mmol/l on day 6 to 16.9 mmol/l on day 7, a value that was significantly lower (P < 0.05) than in animals injected with YK-14 cells. Consistent with the reduction in glucose, the plasma insulin concentration was 2.7.-fold higher in rats carrying the Isl-YK-14 cells (P < 0.05) than in rats with YK-14 cells on day 7 of the study (Fig. 8B). Similar results were obtained using Isl-YK-15 cells. These findings show that Isl-YK cells produced functionally active insulin after being injected into diabetic rats.

Pancreatic cell differentiation from progenitor cells requires both intrinsic and extrinsic influences, such as transcription and growth factors. In this report, we used the cell line IEC-6, derived from rat immature intestinal crypt cells, to test whether these cells can be induced to express insulin. Our previous studies showed that IEC-6 cells given the ability to express a transcription factor, Pdx-1, undergo differentiation into enteroendocrine cells. In culture, these cells form a trabecular network, acquire features of endocrine cells, and synthesize several enteroendocrine hormones, but not insulin (22). After exposure of the Pdx-1−expressing IEC-6 cells (YK cells) to 2 nmol/l betacellulin, these cells synthesized insulin and secreted this hormone into the culture medium. However, the regulation of secreted insulin from the cells in response to glucose concentrations was not observed. Insulin gene expression is known to require the participation of several transcription factors (29,30); in this study, betacellulin treatment of YK cells yielded BYK cells, which have the ability to synthesize insulin. The BYK cells expressed transcription factor(s) that differed from those of YK cells. This finding prompted us to examine whether the specific transcription factors were important for insulin gene expression in the BYK cells. Results of RT-PCR using total RNA from the insulin-producing BYK cells revealed the presence of sequences encoding; NeuroD/Beta2, Isl-1, and Nkx6.1. Among the transcription factors tested, only Isl-1 was detected in the BYK and not in the YK cells before treatment with betacellulin. These results suggest that the combined expression of Pdx-1 and Isl-1 enables intestinal stem cells to produce insulin.

Recent studies in vitro and in vivo have suggested that enteroendocrine cells may share a common lineage with those of the intestinal epithelium, such as the absorptive enterocytes, goblet, or Paneth cells. The origin of both the enteroendocrine and intestinal epithelial cells is believed to arise from a primitive intestinal stem cell located in the intestinal crypt layers (31). Published studies have shown that two transcription factors, cdx-2 and Pdx-1, may be important in deciding the pathway leading to differentiation and thus the formation of absorptive enterocytes or pancreatic cells, respectively (32). The overexpression of cdx-2 in IEC-6 cells resulted in the formation of absorptive enterocytes or goblet cells (33). In contrast, cellular expression of Pdx-1, a homeotic selector in IEC-6 cells, led to the formation of pancreatic β-cells (22). The presence of Pdx-1 enables these immature enterocytes to produce multiple enteroendocrine hormones, such as cholecystokinin, pp, gastrin, serotonin, somatostatin, and chromogranin A, but not insulin (22). Although Pdx-1 is an important factor for cellular expression of insulin, we found that this transcription factor alone is not sufficient to induce expression of the gene in intestinal stem cells. It appears that the combined expression of Pdx-1 and Isl-1 is required to trigger insulin gene expression.

We found that the presence of Isl-1 appears to be necessary for the expression of insulin in BYK cells. Differentiation leading to the formation of enteroendocrine cells is associated with the expression of many transcription factors, including NeuroD/Beta2, Pax4, Pax6, Nkx6.1, and Isl-1 (9,10,34). Despite the expression of these factors in addition to neurogenin 3 (Ngn-3; data not shown) in the Pdx-1−expressing YK cells, they did not produce insulin. However, the induction of Isl-1 expression in the YK cells enabled the production of insulin. These observations are consistent with the idea that cellular expression of both Pdx-1 and Isl-1 is necessary for the cells to produce insulin. In fact, this role of Isl-1 in insulin expression is consistent with our recent data showing that a human brain tumor arising from neuroectoderm cells produced insulin without the expression of Pdx-1. The insulin-producing brain tumor cells also expressed Isl-1, Ngn-3, and NeuroD/Beta2 (26) without expression of Pdx-1. These findings suggest that Isl-1, NeuroD/Beta2 and/or Ngn-3 may play important roles during the production of insulin, more so than Pdx-1 and especially in nonislet cells. In support of this possibility, the Isl-1 knockout mouse was found to have no dorsal pancreatic mesenchyme or differentiated islet cells, despite normal expression of Pdx-1 in these animals (7). In addition, Isl-1, which is one of the LIM homeodomain proteins, may synergistically activate the insulin mRNA transcription rate through binding with basic helix-loop-helix proteins (35,36). Together, these experiments and along with our results suggest that Isl-1 is a key factor for the differentiation of cells that produce insulin.

