Activation of AMPK Stimulates Neurotensin Secretion in Neuroendocrine Cells

Abstract

AMP-activated protein kinase (AMPK), a critical fuel-sensing enzyme, regulates the metabolic effects of various hormones. Neurotensin (NT) is a 13-amino acid peptide predominantly localized in enteroendocrine cells of the small bowel and released by fat ingestion. Increased fasting plasma levels of pro-NT (a stable NT precursor fragment produced in equimolar amounts relative to NT) are associated with an increased risk of diabetes, cardiovascular disease, and mortality; however, the mechanisms regulating NT release are not fully defined. We previously reported that inhibition of the mammalian target of rapamycin (mTOR) complex 1 (mTORC1) increases NT secretion and gene expression through activation of the MEK/ERK pathway. Here, we show that activation of AMPK increases NT secretion from endocrine cell lines (BON and QGP-1) and isolated mouse crypt cells enriched for NT-positive cells. In addition, plasma levels of NT increase in mice treated with 5-aminoimidazole-4-carboxamide riboside, a pharmacologic AMPK activator. Small interfering RNA-mediated knockdown of AMPKα decrease, whereas overexpression of the subunit significantly enhances, NT secretion from BON cells treated with AMPK activators or oleic acid. Similarly, small interfering RNA knockdown of the upstream AMPK kinases, liver kinase B1 and Ca²⁺ calmodulin-dependent protein kinase kinase 2, also attenuate NT release and AMPK phosphorylation. Moreover, AMPK activation increases NT secretion through inhibition of mTORC1 signaling. Together, our findings show that AMPK activation enhances NT release through inhibition of mTORC1 signaling, thus demonstrating an important cross talk regulation for NT secretion.

Neurotensin (NT), a 13-amino acid peptide predominantly localized in specialized enteroendocrine (EE) cells of the small bowel (1) and released by fat ingestion (2, 3), facilitates fatty acid (FA) translocation in rat intestine (4) and stimulates growth of various cancers (5). Melander et al (6, 7) recently reported that elevated fasting plasma levels of pro-NT (a stable NT precursor fragment produced in equimolar amounts relative to NT) are associated with an increased risk of diabetes, cardiovascular disease, mortality, and an increased risk of developing breast cancer. Therefore, NT plays a critical role for normal intestinal physiologic function and also contributes to metabolic disorders and the growth of certain cancers. However, the mechanisms contributing to NT secretion have not been entirely delineated.

AMP-activated protein kinase (AMPK), a serine/threonine kinase comprised of 3 subunits: α (catalytic), β, and γ (regulatory) (6–8), is a critical fuel-sensing enzyme and regulator of metabolism. Upon energy stress, AMP directly binds to the γ-subunit resulting in 3 effects: 1) activation of AMPK allosterically; 2) induction of phosphorylation of a threonine residue (Thr¹⁷²) within the activation domain of the α-subunit by an upstream kinase, the tumor suppressor liver kinase B1 (LKB1); and, 3) inhibition of the dephosphorylation of Thr¹⁷² by protein phosphatase (9). Ca²⁺ calmodulin-dependent protein kinase kinase 2 (CaMKK2) was identified as an additional upstream kinase of AMPK (10–12). By responding to diverse hormonal signals including leptin and adiponectin, AMPK serves as a signal integrator among peripheral tissues and the hypothalamus in the control of food intake, body weight and glucose and lipid homeostasis (8, 9). In addition, AMPK plays a negative role in glucose-stimulated insulin secretion (GSIS) in pancreatic β-cells to maintain glucose homeostasis (13, 14). Activation of AMPK by 5-aminoimidazole-4-carboxamide riboside (AICAR) and metformin, both pharmacologic AMPK activators (15), markedly reduces GSIS from human primary pancreatic islets (16, 17) and β-cell lines (16, 18). Furthermore, overexpression of a constitutively active form of AMPK results in repressed glucose-induced insulin release from β-cell lines (18, 19), whereas overexpression of a dominant-negative form of AMPK leads to increased insulin release (16).

