C57BL/6 wild-type male mice, 5 fasted for 18 h, 4 fasted for 18 h and refed at 4 h, and 5 fasted for 18 h and refed at 7 h, were used for the proposed studies. An additional 3 animals per group, fed regular diet (RD) or HF diet, were utilized for the diet studies. All animals were kept in a 12 h dark/light cycle and control animals were fed RD (ad libitum). All procedures were approved by the University of Toledo Animal Care and Utilization Committee. For the feeding experiments, 3-month-old wild-type mice were fed a RDand a high-fat (HF) diet for 3 months (Research Diets Inc., New Brunswick, NJ, USA; Cat #D12451 and D12079B respectively). The RD contains 69% carbohydrates and 11% fat and the HF diet contains 35% carbohydrates and 45% fat.

The enteroendocrine intestinal cell line, STC-1, was obtained from ATCC with permission from Douglas Hanahan (UCSF-CA) and maintained at 37 °C and 5% CO₂ in DMEM medium containing 1% penicillin and streptomycin. All experiments were performed on cells passaged five times at 80% confluency. LY294002, an inhibitor of the PI3K-AKT kinase signaling pathway, and UO126, a MAP kinase inhibitor, were both obtained from Cell Signaling Technology (Beverly, MA, USA) and were used at concentrations of 10 nℳ from a 10 mℳ stock reconstituted in DMSO. Aliquots were stored at −20 °C and vehicle was used for control experiment.

The 2.9 and 0.210 kb fragments of the human GIP promoter were cloned into the luciferase-containing pGL4 reporter plasmid, as previously described.²⁸ STC-1 cells were plated in 6-well plates (2 × 10⁵ cells/well) and allowed to adhere for 24 h. Thereafter, cells were transfected with a mixture of 4 μg of plasmid DNA (pBABE) containing menin (Addgene Inc., Cambridge, MA, USA), 0.5 μg of the control plasmid (PGL4) or GIPLuc reporter plasmid²⁸ using Lipofectamine 2000 reagent according to the manufacturer's protocol (Invitrogen, Carlsbad, CA, USA). After 48 h of incubation with complete media, luciferase and renilla activities were assayed according to manufacturer's instruction (Dual Luciferase Reporter Assay; Promega-Madison, WI, USA). Firefly luciferase activity was normalized to renilla luciferase expression and is presented as percent activity over samples transfected with pGL4. All experiments were analyzed in duplicate in at least three separate experiments. Menin small interfering RNA (siRNA) oligos and the non-targeting control siRNA and reagents were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA) and the assay was performed according to the manufacturer's protocol.

STC-1 cells were plated in 6-wells in triplicate, serum starved overnight and exposed to 5 or 20 nℳ GIP for 24 h. The concentrations of proteins from total cell lysates were quantified prior to analysis on 10 or 4–12% gradient SDS-PAGE and immunoprobed with specific antibodies. Rabbit anti-menin (Bethyl Labs/Abcam, Cambridge, MA, USA),²⁹ phospho-p38 (Cell Signaling, Beverly, MA, USA),³⁰ goat anti-GIP (Santa Cruz Biotechnology, Santa Cruz, CA, USA)³¹ and GAPDH (Sigma-Aldrich, St, Louis, MO, USA) antibodies were used. Proteins were detected using an Odyssey Infrared Imaging System (Lincoln, NE, USA), using corresponding secondary antibodies conjugated to fluorescent dyes of different wavelengths.

STC-1 cells were plated in 6-well plates in triplicate, serum starved overnight and transfected with siRNA oligos for menin or scrambled controls according to manufacturer's protocol (Santa Cruz Biotechnology). The transfected cells were exposed to LY294002, an inhibitor of PI3K-AKT kinase signaling, or the MAP kinase inhibitor, UO126, overnight prior to harvest and determination of protein concentration. 20 μg of protein in a volume of 10 μl total lysate and 10 μl media were used for a rat/mouse GIP ELISA (Cat. #EZRMGIP-55K, Millipore, Billerica, MA, USA) according to manufacturer's protocol.

