To determine the role of Notch1 signaling in regulation of PDX-1 expression, we transfected Notch1 intracellular domain (NICD), activated form of Notch1, alone or NICD plus NICD shRNA (shRNA^NICD) into PDX-1 stable HEK293 cells [15]. Western blotting analysis showed that overexpression of NICD resulted in a significant increase of PDX-1 expression compared to that of mock transfection (Figure 1A, column 2 vs. column 1), while co-transfection of shRNA^NICD with NICD only resulted in a mild increase of PDX-1 expression (Figure 1A, column 3 vs. column 1). These data indicate that NICD, activated Notch1, acts as a positive regulator for PDX-1 expression and that NICD shRNA inhibits the enhancing effect of NICD on PDX-1 expression. To confirm the positive regulation of PDX-1 expression by Notch1 activation, we examined the effect of NICD shRNA (shRNA^NICD) on endogenous PDX-1 expression in mouse insulinoma β-TC-6 cells which have high expression level of PDX-1 [41]. shRNA^NICD was transfected into β-TC-6 cells. Western blotting analysis showed that NICD was abundant in β-TC-6 cells and transfection of shRNA^NICD resulted in a significant knockdown of endogenous NICD (Figure 1B, top panel). In parallel, PDX-1 expression level was also significantly decreased by transfection of shRNA^NICD in comparison to the mock transfection (Figure 1B, middle panel), further supporting the concept that activated Notch1 (NICD) is a positive regulator for PDX-1 expression.

The underlying mechanism by which Notch1 activation up-regulates PDX-1 expression is not known. It has been well established that, upon ligand binding, NICD is released from the Notch protein and moves to the nucleus, where it regulates gene expression by interacting with and activating transcription factors such as CBF1/RBP-Jκ/Suppressor of Hairless/LAG-1 (CSL) transcription factor and p300 [52,53]. Given that PDX-1 is a potent transcription factor for PDX-1 itself, it is, thus, possible that NICD may up-regulate the transcriptional activity of PDX-1, which, in turn, enhances the expression of PDX-1. However, we cannot exclude the possibility that NICD might enhance PDX-1 expression by up-regulating other transcription factors, which target PDX-1. Further studies warrant a better understanding of the molecular basis for the positive regulation of PDX-1 expression by Notch1 activation.

PDX-1 plays an essential role in insulin expression and secretion [3,4] and cell proliferation [54]. To determine the functional relevance of the positive regulation of PDX-1 by NICD, we sought to examine the effect of NICD on PDX-1-mediated insulin expression/secretion and cell proliferation in β-TC-6 cells. β-TC-6 cells were transfected with NICD or shRNA^NICD. Immunohistochemistry analysis showed that overexpression of NICD further increased NICD expression in β-TC-6 cells, while overexpression of shRNA^NICD inhibited NICD expression (Figure 2A, top panel). As expected, PDX-1 expression level was increased by transfected NICD and decreased by transfected shRNA^NICD (Figure 2A, middle panel). Moreover, we found that NICD-enhanced PDX-1 expression was accompanied by increased insulin expression, and NICD knockdown-induced inhibition of PDX-1 expression was accompanied by decreased insulin expression (Figure 2A, bottom panel). Insulin ELISA analysis showed that NICD overexpression significantly increased glucose stimulated insulin secretion (GSIS) in comparison to mock transfection, whereas shRNA^NICD led to significant inhibition of insulin secretion in β-TC-6 cells (Figure 2B). By performing MTS assays, we found that, at 24, 48 and 72 hour after transfection, overexpression of NICD resulted in significant increases of β-TC-6 cell proliferation by 174 ± 18.2%, 185 ± 30.8% and 166 ± 21.3%, respectively, as compared to the mock transfection (Figure 2C). Conversely, transfection of shRNA^NICD led to significant decreases of cell proliferation by 63 ± 15.3%, 49 ± 8.8% and 56 ± 12.6% at each time point (Figure 2C). All these data indicate that Notch1 activation exerts a positive effect on insulin expression/secretion and cell proliferation likely through a mechanism of up-regulating PDX-1 expression in β-TC-6 cells.