Mammalian target of rapamycin (mTOR) is also a serine/threonine kinase and a central regulator of cell growth (20). mTOR is part of 2 distinct multiprotein complexes (ie, mTOR complex 1 [mTORC1] and mTORC2) (20). The mTORC1 complex is composed of 4 components: raptor (regulatory associated protein of mTOR), the proline-rich Akt substrate of 40 kDa, mTOR associated protein, LST8 homolog, and mTOR. Raptor acts as a scaffold to recruit downstream substrates such as eukaryotic translation initiation factor 4E binding protein and ribosomal S6 kinase (S6K1), to mTORC1 (20). Tuberous sclerosis complex (TSC)1 and TSC2 contain a GTPase-activating protein domain at its carboxyl terminus that inactivates the small Ras-like GTPase Rheb, which has been shown to associate with and directly activate mTORC1 (20, 21). Therefore, TSC is a negative regulator of mTORC1, with loss of TSC1 or 2 leading to hyperactivation of mTORC1 (20). AMPK directly phosphorylates and activates TSC2, leading to suppression of mTORC1 signaling (22). Phosphorylation of raptor by AMPK at 2 highly conserved serines, 722 and 792, induces their direct binding to 14–3-3, which leads to a suppression of mTORC1 kinase activity (23, 24).

Previously, we reported that inhibition of mTORC1 signaling enhances NT secretion and gene expression in the endocrine cell line BON (25). Moreover, we demonstrated that mTORC1 inhibition induces a feedback activation of ERK1/2, which positively regulates NT gene expression and secretion (25, 26). In our present study, we demonstrate that AMPK activation increases NT secretion in endocrine cell lines, isolated primary intestinal crypts, and in vivo. More importantly, we demonstrate that AMPK activation regulates NT release through inhibition of mTORC1 signaling. These findings integrate mTOR and AMPK signaling and further delineate the intracellular mechanisms and feedback pathways regulating NT secretion.

Materials and Methods

Reagents

All the antibodies used in this study, except for the antibodies mentioned below, were from Cell Signaling Technology. Phosphorylation of ERK1/2 (p-ERK1/2) and NT antibodies for immunofluorescence (IF) were from Santa Cruz Biotechnology, Inc. β-Actin antibody was from Sigma-Aldrich. ON-TARGETplus SMARTpool (AMPKα1, AMPKα2, LKB1, CaMKK2, and TSC2) and ON-TARGETplus Nontargeting Control Pool small interfering RNA (siRNA) were purchased from GE Dharmacon. AICAR, 2-deoxyglucose (2-DG), phenformin, Compound C (CC), and PD98059 were from Cayman. FA-free BSA, oleate sodium, and glucose were from Sigma-Aldrich. Wild-type AMPKα1 and AMPKα2 plasmids were provided by Heng-Ye Man (Boston University, Boston, MA) (27). myc-raptor S722A/S792A was a gift from Reuben Shaw (Addgene plasmid 18118) (23).

Cell culture and transfection

The BON cell line was derived from a human pancreatic carcinoid tumor and previously characterized (28, 29). BON cells are maintained in a 1:1 mixture of DMEM and Ham's F-12 Nutrient Mixture, supplemented with 5% fetal bovine serum in 5% CO₂ at 37°C. QGP-1, a pancreatic endocrine cell line purchased from Japan Health Sciences Foundation (30), was maintained in ATCC-formulated RPMI 1640 medium with 10% fetal bovine serum. siRNA and plasmid transfections were performed using RNAiMAX and Lipofectamine LTX Reagent with PLUS Reagent (Life Technologies), respectively. Forty-eight hours after transfection, cells were treated as described below.

Cell treatment and NT enzyme immunoassay (EIA)

For AICAR, 2-DG, and phenformin treatments, both BON and QGP-1 cells were plated in 24-well plates at a density of 1 × 10⁵/cm² and 1.5 × 10⁵/cm², respectively, and grown for 48 hours; cells were treated with these reagents for 3 hours in growth medium. For CC and PD98059 treatments, cells were pretreated with the inhibitors for 30 minutes followed by AMPK activators for another 3 hours. For oleate treatment, cells were treated with either 0.1% FA-free BSA (Sigma-Aldrich) or BSA-conjugated oleate sodium in serum-free medium for 1 hour. Media were collected and stored in −80°C for NT measurement using the NT EIA kit from Phoenix Pharmaceuticals as described previously (31, 32) and cells lysed for Western blotting. Data obtained from NT EIA were normalized by protein concentration from parallel cell lysates.