RNA was extracted using TRIzol (Life Technologies, Grand Island, NY, USA). Following DNAse digestion (DNAfree, Applied biosystems/Ambion, Austin, TX, USA), 1 μg RNA was transcribed into cDNA in a 20 μl reaction using High-Capacity cDNA achieve kit (Applied Biosystems/Ambion) and amplified (ABI 7900 HT system). PCR was performed in a 15 μl reaction, containing 5 μl cDNA (1/10 diluted), 1 × SYBR Green PCR Master Mix (Applied Biosystems/Ambion) and 300 nℳ of each primer. Primers are listed in 5'–3' orientation.

Menin forward primer: TCATTGCTGCCCTCTATGCC

Menin reverse primer: TCCAGTTTGGTGCCTGTGATG

GIP forward primer: GTGGCTTTGAAGACCTGCTC

GIP reverse primer: AAGTCCCCTCTGCGTACCTT

GAPDH forward primer: CCACCAGCCCCAGCAAGAGC

GAPDH reverse primer: GGCAGGGACTCCCCAGCAGT

18S forward primer: ATACATGCCGACGGGCGCTG

18S reverse primer: GGGAGGGAGCTCACCGGGTT

Ct values (cycle threshold) were used to calculate the amount of amplified PCR product relative to GAPDH and 18S. The relative amount of mRNA was calculated as 2^–DCT. Results are expressed as transcript of interest mean normalized to GAPDH and 18S with s.e.m.

One-cm tissue starting from the pyloric sphincter including the brunners gland and into the proximal duodenum was taken from the upper section of the gut from 5 C57BL/6 wild-type male mice fasted for 18 h and 5 C57BL/6 wild-type male mice fed ad libitum. Additional 2 sets of mice for each time point were also used for all described studies and consisted of a group of mice fasted for 18 h, refed and sacrificed after 4 h of feeding, and a second set of mice fasted for 18 h, refed and sacrificed after 7 h.

Tissues were harvested and fixed in 4% paraformaldehyde/phosphate-buffered saline for 18–20 h at room temperature followed by embedding in paraffin. Tissue blocks were obtained and 5 μℳ thick sections were cut and mounted on poly-ℒ-lysine coated glass slides, blocked with 20% normal donkey serum/phosphate-buffered saline and 0.1% Triton X-100 for 30 min after citrate antigen retrieval. The slides were incubated for 1 h with a 1:50 dilution of primary antibodies (Bethyl labs, Montgomery, TX, USA) and a 1:200 dilution of fluorescein isothiocynate-conjugated anti-rabbit or goat (Jackson Laboratories, Bar Harbor, ME, USA) used as secondary antibodies for 1 h, and DAPI for blue staining of nuclei. Negative controls were performed on similar slides using secondary antibodies alone without incubation of primary antibodies.

All colocalization studies were performed on the same sections with specific antibodies raised in different species. Incubations were performed with anti-rabbit menin overnight followed by 1 h incubation with fluorescein isothiocynate-conjugated donkey anti rabbit-green and anti-goat GIP overnight followed by streptavidin-Texas Red-conjugated donkey anti-goat for 1 h. Control staining included (a) replacement of the first layer of antibody by non-immune serum and by the diluent alone, and (b) secondary antibodies tested in relation to the specificity of the species in which the primary antibodies were raised, with the secondary antibody in question being replaced by secondary antibodies from different animal species.

Sections were examined with an Olympus IX70 inverted fluorescence microscope (Olympus; Tokyo, Japan) equipped with filters (Olympus) giving excitation at wavelengths of 475–555 nm for Texas Red and 453–488 nm for fluorescein isothiocynate, with a digital camera. Merged images were viewed by superimposing both photographs at 10 × and 40 × magnification.