To further confirm the up-regulation of PDX-1 expression by Notch1 activation, we generated a transgenic β^NICD+/+ mouse model by crossing LSL-ROSA^NICD mice [55] and RIP1-Cre mice, which resulted in β cell-specific expression of NICD within the islets. To determine the effect of NICD on PDX-1 and PDX-1-mediated cellular functions, the β^NICD+/+ mice were treated at six months of age with or without liposomal shRNA^mNICD at a dose of 35 μg per mouse via tail vein injection. The mice tolerated the treatment well with no toxic effects observed. Forty eight (48) hours after treatment of shRNA^mNICD, the mice were sacrificed and the pancreata were isolated, fixed and prepared for immunohistochemistry analysis using anti-NICD, PDX-1, insulin, PNCA and p27 antibodies, respectively. As shown in Figure 3A, Cre activation induced a significant increase of NICD expression in islet cells of β^NICD+/+ mice as compared to that of control mice, and systemically-delivered shRNA^mNICD efficiently knocked down NICD (top panel). Cre activation-induced overexpression of NICD was accompanied by increased expression of PDX-1, insulin and proliferating cell nuclear antigen (PCNA) and a decreased expression of p27, an inhibitor of cyclin-dependent protein kinase, in islet cells of β^NICD+/+ mice as compared to those of control mice. This is consistent with the finding that Notch directs the transcription of the E3 ubiquitin ligase S phase kinase-associated protein 2 (SKP2), which results in a decrease of p27 protein levels and an increase of cell proliferation [56]. In contrast, knockdown of NICD by systemically-delivered shRNA^mNICD resulted in reduction of the expression of PDX-1, insulin and PCNA and increased expression of p27. These in vivo data further support the concept that Notch1 activation up-regulates insulin expression/secretion and cell proliferation through a mechanism involving enhancing PDX-1 expression. We noticed a previous study that PDX-1 expression is not affected by NICD in both Rosa^N°^tch/+;Pdx1-Cre and Rosa^N°^tch;Pdx1-CreER mice [55]. The discrepancy may be due to the different Cre systems used (Pdx1-Cre and Pdx1-CreER vs. RIP1-Cre). In addition, the mice we analyzed were six months old, while the mice Murtaugh et al. analyzed were newborn or embryo [55].

Given the essential role of PDX-1 in regulation of insulin expression/secretion, we hypothesized that NICD-induced increase of PDX-1 expression may lead to the development of hypoglycemia and hyperinsulinemia in β^NICD+/+ mice. To test this hypothesis, glucose and insulin levels were measured in β^NICD+/+ mice at six months of age. We found that glucose levels were significantly decreased and insulin levels were significantly increased in β^NICD+/+ mice compared to those in control mice, while treatment of the β^NICD+/+ mice with shRNA^mNICD rescued/restored glucose and insulin levels close to wild type levels (Figure 3B).

Next, we performed IPGTT in control (WT) and β^NICD+/+ mice at six months of age to compare glucose tolerance. Basal insulin levels in β^NICD+/+ mice were significantly higher than those in control mice (Figure 3C, top panel), whereas fasting glucose levels in β^NICD+/+ mice were significantly lower than those in control mice (Figure 3C, bottom panel). Following intraperitoneal glucose injection, systemic insulin levels in β^NICD+/+ mice were significantly higher than those in control mice at 15, 30, 120 and 240 minute post-injection (Figure 3C, top panel). Glucose levels in β^NICD+/+ mice were significant lower than those in control mice at 15, 30, 60, 240 minute post injection of glucose (Figure 3C, bottom panel). These data indicate that overexpression of NICD resulted in fasting hypoglycemia and hyperinsulinemia and alterations in insulin and glucose responses to IPGTT.

Taken together, our studies show that NICD-enhanced PDX-1 expression was accompanied by increased insulin expression/secretion and cell proliferation in β cells of transgenic β^NICD+/+ mice, resulting in hyperinsulinemia and hypoglycemia, as well as an altered response to IPGTT in these mice. A single systemic treatment with liposomal NICD shRNA resulted in in situ knockdown of NICD and PDX-1, leading to reversal of hyperinsulinemia and hypoglycemia. Moreover, there were no overt toxic side effects following the single treatment of liposomal NICD shRNA. This leads further support that Notch1 regulates PDX-1 expression and suggests that Notch1 could be a potential therapeutic target for hypoglycemic disorder using an RNAi platform. This is consistent with our recent studies showing that three-cycle treatment of PDX-1 shRNA reverses hyperinsulinemia and hypoglycemia in an immune-competent mouse model of islet neoplasia and results in significant reduction of tumor volume and improved survival in a human pancreatic cancer xenograft mouse model [11]. Our results also show that the transgenic β^NICD+/+ mice with conditional induction of NICD in islet cells may serve as a potential novel mouse model for the development of islet cell neoplasia.