Crypt cell isolation and culture

All procedures were approved by the Institutional Animal Care and Use Committee at the University of Kentucky. Intestinal crypts were isolated from the ileum of mice (8- to 10-wk male C57BL/6 mice) as described (33) with slight modifications. Briefly, ileums were first digested with enzyme mixture containing type 1 collagenase, hyaluronidase, and deoxyribonuclease I for 15 minutes at 37°C in 5% CO₂ and then further digested by enzyme mixture consisting of type IV collagenase and dispase neutral protease for 2 15-minute intervals. The tissues were mechanically disrupted and allowed to sediment in DMEM containing 2% sorbitol. The crypts were collected by centrifuging the supernatant at 250g at room temperature, plated on collagen 1-coated plates (BD Biosciences), and cultured in the serum-free bronchial epithelial growth medium supplemented with growth factors, cytokines, and supplements included in the SingleQuots kit (Lonza). Forty-eight hours after isolation and culture, cells were either used for NT IF staining or treatment for NT EIA.

In vivo studies

All procedures were carried out in an animal facility according to protocols approved by the Institutional Animal Care and Use Committee at the University of Kentucky. For AICAR administration, male C57BL/6 mice (5 mo old) were injected ip with AICAR (500-mg/kg body weight) or saline. Blood glucose was monitored by OneTouch UltraMini Blood Glucose Meter (LifeScan) from a drop of blood from a tail snip before and 1 hour after the injection. One hour after the injection, mice were then anesthetized with isoflurane inhalation; blood was collected from the inferior vena cava using a heparin-coated syringe, and the plasma obtained by centrifuging the blood at 10 000 rpm for 10 minutes at 4°C. Plasma aliquots were stored in −80°C. Plasma (50 μL) was used for NT measurement by NT EIA. For plasma NT measurement in mice fed olive oil, male C57BL/6 mice (5 mo old) were fasted overnight and given saline or olive oil (10-μL/g body weight) by gavage. One hour later, mice were anesthetized, blood obtained, and plasma NT levels measured as described above.

IF staining and confocal microscopy

IF staining was performed as described previously (11). Briefly, cells were grown on glass coverslips (number 1) in 24-well plates for 72 hours. Cells were fixed with 4% paraformaldehyde/PBS and permeabilized with 0.3% Triton X-100/PBS. Cells were incubated with primary antibody for 1 hour followed by Alexa Fluor-conjugated secondary antibody from Invitrogen for 30 minutes and nuclei counterstained with 4′,6-diamidino-2-phenylindole. Images were observed under an FV1000 Olympus confocal microscope with a ×60, 1.35 numerical aperture oil objective. Images were analyzed with Olympus FV10-ASW2.1 software.

Protein preparation and Western blotting

Protein preparation and Western blotting were performed as described previously (31, 34). In brief, the cells were lysed with lysis buffer (Cell Signaling Technology) and equal amounts of protein were resolved on 4%–12% NuPAGE BisTris gels from Invitrogen and electrophoretically transferred to polyvinylidene difluoride membranes; the membranes were incubated with primary antibodies overnight at 4°C followed by secondary antibodies conjugated with horseradish peroxidase. Membranes were developed using an enhanced chemiluminescence Western Blotting System from GE Healthcare Life Science and Thermo Scientific. p-AMPK expression was analyzed by densitometry and normalized to total AMPK expression using NIH ImageJ software.

Statistical analysis

Descriptive statistics including mean and standard deviation were calculated to summarize NT secretion. Bar graphs were generated to represent mean (±SD) NT levels in different cell culture conditions and mice groups such as treatment with AMPK activators and inhibitors, siRNA groups, and dose concentrations. Within each experiment, comparisons across groups were accomplished using one-way ANOVA models, and pairwise comparisons were subsequently performed using contrast statements. Adjustment in P values due to larger numbers of pairwise testing within each experiment (see Figures 3–5 below) was performed using the Holm's procedure. Trend tests for dose comparisons were likewise performed. Normality assumptions of the parametric tests for each outcome were assessed. P < .05 was considered statistically significant.

Figure 1.