Data were analyzed with SPSS software (Armonk, NY, USA) using one-factor analysis of variance analysis or Student's t-test as appropriate. P<0.05 were considered statistically significant.

Previous studies have shown that menin is expressed in the duodenum²² and that K cells express and secrete GIP.¹¹ To examine whether menin in the gut is expressed in GIP-expressing K cells, we performed double stain fluorescent co-immunohistochemistry on the proximal duodenum, 1 cm from the pylorus. We determined that menin and GIP colocalize in cells on the villi as shown in Figure 1.

We have previously shown that 4 and 7 h of refeeding are the times that induce the highest spikes of serum insulin with metabolic events implicated during these times.³² We analyzed menin and GIP levels in the proximal duodenum of mice after an 18 h fast followed by 4 and 7 h of refeeding using western blot analysis and a GIP specific ELISA kit, respectively. We found that menin expression was inversely correlated with GIP levels (Figure 2) such that significantly higher menin levels are observed at fasting compared with refeeding, as opposed to the level of GIP expression at fasting and refeeding in the same tissue samples. Specifically, after 4 and 7 h of refeeding, menin expression declined significantly as the duration of feeding time increased, with the lowest menin levels observed at 7 h of refeeding (Figures 2a and b). However, GIP expression was lowest at fasting and highest at 7 h of refeeding (Figure 2c). Although our data confirms a previous report that menin expression is highest in the gastrointestinal tract during fasting¹⁷ and at feeding, serum GIP is increased,³³ the effect of refeeding on menin levels and the inverse correlation with GIP expression is entirely novel.

We investigated if the effect of feeding on menin and GIP expression is at the level of mRNA by analyzing the proximal duodenum after an 18 h fast followed by 4 and 7 h of refeeding by reverse transcripion-PCR. In Figure 3a, we show that menin mRNA levels steadily decline the with length of refeeding after the 18 h fast, consistent with the levels of protein shown in Figures 2a and b. This steady decline in menin mRNA was associated with an increase in GIP mRNA expression peaking at 7 h of refeeding (Figure 3b), and consistent with the ELISA data in Figure 2c.

To determine the effect of diet on duodenal menin expression in relation to GIP, we subjected C57BL/6 mice to 3 months of HF diet and analyzed menin and GIP expression in the proximal duodenum by western blot analysis and reverse transcripion-PCR. In Figure 4a, we show that HF diet downregulates menin protein levels in the proximal gut of these mice. Furthermore, we observed a significant decrease in menin mRNA expression, (Figure 4b) consistent with reduction in protein levels as shown in Figure 4a. The reduction of menin expression inversely correlated with the upregulation of GIP caused by the HF diet (Figure 4c) relative to expression from mice on RD and consistent with reports that GIP is regulated by dietary fat and glucose.^{34, 35, 36}

To study the mechanism underlying the inverse correlation observed between the expression of menin and GIP, we tested the hypothesis that menin is a negative regulator of GIP in vitro. We overexpressed menin in STC-1 cells by transient transfection using Lipofectamine 2000 (Figure 5a), and determined the effect on endogenous GIP levels. We show in Figure 5b that overexpression of menin in STC-1 cells correlates with significant downregulation of GIP mRNA.

Previously, when a 2.9 kB human GIP promoter luciferase reporter gene construct was transfected into STC-1 cells, there was a 40-fold elevated GIP promoter activity over the response in IEC cells (a cell line derived from the ileum).²⁸ However, the greatest promoter activity was detected with the highly conserved 0.210 kB promoter construct, suggesting that more distal regions may contain repressor elements.²⁸ We therefore compared the effect of menin on both the 0.210 and 2.9 kB GIP promoters to determine if menin is part of a repressor component that regulates GIP expression in STC-1 cells. Using whole cell lysates derived from STC-1 cells overexpressing menin (Figure 5a), we determined the effect of menin on the GIP promoter by luciferase reporter assay. In Figure 6a, we show that the activity of the short 0.210 kB promoter for GIP was not affected by menin overexpression; however, overexpression of menin significantly inhibited the 2.9 kB GIP promoter (Figure 6b), supporting our hypothesis that menin may be part of a repressor component that negatively regulates GIP. These results suggest that menin may negatively regulate the transcription of GIP in cells and explain the in vivo inverse correlation observed with previous results shown.

