Lowering oxidative stress in ghrelin cells stimulates ghrelin secretion

Abstract

Ghrelin is a predominantly stomach-derived peptide hormone with many actions including regulation of food intake, body weight, and blood glucose. Plasma ghrelin levels are robustly regulated by feeding status, with its levels increasing upon caloric restriction and decreasing after food intake. At least some of this regulation might be due to direct responsiveness of ghrelin cells to changes in circulating nutrients, including glucose. Indeed, oral and parental glucose administration to humans and mice lower plasma ghrelin. Also, dissociated mouse gastric mucosal cell preparations, which contain ghrelin cells, decrease ghrelin secretion when cultured in high ambient glucose. Here, we used primary cultures of mouse gastric mucosal cells in combination with an array of pharmacological tools to examine the potential role of changed intracellular oxidative stress in glucose-restricted ghrelin secretion. The antioxidants resveratrol, SRT1720, and curcumin all markedly increased ghrelin secretion. Furthermore, three different selective activators of Nuclear factor erythroid-derived-2-like 2 (Nrf2), a master regulator of the antioxidative cellular response to oxidative stress, increased ghrelin secretion. These antioxidant compounds blocked the inhibitory effects of glucose on ghrelin secretion. Therefore, we conclude that lowering oxidative stress within ghrelin cells stimulates ghrelin secretion and blocks the direct effects of glucose on ghrelin cells to inhibit ghrelin secretion.

INTRODUCTION

The hormone ghrelin binds to its canonical receptor, the growth hormone (GH) secretagogue receptor (GHSR), within certain brain regions and peripheral organs to regulate food intake, body weight, adiposity, and blood glucose, among other biological functions (37, 39). Ghrelin is produced mainly by a distinct set of enteroendocrine ghrelin cells that are localized to the mucosal lining of the stomach and duodenum (9, 27, 39, 45). A portion of ghrelin undergoes a unique posttranslational acylation step, most commonly an octanoylation, that is required for ghrelin binding to GHSR (18, 61). Although ghrelin was originally discovered as a potent GH secretagogue, and also well established for its orexigenic actions, more recent studies have characterized ghrelin as a survival hormone that helps defend the body against challenging conditions such as starvation, hypoglycemia, cachexia, and chronic psychosocial stress (37, 48). For example, genetically engineered mouse models such as ghrelin-knockout mice that are unable to increase ghrelin secretion in response to starvation-like conditions exhibit marked hypoglycemia and increased mortality (15, 30, 58, 62, 63). Also, ghrelin-knockout mice exhibit more pronounced and prolonged hypoglycemia in response to an insulin bolus and require higher glucose infusion rates during hyperinsulinemic-hypoglycemic clamps to maintain the target plasma glucose range than wild-type mice (48).

Given the key glucoregulatory and survival actions of ghrelin in these settings, it is of utmost importance to gain a better understanding of the mechanisms mediating ghrelin secretion at the level of the ghrelin cell. To date, several factors controlling ghrelin secretion have been discovered. For instance, activation of neuronal sympathetic drive to stimulate ghrelin cell-expressed β₁-adrenergic receptors figures prominently in ghrelin secretion resulting from energy deficiency (11, 12, 20, 23, 28, 33, 34, 64). On the other hand, dietary macronutrients, including glucose, fatty acids, and amino acids, all have been reported to suppress ghrelin secretion (1, 24, 39). In particular, administration of glucose by either oral or parental routes rapidly and effectively reduces plasma ghrelin levels in both humans and rodent models (5, 10, 16, 36, 38–40, 42, 49, 52). This effect of glucose to reduce plasma ghrelin levels is due at least in part to direct effects of glucose on ghrelin cells, as evidenced in ex vivo studies using a mouse gastric mucosal cell primary culture model (36, 46). Culturing the gastric mucosal cells, of which ~1 out of every 300 is estimated to be a ghrelin cell, in medium containing high glucose (10 mM; simulating hyperglycemia) inhibits ghrelin secretion (36, 46). In contrast, culturing the gastric mucosal cells in medium containing 0 mM or low glucose (1 mM; simulating hypoglycemia) enhances ghrelin secretion compared with medium containing 5 mM glucose (simulating normoglycemia) (36, 46). Mouse gastric ghrelin cells also express mRNAs of several enzymes, ion channels, and transporters commonly involved in glucose transport, glucose metabolism, and glucose-mediated hormone secretion in perhaps the most well-known glucose-responsive cell, the pancreatic beta cell (46). Here, we took advantage of our previously established mouse gastric mucosal cell primary culture model (46) in combination with an array of pharmacological tools to demonstrate that lowering oxidative stress within ghrelin cells stimulates ghrelin secretion and blocks the direct effects of glucose on ghrelin cells to inhibit ghrelin secretion.