Activation of AMPK increased NT secretion. A–C, BON cells were plated in 24-well plates and grown for 48 hours. Cells were treated with or without AICAR (A), 2-DG (B), or phenformin (phen) (C) at various concentrations for 3 hours; media were collected for NT measurements by EIA (upper panels); cells were collected and AMPK/ACC signaling and β-actin were examined by Western blotting from protein extracts (lower panels). *, P < .05 vs control; n = 6. D, BON cells were plated as in A and pretreated with or without CC (10μM) for 30 minutes followed by AICAR (1mM) for 3 hours; media (upper panel) and cells (lower panel) were analyzed as in A. *, P < .05 vs control; †, P < .05 vs AICAR alone; n = 6. E, BON cells were plated as in A and incubated in DMEM with high glucose (4.5 g/L) or without glucose in the presence or absence of CC (10μM) or AICAR (1mM) for 3 hours; media (upper panel) and cells (lower panel) were analyzed as in A. *, P < .05 vs glucose alone; †, P < .05 vs glucose deprivation; n = 6. F–H, QGP-1 cells were plated in 24-well plate and treated with or without AICAR (F), 2-DG (G), or phen (H) and analyzed as in A. *, P < .05 vs control; n = 6. All experiments were performed at least twice. All data are mean ± SD.

Figure 2.

AMPK activation increases NT secretion from isolated crypts and mice. A, Crypt cells isolated from the ileums of mice are shown at 0 and 48 hours (left and middle panels). Crypt cells at 48 hours were stained with anti-NT antibody (right panel). B, Isolated crypt cells were treated with 2-DG (10mM) or phorbol 12-myristate 13-acetate (10nM) for 3 hours. Media were collected and NT EIA performed. *, P < .05 vs dimethyl sulfoxide (control); n = 6. C, Male C57BL/6 mice (5 mo old) were injected ip with AICAR (500-mg/kg body weight) or saline for 1 hour. Blood glucose was monitored by OneTouch UltraMini Blood Glucose Meter from a drop of blood from tail snip before and 1 hour after the injection (left panel); 1 hour after the injection, blood was collected from the inferior vena cava under anesthesia and plasma (50 μL) was used to measure NT level by NT EIA kit (right panel). *, P < .05 vs saline; n = 5.

Figure 3.

Both AMPKα1 and AMPKα2, LKB1, and CaMKK2 are involved in NT secretion. A, BON cells were plated in 24-well plates and transfected with siRNA (20nM) targeting AMPKα1 or AMPKα2 or with NTC siRNA. Forty-eight hours after transfection, cells were treated with or without AICAR (1mM) for 3 hours; media were collected for NT EIA (upper panel) and cells for Western blotting (lower panel). *, P < .05 vs each control; †, P < .05 vs AICAR in NTC siRNA; n = 6. B, BON cells were plated and transfected with plasmids of wild-type AMPKα1 or AMPKα2 or the control vector. Cells were treated and analyzed as in A. *, P < .05 vs each control; †, P < .05 vs AICAR in control vector; n = 6. C, BON cells were plated, transfected with NTC, LKB1, or CaMKK2 siRNA (40nM), and grown as in A. Cells were treated with 2-DG (10mM) for 3 hours. Media were collected for NT EIA (upper panel) and cells lysed for Western blotting (lower panel). *, P < .05 vs each control; †, P < .05 vs control in NTC siRNA; ‡, P < .05 vs 2-DG in NTC siRNA; n = 6. All experiments were performed at least twice. All data are mean ± SD.

Figure 4.

Cross talk of AMPK with mTORC1 signaling. A, BON cells were plated in 6-well plates and grown for 48 hours. Cells were treated with various doses of AICAR (left panel) or 1mM AICAR in the presence or absence of CC (10μM) (right panel) for 3 hours and lysed for Western blotting. B, BON cells were plated, grown, treated, and analyzed as in A (left panel). C, BON cells were plated on coverslips in 24-well plate and grown for 72 hours. Cells were treated as in A (right panel), and IF staining using p-ERK1/2 antibody and confocal microscopic analysis were performed. D and E, BON cells were plated as in A. Cells were transfected with control vector and myc-raptor S722A/S792A (D) or NTC and TSC2 siRNA (E) and grown for 48 hours. Cells were treated with AICAR (1mM) for 3 hours; media and cells were collected for NT EIA and Western blotting, respectively. *, P < .05 vs control in vector and †, P < .05 vs AICAR in vector (D); *, P < .05 vs control in NTC siRNA; †, P < .05 vs control in NTC siRNA; ‡, P < .05 vs AICAR in NTC siRNA (E); n = 6. F, BON cells were treated with or without AICAR (1mM) or phenformin (1mM) in the presence or absence of PD98059 (10μM) for 3 hours. Media were collected and NT EIA performed. *, P < .05 vs control; †, P < .05 vs control; ‡, P < .05 vs AICAR alone; #, P < .05 vs phen alone; n = 6. All experiments were performed at least twice. All data are mean ± SD.