It has been reported that Notch1 is over-expressed in neuroendocrine tumors. We have recently shown that PDX-1 is markedly overexpressed in PNETs [41]. We, therefore, wanted to examine whether Notch1 activation is correlated with PDX1 overexpression in PNETs by performing IHC analysis of 35 human PNET specimens. We found that Notch1 was activated in 31 of the 35 specimens (89%) as evidenced by NICD expression and that PDX-1 was expressed in all PNET specimens studied. Semi-quantitative analysis of the immunofluorescence images for NICD and PDX-1 using software ImageJ revealed that the median expression levels of PDX-1 were 85.2% and that of NICD was 63.8%. Moreover, we found that higher expression levels of NICD were accompanied by higher expression levels of PDX-1 (Figure 4A, right panel vs. left panel). Pearson correlations analysis showed significant correlation between PDX-1 and NICD expression (Figure 4B, R = 0.933; p < 0.01). These data further support the concept that Notch1 activation up-regulates PDX-1 expression and that Notch1 activation may contribute to the overexpression of PDX-1 in PNETs. Notch1 has been reported to be expressed in PNET with accompanied expression of HES1, a downstream target of Notch1 [57]. Our studies showed that the majority of human PNET specimens expressed NICD in correlation with PDX-1 overexpression. Moreover, we found that NICD-enhanced PDX-1 expression was accompanied by increased cell proliferation in mouse insulinoma β-TC-6 cells and resulted in increased PCNA in the islets of transgenic β^NICD+/+ mice. Given the oncogenic function of PDX-1 [15], our study suggests that Notch1 may play an oncogenic role in PNETs via up-regulating PDX-1. Cooperation with other oncogenes is one mechanism by which Notch1 exerts its oncogenic function. Our studies, thus, provide that up-regulation of PDX-1 is one novel mechanism by which Notch1 exerts its oncogenic functions.

Human embryonic kidney 293 (HEK 293) and mouse insulinoma β TC-6 cells were purchased from the American Type Culture Collection (ATCC, Bethesda, MD, USA) and maintained in Dulbecco’s modified Eagle medium (Gibco-BRL, Bethesda, MD, USA) supplemented with 100,000 U/L of penicillin, 100,000 µg/L of streptomycin, and 10% fetal bovine serum (FBS). pBS-Notch1-IC (mouse) was purchased from Addgene (plasmid 15079) and subcloned into pcDNA3.1 expressing vector to produce pcDNA^mNICD. Human PDX-1 cDNA was polymerase chain reaction (PCR)-amplified from human islet cells and cloned into the FLAG epitope containing pCMV5 expression vector which was provided by Dr. Narasimhaswamy Belaguli (Baylor College of Medicine). Both mouse Notch1-IC and human PDX-1 were subcloned into retroviral vector pQICXIP (Clontech, Mountain View, CA, USA) for the purpose of retrovirus production. Five mouse Notch shRNAs targeting NICD (shRNA^mNICD) were designed, synthesized and cloned into pSuper vectors. One of them targeting AGGCAACAGTGAAGAAGAA (starting 5661 of open reading frame of Notch1) was selected for the current study based on its most efficient knockdown effect. A scrambled shRNA was used as control. Goat anti-PDX-1 and rabbit anti-p27 and anti-PCNA antibodies were purchased from Santa Cruz Biotechnology Inc (Santa Cruz, CA, USA). Rabbit anti-activated Notch1 antibody that only recognizes the cleaved intracellular (activated) form of Notch1 (NICD) was purchased from Abcam (Cambridge, MA, USA). Monoclonal anti-insulin antibody was purchased from Sigma-Aldrich (St. Louis, MO, USA). Goat anti-rabbit antiserum and sheep anti-mouse antiserum conjugated with horseradish peroxidase were purchased from Amersham Life Science Inc. (Arlington Heights, IL, USA). Rabbit anti-goat IgG was obtained from Zymed Laboratories Inc. (South San Francisco, CA, USA). PNET specimens were as previously described [58].

Transient transfection in HEK293 and β-TC-6 cells was performed with 24 µg of plasmid DNAs using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA). Stable cell lines were established using retroviral expression system with pQCXIP retroviral vectors (Clontech, Mountain View, CA, USA). pQCXIP expressing PDX-1 was constructed and confirmed, followed by transfection into the AmphoPack 293 cell line (Clontech, Mountain View, CA, USA) with Lipofactamine 2000 to produce retrovirus. The cells were infected with the supernatant containing the reconstructed retrovirus and selected with 10 mg/mL puromycin for two weeks. pcDNA^mNICD with shRNA^mNICD or shRNA^{mNICD_scrambled} were co-transfected into β-TC-6 cells.