MATERIALS AND METHODS

Animals.

All animal experiments were approved by the University of Texas (UT) Southwestern Medical Center Institutional Animal Care and Use Committee. Mice were housed under a 12:12-h dark-light cycle in standard environmentally controlled conditions with free access to water and standard chow diet [2016 Teklad Global 16% protein diet; Envigo, Indianapolis, IN)].

Mouse gastric mucosal cell primary culture.

Gastric mucosal cell primary cultures were established from 8- to 12-wk-old male C57BL/6N mice, as reported previously (34, 46). The mice were anesthetized with chloral hydrate (700 mg/kg ip) followed by laparotomy to dissect out the stomachs. The stomachs were tied off at both ends with surgical sutures, excised and placed in ice-cold PBS, and then incised at the nonglandular portion to clear the digesta. The stomachs were inverted inside-out by passing the distal part of the stomach through the incision and inflated with and briefly placed in ice-cold DMEM medium without glucose (Life Technologies, Grand Island, NY). The residual digesta adhering to the stomach mucosa was cleared off gently by using soft paper pads. The inverted stomachs were incubated for 1.5 h at 37°C in 35 U dispase II/15 mL PBS (Roche Diagnostics, Indianapolis, IN) in a 50-mL centrifuge tube, and then the mucosa was scraped off with a transfer pipette into sterile DMEM-F-12 (1:1) medium (Mediatech Inc. Manassas, VA) containing 10% (vol/vol) FBS (Atlanta Biologicals, Lawrenceville, GA) supplemented with 100 U/mL penicillin and 100 μg/mL streptomycin sulfate. The cells were centrifuged at 310 g for 3 min, and the medium was removed and subjected to digestion with 0.25% trypsin EDTA (Mediatech Inc.). After 5 min, the trypsin was inactivated by adding FBS containing DMEM-F-12 medium. The mucosal cells in the medium were triturated by using a sterile glass pipette to disperse them, filtered through a 100-µm Falcon nylon mesh filter, and then centrifuged again at 310 g for 3 min to remove the supernatant medium. The cells were resuspended in FBS containing DMEM-F-12 medium supplemented with sodium octanoate-bovine serum albumin (BSA) to achieve a final concentration of 50 μM sodium octanoate and then plated at a density of 1 × 10⁵ cells·mL⁻¹·well⁻¹ in poly-d-lysine-precoated 24-well plates. The cells were placed overnight in a 37°C incubator with 5% CO₂. The next day, the cells were treated with the test substance in serum-free DMEM (Life Technologies, Grand Island, NY) containing 50 μM sodium octanoate-BSA and different concentrations of d-glucose, as indicated, for 6 h. At the end of the 6-h incubation period, the medium was collected, placed on ice, and immediately centrifuged at 4°C and 800 g for 5 min. Hydrochloric acid was added to the supernatant to achieve a final concentration of 0.1 N (for stabilization of acyl ghrelin) and stored at −80°C until analysis for acyl ghrelin.