Figure 5.

AMPK activation increases FA-stimulated NT secretion. A, Male C57BL/6 mice (5 mo old) under normal chow (NC) were given saline (control) or olive oil (10-μL/g body weight) by gavage. One hour after the gavage, blood was collected, and plasma NT was measured by NT EIA. *, P < .05 vs NC; n = 5. B and C, BON (B) and QGP-1 (C) cells were plated in 24-well plates and treated with 0.1% BSA or various concentrations of BSA-conjugated oleate for 1 hour. Media and cells were collected and analyzed by NT EIA and Western blotting, respectively. *, P < .05 vs control (BSA); n = 6. D. BON cells were plated in 24-well plate and pretreated with or without AICAR (1mM) for 2 hours followed by combined with or without oleate (0.5mM) for 1 hour. Media and cells were analyzed as in B. *, P < .05 vs control (BSA); †, P < .05 vs control (BSA); ‡, P < .05 vs oleate alone; n = 6. E and F, BON cells were plated in 24-well plates and transfected with siRNA directed against AMPKα1 or AMPKα2 or NTC siRNA (E) or LKB1 or CaMKK2 (F) siRNA and grown for 48 hours. Cells were treated with or without oleate (0.5mM) for 1 hour; media and cells were collected and analyzed as in B. *, P < .05 vs each control (BSA); †, P < .05 vs oleate in NTC siRNA; n = 6. G, BON cells were plated in 24-well plate and pretreated with or without EGTA (0.5mM) for 20 minutes followed with BSA or oleate (0.5mM) for 1 hour. Cells were collected and Western blotting of the cell extracts performed. Densitometric analysis (upper panel) of p-AMPK is normalized to total AMPK from 3 separate experiments and Western blot analyses (lower panel); graph demonstrates the fold change of p-AMPK vs BSA. *, P < .05 vs BSA; †, P < .05 vs oleate alone; n = 6. H, BON cells were plated on coverslips in 24-well plate. Seventy-two hours after plating, cells were treated with BSA or oleate (0.5mM) for 1 hour and processed for IF. Double staining was performed using antibodies of CaMKK2 and NT. Nuclei were stained by DAPI. Experiments in vitro were performed at least twice. All data are mean ± SD.

Results

AMPK activation increases NT secretion

To first determine whether AMPK regulates NT secretion, we treated BON cells with different doses of AICAR, a pharmacologic AMPK activator. As shown in Figure 1A, AICAR treatment increased NT secretion in a dose-dependent fashion and, concurrently, p-AMPK and p-acetyl-coenzyme A carboxylase (ACC), a direct downstream target of AMPK (8). Similarly, the AMPK activators, 2-DG and phenformin (an analog of metformin), increased NT secretion and AMPK and ACC activation (Figure 1, B and C). CC (6-[4-(2-piperidin-1-ylethoxy) phenyl]-3-pyridin-4-ylpyrazolo [1,5-a]pyrimidin), a cell-permeable AMPK inhibitor (35), inhibited AICAR-stimulated NT secretion, p-AMPK, and p-ACC (Figure 1D). Consistent with findings using pharmacologic agents, deprivation of glucose (5.4 g/L) from BON cells increased NT secretion and AMPK/ACC activation, which were attenuated by treatment with CC (Figure 1E). To further confirm these findings, we repeated experiments using another endocrine cell line, QGP-1, which expresses high levels of NT and secretes NT peptide in response to stimulation (36). Treatment of QGP-1 cells with AICAR, 2-DG, and phenformin consistently increased NT secretion and AMPK/ACC activation (Figure 1, F–H). Therefore, we demonstrate that AMPK activation stimulates NT release using 2 different endocrine cell line models.

AMPK activation increases NT release from isolated intestinal crypts and from mice in vivo

To confirm our findings using the tumor-derived cell lines, we used a primary intestinal crypt cell model enriched in EE cells as described previously (33). Intestinal crypts isolated from the ileums of mice were cultured in serum-free bronchial epithelial growth medium, in which EE cells are enriched and fibroblast growth is inhibited. IF staining using anti-NT antibody showed multiple NT-positive cells in the isolated crypts cultured for 48 hours (Figure 2A). Treatment of the crypt cells with 2-DG or phorbol 12-myristate 13-acetate (a secretory agonist used as a positive control) increased NT secretion (Figure 2B). Finally, we evaluated the effect of AMPK activation on NT release in vivo. As shown in Figure 2C, left panel, blood glucose levels were similar between the 2 groups before AICAR injection; however, the levels were significantly decreased in the AICAR-treated group compared with control mice at 1 hour after injection. This result is consistent with the notion that AMPK activation controls glucose homeostasis (37). Importantly, AICAR administration enhanced plasma NT levels (Figure 2C, right panel), consistent with our in vitro findings. Taken together, these results demonstrate that activation of AMPK positively regulates NT release from EE cells both in vitro and in vivo and further indicates a physiologic role for AMPK signaling on NT secretion.

