Promotion of β-Cell Differentiation by Conophylline in Fetal and Neonatal Rat Pancreas

Conophylline was isolated from the leaves of Ervatamia microphylla and purified as previously described (18). Briefly, a chloroform extract of the leaves was chromatographed on a silica gel column with hexane:ethyl acetate (1:2) to produce the active fraction. After the active fraction was dissolved in MeOH and the insoluble fraction removed, it was then applied onto a Toyopearl HW-40 column and eluted with methanol to produce the active material. Finally, the crude material was purified by centrifugal partition chromatography and freeze dried from t-butyl alcohol to produce white powder. Recombinant human activin A was provided by Dr. Eto of Ajinomoto (Kawasaki, Japan).

Pregnant Wistar rats were purchased from Japan SLC (Shizuoka, Japan). The morning of the discovery of the vaginal plug was designated as embryonic day 0.5 (E0.5). Pregnant rats at 14.5 days of gestation were killed by cervical dislocation. Embryos were removed from the uterus. Embryonic pancreata were dissected from the digestive tracts under a stereoscopic microscope.

Pancreatic rudiments were grown in three-dimensional collagen gels in RPMI-1640 medium supplemented with 10% heat-inactivated FCS, 2 mmol/l glutamine, 50 units/ml penicillin, and 75 mg/ml streptomycin. In some experiments, 2 nmol/l activin A or 0.1 μg/ml conophylline was added to the culture medium. Cultures were maintained at 37°C in a humidified atmosphere of 95% air and 5% CO₂. Medium was replaced every 48 h. At the indicated times, the rudiments were photographed and fixed for immunohistochemistry as described below. The DNA content of the rudiment was measured by fluorometric assay (19).

One-day-old male neonatal Wistar rats received a single intraperitoneal injection of 85 μg/g of streptozoficin (STZ) (Wako, Tokyo, Japan) freshly dissolved in 0.05 mmol/l citrate buffer (pH 4.5). The day of STZ injection was designated as day 0. The pups were left with their mothers until 4 weeks. All neonates were tested 1 day after the STZ injection (day 1) for blood glucose using Accu-Chek Active (Roche Diagnostics, Mannheim, Germany). The animals were included in this study only if their blood glucose concentration was between 200 and 350 mg/dl on day 1. Neonatal STZ-induced diabetic rats were injected subcutaneously with 5 μg/g conophylline or vehicle on days 1, 3, 5, and 7. Fasting blood glucose concentration and body weight were measured daily for the first week and then once a week for up to 8 weeks. The plasma insulin concentration was measured on day 8 and week 8 with an insulin assay kit (Morinaga, Yokohama, Japan), using rat insulin as standard. Eight weeks after the STZ treatment, an intraperitoneal glucose tolerance test (2 g/kg body wt) was performed after 14 h fasting. The experimental protocol was approved by the animal care committee of Gunma University.

The pancreatic rudiments and pancreas were fixed at 4°C in 10% formaldehyde in PBS for 24 h and embedded in Paraplast (Labonord, Templemars, France). Each pancreatic block was serially sectioned (6 μm) throughout its length. Immunostaining was performed as previously described (6). The sections were examined and photographed with Axiophoto (Carl Zeiss, Thornwood, NY). The antibodies used in this study had the following dilutions: guinea pig anti-porcine insulin (provided by Dr. T. Matozaki, Gunma University, Maebashi, Japan) 1:1,000, rabbit anti-human pancreatic polypeptide (PP) (Peptide Institute, Mino, Japan) 1:1,000, monoclonal anti-amylase (Sigma, St. Louis, MO) 1:200, and rabbit anti-human pancreatic duodenal homeobox (PDX)-1 (Dr. Y. Kajimoto, Osaka University, Osaka, Japan) 1:5,000. The fluorescent secondary antibodies were as follows: fluorescein anti–guinea pig antibodies and anti-rabbit antibodies (Alexa Fluor; Molecular Probes) 1:250.

To determine the number of cell types per pancreatic rudiment, serial 6-μm sections were cut and collected on multiwell glass slides. One of two consecutive sections was analyzed by immunocytochemistry for a given antigen (20). A total of 50 sections were analyzed per rudiment. Only the immunoreactive cells for which a nucleus was clearly visible were counted. The results of each experimental point were obtained by measuring the absolute number of each cell type in at least four pancreatic rudiments. An islet-like cell cluster (ICC) is designated as a cluster of insulin-positive cells containing less than eight cells across. To measure the number of ICCs in neonatal STZ-induced diabetic rats, ICCs in 20 fields per section were counted and at least 5 sections were counted per rat. To determine the β-cell mass, the area of insulin-producing cells, as a percentage of the pancreatic area, was measured by using National Institutes of Health Image software. To calculate the β-cell mass, this relative volume of β-cells was multiplied by the pancreatic weight (6). The insulin content was measured after extraction with 0.1 mol/l HCl as previously described (6). Data are presented as means ± SE, and statistical analysis was performed using Student’s t test.