Some of the test compounds used for the ghrelin secretion studies [d-glucose, 1-isothiocyanato-4-(methylsulfinyl)-butane (dl-sulforaphane), 6-chloro-2,3,4,9-tetrahydro-1H-carbazole-1-carboxamide (EX-527), 3,5,4′-trihydroxy-trans-stilbene (resveratrol), tunicamycin, and (E,E)-1,7-bis(4-hydroxy-3-methoxyphenyl)-1,6-heptadiene-3,5-dione (curcumin)] were purchased from Millipore Sigma (St. Louis, MO). d-mannoheptulose was from MP Biomedicals (Solon, OH). N-(2-(3-(1-piperazinylmethyl)imidazo[2,1-b]thiazol-6-yl)phenyl)-2-quinolinecarboxamide (SRT1720) was from EMD-Calbiochem (San Diego, CA). 2-Cyano-3,12-dioxooleana-1,9-dien-28-oic acid methyl ester (CDDO-Me) was from Cayman Chemicals (Ann Arbor, MI). N¹-(β-D-ribofuranosyl)-5-aminoimidazole-4-carboxamide (AICAR) and (3S)-1-[4-[(2,3,5,6-tetramethylphenyl) sulfonylamino]-1-naphthyl]pyrrolidine-3-carboxylic acid (RA839) were from Tocris Bioscience (Bristol, UK). All compounds were dissolved in DMSO, except resveratrol, which was dissolved in ethanol, and d-glucose and d-mannoheptulose, which were dissolved in sterile water. Since the baseline ghrelin secretion varies considerably from experiment to experiment, the effect of test compounds on acyl ghrelin secretion was calculated by dividing acyl ghrelin concentration in media after incubation with the test compounds by the mean acyl ghrelin concentration in the media of vehicle-treated wells in the same experiment and is described as fold change in acyl ghrelin secretion.

ELISA.

Acyl ghrelin concentrations in the cell culture supernatants were determined by ELISA (catalog no. EZRGRA-90K; Millipore-Merck, St. Charles, MO). The end-point calorimetric measurements for the ELISA were performed with a PowerWave XS Microplate spectrophotometer and KC4 junior software (BioTek Instruments, Inc., Winooski, VT).

FACS purification of mouse gastric ghrelin cells and quantitative real-time polymerase chain reaction to determine gene expression in gastric ghrelin cells.

To study gene expression in gastric ghrelin cells, we first crossed our previously characterized ghrelin-cre line (which faithfully express cre recombinase in ghrelin cells) (34) with a Rosa26-lox-STOP-lox-tdTomato reporter line [B6.Cg-Gt(ROSA)26Sor^{tm14(CAG-tdTomato)Hze}/J; The Jackson Laboratory] (32). The tdTomato+ gastric ghrelin cells were separated from tdTomato− gastric mucosal cells by fluorescence-activated cell sorting (FACS) using a MOFLO high-speed cell sorter (Beckman Coulter, Brea, CA) at the UT Southwestern Flow Cytometry Multi-User Core Facility by procedures described previously (34, 46). Thereafter, RNA was isolated from the cells with an Arcturus PicoPure RNA isolation kit (Applied Biosystems by Thermo Fisher Scientific, Wilmington, DE) and quantified with a Nanodrop Spectrophotometer (Thermofisher Scientific). Complementary DNA was synthesized by reverse transcription using SuperScript III (Invitrogen, Carlsbad, CA). Quantitative real-time polymerase chain (qPCR) was performed with an Applied Biosystems 7900HT Fast Real-Time PCR System (Applied Systems, Foster City, CA) and SYBR Green chemistry. The following primer sets designed with Primer Express Software (PerkinElmer Life Sciences, Boston, MA) were used to amplify ghrelin: 5′-GTCCTCACCACCAAGACCAT-3′, 5′-TGGCTTCTTGGATTCCTTTC-3′, Keap1: 5′-TGCCCCTGTGGTCAAAGTG-3′, 5′-GGTTCGGTTACCGTCCTGC-3′, and Nrf2: 5′-TGCAGCTTTTGGCAGAGACA-3′, 5′-TTGAGGGACTGGGCCTGAT-3′. The mRNA levels in tdTomato+ gastric ghrelin cells are represented relative to levels in tdTomato− gastric mucosal cells and were calculated by the comparative threshold cycle (∆∆C_t) method. The invariant control for the qPCR reaction was 18S rRNA with the Taqman Gene Expression Master Mix. Successful enrichment for ghrelin cells versus other gastric mucosal cell types was confirmed by demonstrating 2 × 10⁷-fold higher ghrelin mRNA expression in the tdTomato+ gastric mucosal cell pools compared with tdTomato− gastric mucosal cell pools.