Both AMPKα1 and AMPKα2 are involved in AMPK-mediated NT secretion

BON cells express equal amounts of AMPKα1 and AMPKα2 catalytic subunits (data not shown). To determine whether these subunits are involved in AMPK-regulated NT secretion, BON cells were transfected with siRNA targeting either AMPKα1 or AMPKα2 or with nontargeting control (NTC) siRNA. Either knockdown of AMPKα1 or AMPKα2 decreased AICAR-stimulated NT secretion (Figure 3A). In contrast, overexpression of wild-type AMPKα1 or AMPKα2 further increased AICAR-stimulated NT release (Figure 3B). Thus, AMPKα1 and AMPKα2 are both involved in AMPK-mediated NT secretion.

LKB1 and CaMKK2 are the most common upstream kinases of AMPK (8). Compared with the ubiquitous expression of LKB1 in various tissues, the expression of CaMKK2 is predominantly localized to the brain, liver, and skeletal muscle (10–12). Whether CaMKK2 is expressed in EE cells is unknown. We demonstrate that both LKB1 and CaMKK2 proteins are abundantly present in BON cells and that CaMKK2 appears as a doublet (Figure 3C, lower panel), consistent with findings in HeLa cells (12) and mouse hypothalamus (38). 2-DG-stimulated NT secretion was attenuated in BON cells transfected with either LKB1 or CaMKK2 siRNA (Figure 3C, upper panel). Furthermore, knockdown of either LKB1 or CaMKK2 decreased 2-DG-stimulated AMPK activation (Figure 3C, lower panel), indicating that both are involved in NT secretion stimulated by AMPK activators.

Cross talk of AMPK and mTOR signaling

Cross talk of AMPK and mTOR signaling has been implicated in multiple biological functions (23, 24). AMPK activation inhibits mTORC1 through raptor (23, 24) or TSC2 (22), and we have previously demonstrated that inhibition of mTORC1 increased NT secretion through feedback activation of ERK1/2 (25). Consistently, AICAR treatment decreased S6K1 and concurrently increased ERK1/2 phosphorylation in a dose-dependent fashion in BON cells (Figure 4A, left panel). In addition, CC treatment attenuated this effect (Figure 4A, right panel). mTORC1 negatively regulates phosphatidylinositol 3-kinase (PI3K)/V-akt murine thymoma viral oncogene homolog 1 (Akt1) signaling through a feedback loop. Inhibition of mTOR/S6K1 pathway decreases phosphorylation of insulin receptor substrate 1 at S307 and S636/S639 (39), which inhibits PI3K/Akt signaling (40). Consistently, a decrease of insulin receptor substrate 1 phosphorylation at S636/S639 and a subsequent increase of Akt phosphorylation were noted in BON cells treated with AICAR (Figure 4B), demonstrating that AMPK activation resulted in mTOR/S6K1 inactivation, thus releasing the negative feedback loop on PI3K/Akt signaling pathway. Furthermore, as shown by IF staining and confocal analysis, p-ERK1/2 was localized in the cytoplasm of control cells; AICAR treatment resulted in p-ERK1/2 nuclear localization, which was prevented by CC treatment (Figure 4C).

To examine the physiologic role of raptor, we used the AA mutant in which both Ser⁷²² and Ser⁷⁹² (AMPK phosphorylation sites) are replaced by alanine; mutation of these sites prevents AMPK agonists from fully suppressing mTORC1 in cells with normal mTOR signaling (23). Overexpression of AA raptor prevented AICAR-stimulated NT secretion in BON cells (Figure 4D). siRNA-mediated knockdown of TSC2 decreased both basal and AICAR-stimulated NT release (Figure 4E). Furthermore, as shown in Figure 4F, both AICAR- and phenformin-stimulated NT secretion was blocked by PD98059, an inhibitor of MEK (ERK1/2 kinase). Together, these findings demonstrate that AMPK stimulates NT secretion through inhibition of mTORC1, which results in feedback activation of ERK1/2 signaling.