Twenty-four hours after transfection, cells were replated into 96-well cell culture plates at 5 × 10³ cells/well. Cell proliferation was determined by MTS assay (Promega, Madison, WI, USA) at 24, 48, and 72 hour after transfection. The absorbance was read using a Multiskan EX plate reader (Thermo Electronic Corp, Franklin, MA, USA) at 492 nm. Levels of proliferation were determined according to the following formula: percentage of proliferation = 100 × B/A, where A is the absorbency at 492 nm of control cells, and B is the absorbency at 492 nm of treated cells.

Transfections were performed at 10-cm plates for 48 hours as described as previously. Cells were collected using a cell scrapper in an ice-cold phosphate-buffered saline (PBS) solution containing 5 mM EDTA. Cells were lysed in a buffer containing 20 mM HEPES (pH 7.4), 2 mM EGTA, 50 mM β-glycerophosphatase, 1% Triton X-100, 10% glycerol, 1 mM DTT, 2 mg/mL of leupeptine, 5 mg/mL of aprotinin and 1 mM phenylmethylsulphonyl fluoride after two washes with PBS. Supernatants were collected after centrifugation at 100,000 g for 15 minutes at 4 °C. Twenty micrograms of cell lysates were analyzed using a sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel (SDS-PAGE). After electrophoresis, gels were removed and blotted onto polyvinylidene fluoride (PVDF) membranes. Membranes were then incubated with 5% non-fat milk buffer and probed with antibodies against NICD and PDX-1. Immunocomplexes were visualized by enhanced chemiluminescence detection using horseradish peroxidase-conjugated secondary antibodies. Images were captured using the UVP imaging system, and the band was analyzed using ImageJ software.

LSL-Rosa^NICD transgenic mice were kindly provided by Dr. Brendan Lee (Baylor College of Medicine, Houston, TX, USA) with permission of use from Dr. Douglas Melton (Harvard University, Boston, MA, USA) [55]. The mice were maintained on a hybrid 129× C57BL/6 background. β^NICD+/+ mice were generated by crossing LSL-Rosa^NICD and RIP1-Cre mice with successive generation screening. The β^NICD+/+ mice at age of 6 months were treated with intravenous liposomal shRNA^mNICD at dose of 35 μg per mouse via tail vein injection. Forty-eight (48) hours after injection, blood was collected and mice were sacrificed for tissues collection. Animals were used in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All mice were housed in a specific pathogen-free facility and under light-, temperature-, and humidity-controlled conditions. These studies were performed under IRB approval.

50 μL whole blood samples were collected and spun to separate the serum. Serum samples were stored at −20 °C until completion of experiments. Glucose levels were measured using a Beckman-Coulter Glucose Analyzer 2 (Coulter-Beckman, Fullerton, CA, USA), and presented as mean ± S.E.M. in mg/dL. Insulin levels were determined using a mouse insulin ELISA kit from Mercodia (Linco Research, St. Charles, MO, USA) and presented as mean ± S.E.M. in μg/L.

Control (WT) and β^NICD+/+ mice at age of six months were fasted 6 hours before collection of blood samples as T₀. Grouped mice were then given 1.2 g glucose/kg body weight via ip injection, followed by collection of blood samples at 15, 30, 60, 120 and 240 minute after injection of glucose. Glucose and insulin levels were measured as described above.

Fluorescein isothiocyanate-conjugated anti-rabbit IgG antibodies were purchased from Sigma (St. Louis, MO, USA). PNET tumors were fixed in 4% paraformaldehyde at 4 °C overnight. After treatment with 70% alcohol, tissue blocks were embedded in paraffin, and tissue sections were prepared. For immunostaining, sections were deparaffinized in xylene and hydrated gradually through graded alcohol. Slides were then placed in a humidified chamber, overlaid with diluted antibodies (1:100) against activated Notch1, PDX-1, insulin, PCNA or p27, and incubated overnight at 4 °C. After washing with PBS, slides were incubated with FITC-conjugated anti-rabbit or Cy3-conjugated anti-goat secondary antibody for 1 hour, then washed with PBS and mounted with cover slides. Images were recorded using a digital camera (Diagnostic Instruments Inc., Sterling Heights, MI, USA) on a fluorescent microscope (Olympus IX70; Olympus Optical Co Ltd., Japan). Immunostaining for PDX-1 expression was semi-quantified using ImageJ.

Student’s t-test was used to analyze difference in means of the continuous data. All numeric data are expressed as mean ± SEM, with p < 0.05 indicating significance. Pearson correlation analysis was used to analyze the correlation between PDX-1 and NICD expression