Statistics.

All data are expressed as means ± SE. The statistical tests (2 sided) used are indicated in the figure legends and together with the graph preparations were performed with GraphPad Prism v8.3.0 (GraphPad, San Diego, CA). P values < 0.05 were considered statistically significant.

RESULTS

Inhibition of glucose metabolism attenuates glucose suppression of ghrelin secretion.

In our previous studies, we showed that glucose inhibits ghrelin secretion from cultured mouse gastric mucosal cells (36, 46). Also, as simultaneous application of a glucoprivic agent, 2-deoxy-d-glucose, prevented glucose suppression of ghrelin secretion, we proposed that the direct action of glucose on ghrelin cells was likely dependent on its entry into and metabolism within ghrelin cells (46). Here, we extended those findings by investigating the effects of d-mannoheptulose. d-mannoheptulose blocks the first step in glucose metabolism, phosphorylation of d-glucose to glucose-6-phosphate, by acting as a competitive inhibitor of hexokinases and glucokinase (2, 8). Glucokinase is highly enriched within FACS-separated gastric ghrelin cells, whereas hexokinase 1 and hexokinase 2 are enriched in ghrelinoma cell lines (46). As observed previously, d-glucose dose-dependently inhibited ghrelin secretion from cultured gastric mucosal cells (Fig. 1) (36, 46). No (0 mM) or low (1 mM) d-glucose increased ghrelin secretion by 93% and 42%, respectively compared with ghrelin secretion in 5 mM d-glucose (Fig. 1). High (10 mM) d-glucose reduced ghrelin secretion by 13% (Fig. 1). Addition of d-mannoheptulose to the culture medium containing the different d-glucose concentrations (0, 1, 5, and 10 mM) increased ghrelin secretion, with the increases reaching statistical significance at the 5 and 10 mM d-glucose concentrations (Fig. 1). Thus, these data suggest that glucose must be metabolized to glucose-6-phosphate for it to inhibit ghrelin secretion.

Fig. 1.

Effects of d-glucose and d-mannoheptulose on ghrelin secretion from primary cultures of mouse gastric mucosal cells: effects of 10 mM d-mannoheptulose on acyl ghrelin secretion in the presence of cell culture medium containing 0, 1, 5, or 10 mM d-glucose. Acyl ghrelin concentrations are expressed as fold change compared with acyl ghrelin measured in cells cultured in 5 mM d-glucose (316 ± 50 pg/mL). n = 6 each. Data were analyzed by 2-way ANOVA followed by Sidak’s post hoc test. The main effects of d-mannoheptulose and d-glucose were both significant (P < 0.0001). d-mannoheptulose-induced differences in acyl ghrelin concentrations measured by post hoc analysis: *P < 0.05, **P < 0.01. n.s., No significant difference.

Oxidative stress regulates ghrelin secretion.