Total RNA was extracted from cultured pancreatic rudiments at different time points using TRIzol Reagent (Invitrogen Japan, Tokyo, Japan). Messenger RNA was extracted using the Quick Prep Kit (Pharmacia LKB Biotechnology, Piscataway, NJ). RNA or mRNA samples were pretreated with DNase to remove contamination of genomic DNA. First-strand cDNA was synthesized by using a preamplication system for the SUPERSCRIPT First-Strand Synthesis System (Invitrogen Japan). Real-time PCR was performed as previously described (21). The expression of rat insulin, PP, and glyceraldehyde-3-phosphate dehydrogenase in each sample was quantitated in separate tubes with the following primers: rat insulin I (sense and antisense were 5′-GGCTTTTGTCAAACAGCACCTTTG3′ and 5′-AGTTGCAGTAGTTCTCCAGTTGG-3′, respectively), rat PP (sense and antisense were 5′-TGAACAGAGGGCTCAATACGAAAC-3′ and 5′-AGACGAAGGGAGGCTACAAATCC-3′, respectively), rat glyceraldehyde-3-phosphate dehydrogenase (sense and antisense were 5′-CATGACCACAGTCCATGCCATC-3′ and 5′-CACCCTGTTGCTGTAGCCATATTC-3′, respectively). No PCR products were detected in the real-time PCR procedure without reverse transcription, indicating that the contamination of genomic DNA was negligible. Gels of the PCR products after quantitation of the indicated genes by real-time PCR showed single bands (264, 236, and 451 bp, respectively), the sizes of which were the same as expected. To measure the changes in the mRNA expression, mRNA was obtained from control and activin- and conophylline-treated tissues every 24 h, and each datum was standardized by the ratio of E14.5 rat mRNA.

Figure 1A shows the macroscopic morphology of pancreatic rudiments cultured in collagen gel for 72 h. The size of the rudiments was almost the same in the control and activin- and conophylline-treated rudiments. The ratio of the DNA content of the control and activin- and conophylline-treated rudiments was 1:0.94:1.06 (n = 3). At E14.5, the pancreatic rudiments consisted of duct-like epithelia surrounded by a considerable amount of mesenchymal tissue (Fig. 1B, b). A few insulin-positive cells were scattered around the epithelia (Fig. 1B, a). After a 72-h culture, duct-like epithelia expanded and formed cysts (Figs. 1B–D). The number of insulin-positive cells increased slightly (Fig. 1B and C). In activin-treated tissues, expansion of duct-like epithelia was moderate, and insulin-positive cells formed small clusters (Fig. 1B–E, f). In conophylline-treated tissues, expansion of duct-like epithelia was scarce and many islet-like cell clusters were observed (Fig. 1B–G, h). In addition, formation of small ducts was observed.

Since activin A converts amylase-secreting AR42J into PP-secreting cells (22), we first measured the changes in the number of PP-positive cells. In specimens obtained at 72 h, the number of PP-, insulin-, or amylase-positive cells was measured. Regarding the number of PP-positive cells, no statistical significance was obtained among the three groups (Fig. 1C). The number of insulin-positive cells in activin-treated tissue was significantly higher than that of the control. Likewise, the number of insulin-positive cells in the conophylline-treated rudiment was significantly greater than that of the control (Fig. 1C). Activin A and conophylline significantly reduced the number of amylase-positive cells. Real-time RT-PCR analysis further confirmed these results. The PP expression of the control group increased fourfold at 72 h, whereas that of the activin- and conophylline-treated groups did not increase significantly (Fig. 2). The insulin expression in the control group increased only slightly. In contrast, the expression of insulin in activin- and conophylline-treated tissues increased ∼200-fold at 72 h.

We then studied the effect of conophylline in rudiments cultured for a longer period. Conophylline is light sensitive and degraded by ultraviolet rays. We tested various conditions and found that an addition of conophylline every 48 h provided the maximal effect (data not shown). Accordingly, we added conophylline every 2 days and cultured pancreatic rudiments for 10 days. In the control, epithelia expanded and formed cysts and insulin- or PDX-1–positive cells were scattered around them (Fig. 3A). In conophylline-treated tissues, epithelial expansion was scarce and the numbers of insulin- and PDX-1–positive cells increased markedly around duct-like structures (Fig. 3B). Moreover, many ICCs were observed. Quantitative analysis using the National Institutes of Health Image software showed that the relative area for PDX-1–and insulin-positive cells in conophylline-treated tissues were significantly greater than those of the control, and the ratio of area of insulin to PDX-1 in conophylline-treated tissues was lower than that of the control (Fig. 3C).