AMPK is involved in long-chain FA-stimulated NT secretion

NT release is increased in response to FA administration in rats and humans (2, 3). To confirm this effect, we measured plasma NT levels in mice fasted overnight and then fed either saline or olive oil (10-μL/g body weight) by gavage. Blood was collected from mice 1 hour after gavage and plasma purified. As shown in Figure 5A, higher plasma NT levels were detected in mice fed olive oil compared with the control mice.

With confirmation that FA administration stimulates NT release in vivo, we next determined whether AMPK is involved in FA-mediated NT secretion using our established cell models. Treatment of oleate, a long-chain FA, increased NT secretion and concomitant activation of AMPK and ACC in a dose-dependent fashion in both BON (Figure 5B) and QGP-1 (Figure 5C) cells. NT secretion was increased in BON cells treated with oleate or AICAR alone and further increased by the combination treatment (Figure 5D), indicating the involvement of AMPK. siRNA-mediated knockdown of AMPKα1 or AMPKα2 attenuated oleate-stimulated NT secretion (Figure 5E). Surprisingly, oleate-stimulated NT secretion was decreased only in cells transfected with CaMKK2 but not LKB1 siRNA, even though p-AMPK was decreased in cells transfected with both siRNAs (Figure 5F). Oleate increases intracellular Ca²⁺ (41, 42); phosphorylation and activation of AMPK occur in response to an increase in intracellular Ca²⁺, which is dependent on CaMKK2 (10). To determine whether Ca²⁺ is required in CaMKK2-mediated AMPK phosphorylation stimulated by oleate, we treated BON cells with or without oleate for 1 hour in the presence or absence of EGTA (0.5mM), a specific chelator for Ca²⁺. As shown in Figure 5G, AMPK phosphorylation was increased by oleate; this induction was attenuated by EGTA treatment, indicating the involvement of Ca²⁺ in oleate-mediated AMPK phosphorylation. Moreover, CaMKK2 was expressed in the process tips in BON cells after oleate treatment and colocalized with NT vesicles (Figure 5H), further demonstrating the regulation of CaMKK2 on FA-mediated NT release. Together, these results suggest that CaMKK2 is functionally involved in AMPK-regulated NT secretion stimulated by FA through Ca²⁺/CaMKK2 signaling upstream of AMPK.

Discussion

An increasing number of hormones including leptin, adiponectin, ghrelin and insulin have been implicated in the regulation (ie, stimulation or inhibition) of AMPK activity in vivo (43–46). For example, leptin and adiponectin increase AMPK activity in liver and muscle (45), whereas ghrelin stimulates heart AMPK activity but inhibits adipose tissue and liver AMPK activity (45) and insulin has an inhibitory effect on fat and myocardial AMPK activity (45). Conversely, AMPK also regulates the secretion and biologic functions of various hormones, such as insulin and GH (47). These findings establish AMPK as a key molecular regulator of hormonal signaling. Here, we demonstrate that: 1) AMPK activation, mediated by either pharmacologic AMPK activators or FA, increases NT secretion from BON and QGP-1 cells; 2) AMPK activation enhances NT release from primary intestinal crypts enriched for NT-positive cells; 3) administration of AICAR in mice elevates plasma NT; and, 4) AMPK activation stimulates NT secretion through inhibition of mTORC1 signaling.

Similar to our findings demonstrating a stimulatory effect of AMPK on NT release in vitro and in vivo, activation of AMPK by AICAR increased GH secretion in pituitary tumor cells (47). In human adipose tissue, AICAR has been shown to increase adiponectin expression (48, 49). The role of AMPK in the regulation of insulin secretion is still debated (13). Perfusion of the pancreas with glucose (6mM) together with AICAR enhances insulin secretion (50). AICAR treatment of β-cells and rat islets increases GSIS (17, 51). In contrast, AMPK activation negatively regulates GSIS from β-cells in MIN6 insulinoma cells (52). Collectively, these studies indicate that AMPK plays either a positive or negative role in regulating hormone secretion, which is dependent on cell type.