Given the many studies indicating that glucose metabolism within cells changes oxidative stress within those cells (4, 6, 13, 17, 21, 29, 43, 44, 47, 55), we determined the effect of a well-known antioxidant, resveratrol (3,5,4′-trihydroxy-trans-stilbene), on ghrelin secretion with cultured mouse gastric mucosal cells. Resveratrol dose-dependently increased ghrelin secretion from cells cultured in 5 mM d-glucose (Fig. 2A). The highest resveratrol concentration tested (300 µM; log −3.5) increased ghrelin secretion by 4.2-fold (Fig. 2A). This dose of resveratrol was also highly effective at increasing ghrelin secretion above the levels observed in 0 mM, 1 mM, and 10 mM d-glucose (Fig. 2B). An analog of resveratrol, SRT1720 [N-(2-(3-(1-piperazinylmethyl)imidazo[2,1-b]thiazol-6-yl)phenyl)-2-quinolinecarboxamide], also dose-dependently increased ghrelin secretion at 5 mM d-glucose (Fig. 2C). Additionally, we tested whether the unrelated, structurally different antioxidant curcumin [(E,E)-1,7-bis(4-hydroxy-3-methoxyphenyl)-1,6-heptadiene-3,5-dione] impacts ghrelin secretion. Similar to the observations with resveratrol and SRT1720, curcumin, which is a polyphenol found naturally in turmeric, dose-dependently increased ghrelin secretion at 5 mM d-glucose (Fig. 2D). Collectively, these data indicate that lowering oxidative stress within ghrelin cells increases ghrelin secretion and blocks glucose-induced inhibition of ghrelin secretion.

Fig. 2.

Effects of antioxidants and compounds targeting resveratrol effectors on ghrelin secretion from primary cultures of mouse gastric mucosal cells. A: effects of different concentrations of resveratrol on acyl ghrelin secretion from cells grown in cell culture medium containing 5 mM d-glucose (n = 6). B: effect of 300 µM (log −3.5) resveratrol on acyl ghrelin secretion in the presence of different d-glucose concentrations (n = 6). C and D: effects of different concentrations of SRT1720 (C; n = 9) and curcumin (D; n = 9) on acyl ghrelin secretion from cells grown in medium with 5 mM d-glucose. E: effect of 1 µM EX-527 on resveratrol-induced increases in acyl ghrelin secretion from cells grown in 5 mM d-glucose (n = 3). F: effect of 500 µM AICAR on acyl ghrelin secretion from cells grown in the presence of different d-glucose concentrations (n = 9). Data were normalized to acyl ghrelin measured in vehicle-treated cells cultured in 5 mM d-glucose (A: 229 ± 66 pg/mL, B: 312 ± 36 pg/mL, C: 334 ± 29 pg/mL, D: 326 ± 25 pg/mL, E: 378 ± 10 pg/mL, F: 244 ± 17 pg/mL). Data in A, C, and D were analyzed by 1-way ANOVA followed by Sidak’s post hoc test, and data in B, E, and F were analyzed by 2-way ANOVA followed by Sidak’s post hoc test: *P < 0.05, **P < 0.01, ****P < 0.0001. n.s., No significant difference, No significant effects of EX-527 or AICAR were observed; a significant effect of d-glucose (increasing concentrations lower acyl ghrelin secretion) was observed in the AICAR experiment (F; P < 0.0001).

Notably, since the pharmacological actions of resveratrol are multifaceted, including not only antioxidant effects but also well-characterized effects to activate Sirtuin 1 (SIRT1) and 5′ adenosine monophosphate-activated protein kinase (AMPK), we tested whether either SIRT1 or AMPK activity modulates ghrelin secretion. Addition of 1 µM EX-527 (6-chloro-2,3,4,9-tetrahydro-1H-carbazole-1-carboxamide; Selisistat), a selective SIRT1 inhibitor (14), did not prevent resveratrol-induced increases in acyl ghrelin secretion, suggesting that activation of SIRT1 does not mediate resveratrol’s induction of ghrelin secretion (Fig. 2E). In addition, selective activation of AMPK with 500 µM AICAR [N¹-(β-D-ribofuranosyl)-5-aminoimidazole-4-carboxamide] (26, 51) did not increase ghrelin secretion, suggesting no role for activation of AMPK in mediating resveratrol’s induction of ghrelin secretion (Fig. 2F). We also ruled out endoplasmic reticulum (ER) stress as a cause of the glucose-induced changes in acyl ghrelin secretion, as pharmacological induction of ER stress with 1 µM tunicamycin did not alter acyl ghrelin secretion in the primary gastric mucosal cell culture model (data not shown).