With regard to Pax6, the expression of this gene was found in YK cells, but then disappeared in BYK and Isl-YK cells. Pax-6 is required for the development of pancreatic endocrine cells, duodenal cells positive for gastric inhibitory peptide, and gastrin- and somatostatin-producing cells in the stomach (37). Above all, glucagon gene expression is shown to require this gene expression (38). In fact, although the promoter of glucagon contains two AT-rich sequences that are recognized by the homeodomain containing Cdx-2 and Pax-6 (39), these genes have been shown to bind directly to each other and transactivate synergistically the glucagon gene via their interaction with the G1 element (40). Therefore, our data show that the lack of glucagon expression observed in BYK cells or Isl-YK cells may suggest these cells are more characteristic of β- rather than α-cells.

The finding that both BYK and Isl-YK cells produce multiple hormones except glucagons may point to the lack of terminal differentiation. A potential explanation arises from the known abilities of the transcription factors Pdx-1 and Isl-1 to regulate the expression of the insulin, somatostatin (41), and pp genes (42). To understand how cells restrict their function to the production of insulin, it will be necessary to further understand the control of the noninsulin genes and possibly identify factor(s) that repress their expression. For example, the transcription factor, Pax4 is a potential candidate. This protein is a paired-domain, PD-containing transcription factor that plays an important role in pancreatic β/δ-cell development (43). Overexpression of Pax-4 represses the basal transcription of several genes, including Pdx-1 (44), and this action exerts a negative effect on cellular differentiation leading to the formation of α-cells, but at the same time permits the formation of β-cells (43). Therefore, the absence of Pax-4 may underlie the ability of these cells to express multiple hormones.

Another feature of BYK and Isl-YK cells is the absence of insulin production in response to glucose concentrations in the media. Previous studies have shown that control of the insulin promoter requires at least two different cis-acting motifs designed E and A (45), bound by Pdx-1 (28) and Isl-1 (30) in vitro. These motifs are not unique to the insulin gene and are also found in other pancreatic β-cell genes encoding glucokinase (GK) (46), GLUT2 (47), and Kir. The expression of GK, GLUT2, and Kir is present in both rat pancreatic and Isl-YK cells (data not shown). The high-km GLUT2 and GK have key roles in regulating the rate of glucose flux in β-cells (48) and presumably generate the necessary signals for glucose-regulated insulin secretion (49). However, insulin released from BYK cells into the cell culture medium containing a high concentration of glucose (22.0 mmol/l) was not different from that exposed to low concentrations (5.6 mmol/l) of glucose. The lack of glucose-regulated expression of insulin in the BYK cells did not parallel that present in rat pancreatic cells and will need further studies using Isl-YK cells.

The significance of our findings is better understood when viewed in the light of recent developments in improving treatment for diabetes. The goal of producing surrogate cells that have the ability to produce and secrete insulin is an attractive approach for treating the disease. For example, the introduction of an adenovirus containing Pdx-1 into animals has been shown to facilitate the expression of the insulin gene in liver and ameliorate STZ-induced hyperglycemia (50). The results of other studies have demonstrated glucose-regulated expression of insulin in animals infected with an adeno-associated virus carrying a modified insulin gene fused to the promoter sequence of l-type pyruvate kinase (51). More recently, glucose regulation of a chimeric gene comprised of an insulinotropic polypeptide promoter fused to the insulin gene was introduced into gastric K cells (52). These animals successfully acquired glucose-induced insulin secretion.

Our data represent a small but significant step in developing a novel approach to modifying intestinal stem cells and equipping them with the ability to produce and secrete insulin. For cells to produce insulin, they must express Pdx-1 after treatment with betacellulin or be given the ability to co-express Isl-1. Immature intestinal cells treated in this fashion produce insulin, but the expression and secretion of the hormone is not regulated by glucose. Glucose-regulated expression of insulin in the cells is the focus of future studies and is essential for further development of the use of this approach to treat diabetes.

This work was supported, in part, by a grant for scientific research from the Ministry of Education, Science, and Culture of Japan (no. 10671064).

We would like to thank Dr. Yoshitaka Kajimoto (Department of Internal Medicine and Therapeutics, Osaka University Graduate School of Medicine) for providing us with Pdx-1 cDNA and anti-serum. We also thank Dr. P.G. Traber (University of Pennsylvania) for providing the Cdx-2 cDNA, Dr. Wakabayashi (Gunma University) for anti-insulin antibodies, and Dr. C.V.E. Wright (Vanderbilt University School of Medicine) for kind advice on Pdx-1. We also thank Takefumi Yamamoto and Hiroshi Okuno in the Central Research Laboratory of Shiga University of Medical Science for their technical support in fluorescence and electron microscopy.