We then examined the effect of conophylline in vivo using neonatal STZ-induced diabetic rats. As shown in Fig. 4A, the plasma glucose concentration peaked on day 2 in neonatal STZ-induced diabetic rats, with a peak value of ∼400 mg/dl. The plasma glucose concentration then decreased gradually but remained significantly higher than that in the normal rats at 2 months of age (Table 1). We administered conophylline on days 1, 3, 5, and 7. As shown in Fig. 4A, the plasma glucose concentration was significantly lower on day 2 (P < 0.01) and thereafter. At 2 months of age, the plasma glucose concentration was significantly lower in conophylline-treated rats compared with STZ-induced diabetic rats (Table 1). An intraperitoneal glucose-loading test was done at 2 months. As shown in Fig. 4B, the plasma glucose response to glucose loading was greatly improved in conophylline-treated rats. The insulin secretion in response to glucose loading was also improved in conophylline-treated rats, but the response was still delayed compared with normal rats. As shown in Table 1, the insulin content and the β-cell mass were significantly increased in conophylline-treated rats. Immunohistochemical analysis revealed that the percent of PDX-1–positive ductal cells was markedly increased (Figs. 5A and B), and the number of ICCs was also significantly greater in conophylline-treated rats compared with vehicle-treated rats. Note that conophylline did not affect the number of BrdU-positive ductal cells or BrdU-positive β-cells in the islets (data not shown).

Here, we demonstrate that conophylline induced differentiation of fetal rat pancreas in organ culture. Previous studies using pancreatic anlage showed that mesenchymal factors promoted differentiation of pancreatic progenitors to exocrine lineage (22), and removal of mesenchymal tissue promoted differentiation to the endocrine lineage (20). Miralles et al. (20) showed that conversion of progenitor cells to endocrine or exocrine cells was modulated by mesenchyme-derived follistatin, an inhibitor of activin A. Collectively, these results imply that differentiation of progenitor cells is regulated by the activin-follistatin system, which acts as an autocrine/paracrine mechanism. In agreement with this notion, activin A induces differentiation of AR42J cells, a model of pancreatic progenitor cells, into endocrine cells (22). Conophylline is also capable of inducing differentiation of AR42J cells into insulin-producing cells. It had been postulated that conophylline reproduces the action of activin A (16). In an organ culture system, conophylline markedly augmented the expression of the insulin gene and significantly increased the number of insulin-positive cells (Figs. 1 and 2). This was also the case in the activin-treated group, which suggests that activin A and conophylline suppressed the effect of mesenchymal cells and promoted differentiation to the endocrine lineage. Because conophylline activates p38 mitogen–activated protein kinase and induces the expression of neurogenin 3, as does activin A (16), conophylline may induce differentiation of pancreatic progenitor cells using the same pathway as activin (16). Hence, conophylline also reproduces the action of activin A in organ culture. Although conophylline increased the number of insulin-positive cells, unlike activin A, it did not significantly induce cystic expansion of the duct-like structure. Instead, there were many small ducts. This indicates that conophylline increased branching morphogenesis of the ducts. It is known that activin A inhibits branching morphogenesis of the pancreatic ducts (14). Conophylline may be less potent in inhibiting branching morphogenesis than activin A. Previous studies by Gittes et al. (23) and Miralles et al. (20) showed that mesenchymal tissue promoted differentiation of pancreatic anlage to the exocrine lineage. On the other hand, typical acinar structure was not formed and the duct-like structures expanded as cystoids in our experiments. This may have resulted from the cultivation conditions of the pancreatic anlage. Because insulin gene expression at 72 h measured by real-time PCR increased remarkably, we stimulated pancreatic anlage with conophylline every 2 days to induce them into insulin-positive cells. As a result, insulin-positive cells increased in the duct-like structures and ICCs invaginated from them. These ICCs were PDX-1 positive, and many of their cells expressed both insulin and PDX-1. These findings suggest that conophylline acts on pancreatic progenitor cells in the duct-like structures and promotes differentiation into insulin-producing cells. Although conophylline increased the number of insulin- and PDX-1–positive cells, the ratio of insulin to PDX-1 was reduced. Conophylline may act on the early step of β-cell differentiation. In this study, we detected the expression of PP in pancreatic rudiments. This result appears to be contradictory to the generally accepted view that PP is first expressed in the pancreas at birth. However, Herrera and colleagues (24,25) have shown that PP promotor is activated in earlier time points, and PP-positive cells are detected in some studies (26,27). Our result is in line with these reports.

An in vivo study using neonatal STZ-induced diabetic rats further indicates that conophylline promotes differentiation of β-cells and improves glucose metabolism. This model has been widely used to study regeneration of pancreatic β-cells (28–33). Glucagon-like peptide 1 and exendin 4 have been shown to ameliorate diabetes in this model (32). The present results demonstrate that conophylline increased β-cell mass and insulin content and is effective in the promotion of β-cell regeneration. Since conophylline increased the numbers of PDX-1–positive ductal cells and ICCs, it is quite likely that conophylline promotes neogenesis of β-cells by acting on the early stage of β-cell differentiation.

In summary, conophylline reproduces the effect of activin A and acts on pancreatic progenitor cells located in the vicinity of duct epithelium. It promotes differentiation to endocrine cells. Conophylline may be useful for promotion of differentiation of pancreatic β-cells.

The present study was supported by a grant from the Project for Realization of Regeneration Medicine and a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture of Japan.

The authors are grateful to Mayumi Odagiri for secretarial assistance.