Nrf2 regulates ghrelin secretion.

On the basis of the results above, we hypothesized that the transcription factor Nuclear factor erythroid-derived-2-like 2 (Nrf2), which is a master regulator of antioxidant response to counter oxidative stress (reviewed in 19), might be involved in ghrelin secretion. Of note, all of the antioxidants used above have been shown to activate Nrf2, which reduces oxidative stress via inducing gene expression of a variety of cytoprotective proteins (3, 7, 19, 25, 54, 56, 60). To more specifically assess the role of Nrf2, we first examined gene expression of Nrf2 and the related protein Kelch-like ECH-associated protein 1 (KEAP1) within ghrelin cells. Under basal conditions, the cytosolic protein KEAP1 negatively regulates Nrf2 by directing it toward ubiquitination and proteasomal degradation (19). In settings of oxidative stress, though, KEAP1 releases Nrf2, allowing it to orchestrate the aforementioned gene transcriptional changes and resulting robust antioxidative response (19). Within FACS-purified tdTomato+ gastric ghrelin cells derived from mice resulting from crosses of ghrelin-cre mice (34) with a Rosa26-lox-STOP-lox-tdTomato reporter line, Nrf2 and KEAP1 mRNA expression were enriched by 2.1-fold and 4.2-fold, respectively, compared with their expression within tdTomato− gastric mucosal cells (Table 1).

Table 1.

Relative mRNA expression within FACS-separated tdTomato+ gastric ghrelin cells compared with tdTomato− gastric mucosal cells

Gene	tdTomato− Cells	tdTomato+ Cells	Fold Increase in Expression	P Value
Keap1 (keap1)	3.192 ± 1.516 (36.04 ± 1.73)	13.47 ± 1.513 (34.34 ± 0.74)	4.2	0.001
Nrf2 (nfe2l2)	1.135 ± 0.481 (23.79 ± 1.41)	2.34 ± 0.945 (24.41 ± 1.03)	2.1	0.33
Ghrelin (ghrl)	2.488 ± 1.31 (44.98 ± 1.23)	50,250,641 ± 19,050,000 (22.12 ± 0.79)	2 × 107	>0.0001
18S rRNA	(10.10 ± 0.13)	(12.42 ± 0.68)

Values are expressed as means ± SE with threshold cycle (C_t) values in parentheses; n = 5 for tdTomato− cells, n = 4 for tdTomato+ cells. All values are normalized to 18S rRNA. Data were analyzed by Student’s t test. Keap1, Kelch-like ECH-associated protein 1; Nrf2, Nuclear factor erythroid-derived-2-like 2.

Since resveratrol, SRT1720, and curcumin all reduce oxidative stress via effects on multiple targets in addition to activating Nrf2, we also tested the effects on ghrelin secretion of three compounds that selectively modify the KEAP1-Nrf2 system. dl-sulforaphane [1-isothiocyanato-4-(methylsulfinyl)-butane], an isothiocyanate found in cruciferous vegetables that covalently modifies KEAP1 to activate Nrf2 (22), dose-dependently increased acyl ghrelin secretion from mouse gastric mucosal cells cultured in 5 mM d-glucose (Fig. 3A). CDDO-Me (2-cyano-3,12-dioxooleana-1,9-dien-28-oic acid methyl ester; also called bardoxolone methyl), a synthetic triterpenoid with a mechanism of action similar to dl-sulforaphane (22), overcame d-glucose-induced inhibition of ghrelin secretion, increasing ghrelin secretion under both low (1 mM)- and high (10 mM)-d-glucose conditions (Fig. 3B). RA839 [(3S)-1-[4-[(2,3,5,6-tetramethylphenyl) sulfonylamino]-1-naphthyl]pyrrolidine-3-carboxylic acid], a recently developed small molecule that noncovalently binds KEAP1 to activate Nrf2 (59), increased ghrelin secretion in high-glucose (10 mM) conditions (Fig. 3C). Overall, these results suggest that activation of Nrf2 stimulates ghrelin secretion, including in a high-glucose environment, which usually inhibits ghrelin secretion.