We demonstrate that CaMKK2, but not LKB1, mediates the effect of AMPK in FA-mediated NT secretion. Either knockdown of LKB1 or CaMKK2 decreased FA-stimulated AMPK phosphorylation; however, only CaMKK2 siRNA attenuated NT release mediated by oleate. CaMKK2-dependent activation of AMPK operates independently of AMP due to the absence of the AMP-binding subunit (53); in contrast, CaMKK2 has considerable constitutive activity and can be stimulated by an increase in Ca²⁺ (9, 54–56). Moreover, application of oleic acid increases the intracellular calcium concentration in ventricular myocytes and cultured human keratinocytes (41, 42). Consistent with these studies, we found that AMPK phosphorylation was decreased by EGTA, a Ca²⁺ chelator. LKB1 and CaMKK2 form individual signaling complexes. LKB1 is present in the nucleus, and its cytosolic translocation is an essential step for AMPK activation (57, 58). CaMKK2 also forms a unique physical complex with AMPKα and AMPKβ and activates AMPK (38, 53). We found that CaMKK2 colocalized with NT-positive vesicles in BON cells after stimulation with oleate. This unique expression pattern further suggests the specific involvement of CaMKK2 in FA-stimulated NT secretion. Based on our findings, we speculate that oleate treatment activates Ca^2+/CaMKK2 signaling and subsequent AMPK activation, leading to increased NT secretion from endocrine cells.

We demonstrate that AMPK activation regulates NT release through inhibition of mTORC1 signaling, which we have previously shown to play a negative role in NT gene expression and NT secretion (25). The cross talk of these 2 pathways is well established in regulating cellular metabolism, energy homeostasis and cell growth (59); however, the role of AMPK/mTORC1 cross talk on hormone secretion has not been studied in-depth. Gleason et al (51) reported that inhibition of AMPK in pancreatic β-cells by high glucose inversely correlates with mTOR activation, and in contrast, amino acids suppressed the activity of AMPK, indicating the cooperative regulation of both mTOR and AMPK in response to intracellular energy status. In our present study, we showed that AMPK activation decreased S6K1 phosphorylation and that overexpression of raptor S722A/S792A or transfection of TSC2 siRNA blocked AMPK-mediated NT release. Taken together, our findings further demonstrate the important effect of the interaction of AMPK and mTORC1 signaling pathways on hormone secretion.

In summary, we demonstrate that AMPK activation by a variety of mechanisms stimulates NT secretion (Figure 6). Furthermore, AMPK activation results in inhibition of mTORC1, suggesting a novel role of AMPK/mTORC1 cross talk on intestinal hormone secretion. These findings integrate the mTOR and AMPK signaling networks, the 2 most important pathways in glucose, lipid, and protein metabolism, and further define the molecular mechanisms regulating hormone secretion from intestinal endocrine cells.

Figure 6.

Summary diagram. AMPK activation stimulates NT secretion through inhibition of mTORC1 signaling.

Acknowledgments

We thank Donna A. Gilbreath, Catherine E. Anthony, Heather N. Russell-Simmons, and Jennifer F. Rogers for manuscript preparation and the Biostatistics and Bioinformatics Shared Resource Facility of the University of Kentucky Markey Cancer Center for statistical analyses.

This work was supported by National Institutes of Health Grants R37 AG10885 and R01 DK48489 and by National Cancer Institute Grant P30 CA177558.

Disclosure Summary: The authors have nothing to disclose.

Abbreviations

 
• ACC

  acetyl-CoA carboxylase

 
• AICAR

  5-aminoimidazole-4-carboxamide riboside

 
• AMPK

  AMP-activated protein kinase

 
• Akt1

  V-akt murine thymoma viral oncogene homolog 1

 
• CaMKK2

  Ca²⁺ calmodulin-dependent protein kinase kinase 2

 
• CC

  Compound C

 
• 2-DG

  2-deoxyglucose

 
• EE

  enteroendocrine

 
• EIA

  enzyme immunoassay

 
• FA

  fatty acid

 
• GSIS

  glucose-stimulated insulin secretion

 
• IF

  immunofluorescence

 
• LKB1

  liver kinase B1

 
• mTOR

  mammalian target of rapamycin

 
• mTORC1

  mTOR complex 1

 
• NT

  neurotensin

 
• NTC

  nontargeting control

 
• p-ERK1/2

  phosphorylation of ERK1/2

 
• PI3K

  phosphatidylinositol 3-kinase

 
• S6K1

  ribosomal S6 kinase

 
• siRNA

  small interfering RNA

 
• TSC

  tuberous sclerosis complex.