In line with our hypothesis and consistent with menin's possible role as a GIP repressor, we found that menin siRNA (si^menin) significantly increased GIP expression as determined by ELISA performed on STC-1 whole cell lysates (Figure 6c). The PI3K-AKT inhibitor LY294002 (LY) caused downregulation of GIP expression; however, the effect was inhibited with menin siRNA (LYsi^menin) (Figure 6c). Furthermore, the MAP kinase inhibitor UO126 (UO) inhibited GIP expression independent of menin, since menin siRNA did not reverse the effect of UO as observed with LY294002. Expression of secreted GIP in the media followed a similar trend as observed in the lysates (Figure 6d). In Figure 6e, we demonstrate that the addition of menin-specific siRNA to STC-1 cells indeed leads to a significant decrease in menin expression and that the changes observed with si^menin and LYsi^menin can be attributed to the significant downregulation of menin expression in the transiently transfected STC-1 cells. With this data, we conclude that the PI3K-AKT-mediated regulation of GIP is dependent on menin whereas MAP kinase-mediated GIP expression is via a menin-independent pathway.

GIP is known to suppress p38 MAPK activation via AKT to induce pro-survival responses and inhibit apoptosis in pancreatic beta cells.³⁷ Menin induces apoptosis and when down-regulated has also been associated with proliferation of beta cells.²² In contrast, GIP directly stimulates proliferation by binding to its receptor on target beta cells.³⁸ Since it has been reported that GIP exerts its effects on cell survival via an AKT-dependent pathway,³⁷ we postulated that GIP inhibits menin expression via the PI3K-AKT signaling cascade. In Figure 7a, we show that STC-1 cells exposed to GIP resulted in significant upregulation of pAKT associated with the downregulation of menin protein. This activation was abrogated with LY294002. Since GIP is known to inhibit MAPK signaling via downregulation of phospho-p38, we demonstrated that 20 nℳ GIP not only downregulates menin expression but also reduces phospho-p38 levels in the STC-1 cells (Figure 7b). Furthermore, we confirm that the ability of PI3K/AKT inhibitor LY294002 (LYg²⁰) to abrogate GIP (g) downregulation of menin protein expression is via inhibition of pAKT (Figure 7a), as phospho-p38 levels did not change with LY treatment as shown in Figure 7b. Indeed the AKT-mediated GIP suppression of p38 is confirmed with the activation of pAKT (Figure 7a) and the significant reduction in phospho-p38 (Figure 7b). This was followed by LY294002's ability to abrogate the inactivation of p38 by GIP, as demonstrated by the high expression of phospho-p38 in the LYg²⁰ lane compared to g²⁰ in Figure 7b.

To further elucidate if phospho-p38 is a component of the GIP-mediated regulation of menin, we used the MAP kinase inhibitor UO126 (UO) to effectively downregulate the expression of phospho-p38 with or without GIP and analyzed menin levels. As previously shown, GIP effectively reduced menin expression at a concentration of 20 nℳ (g²⁰) and phospho-p38 levels (Figure 7c). Interestingly, pretreatment of cells with UO126 prior to 20 nℳ GIP abrogated GIP's ability to downregulate menin protein, even though phospho-p38 was still significantly inhibited by UO126 (Figure 7c) as previously described.³⁷ This suggests that UO126-inhibition of GIP regulation on menin is independent of phospho-p38. Collectively, Figure 6 demonstrates that GIP inhibition of menin expression is in part via a PI3K-AKT-dependent signaling pathway independent of phospho-p38.