Fig. 3.

Effects of selective activators of Nuclear factor erythroid-derived-2-like 2 (Nrf2) on ghrelin secretion from primary cultures of mouse gastric mucosal cells. A: effects of dl-sulforaphane on acyl ghrelin secretion from cells grown in culture medium containing 5 mM d-glucose (307 ± 34 pg/mL). B: effects of CDDO-Me on acyl ghrelin secretion from cells grown in medium containing either 1 mM or 10 mM ambient d-glucose. C: effects of RA839 on acyl ghrelin secretion from cells grown in medium containing either 1 mM or 10 mM ambient d-glucose. Veh, vehicle. Data in B and C were normalized to acyl ghrelin measured in cells cultured in 5 mM d-glucose (323 ± 19 pg/mL). n = 6 for each condition. Data in A were analyzed by 1-way ANOVA followed by Sidak’s post hoc test, and data in B and C were analyzed by 2-way ANOVA followed by Sidak’s post hoc test: ***P < 0.001, ****P < 0.0001. n.s., Not significant.

DISCUSSION

Glucose regulation of ghrelin secretion is well established (35). Oral or parental administration of glucose lowers plasma ghrelin in humans and mouse models (5, 10, 16, 36, 38–40, 42, 49, 52). Also, primary cultures of gastric mucosal cells increase ghrelin secretion in conditions simulating hypoglycemia, whereas they decrease ghrelin secretion in conditions simulating hyperglycemia (36, 46). Here, using the mouse gastric mucosal cell primary culture model that we reported previously, we extend our previous findings of direct glucose inhibition of ghrelin secretion and provide insights into a possible mechanism mediating the direct glucose inhibition of ghrelin secretion (34, 46). Just as previous ex vivo studies with 2-deoxy-d-glucose suggested that d-glucose must be metabolized within ghrelin cells to directly inhibit ghrelin secretion, so too do the findings here with d-mannoheptulose. Moreover, we show that decreasing oxidative stress levels within ghrelin cells, either via compounds with multiple potential intracellular targets (resveratrol, SRT1720, and curcumin) or via compounds that selectively activate the transcription factor and master antioxidative response regulator Nrf2 (dl-sulforaphane, CDDO-Me, and RA839), stimulate ghrelin secretion and block the effect of glucose to inhibit ghrelin secretion (Fig. 4).

Fig. 4.

Model for regulation of ghrelin secretion by oxidative stress. The antioxidants resveratrol, SRT1720, and curcumin, which are known to activate Nuclear factor erythroid-derived-2-like 2 (Nrf2) and other biological targets to reduce oxidative stress, potently stimulate ghrelin secretion. dl-sulforaphane, CDDO-Me, and RA839—all selective activators of the transcription factor Nrf2, which is a master regulator of antioxidant response, acting to reduce oxidative stress by inducing gene expression of a variety of cytoprotective proteins—also stimulate ghrelin secretion. This collection of pharmacological antioxidants also attenuate glucose-induced inhibition of ghrelin secretion, suggesting higher oxidative stress levels within ghrelin cells resulting from glucose metabolism as a potential mechanism for glucose inhibition of ghrelin secretion.

Despite these new revelations, several questions remain unanswered. The first question relates to how glucose is metabolized once it undergoes the initial hexokinase/glucokinase-induced phosphorylation. Presumably, this might include further processing via the glycolytic pathway and/or perhaps further processing via the pentose phosphate pathway. Yet, if glucose’s effects on ghrelin secretion were mediated selectively via its further metabolism through the pentose phosphate pathway, we might have expected the resulting increased production of NADPH to reduce oxidative stress. Indeed, two of the known substrates for Nrf2 are glucose-6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase, the two pentose phosphate pathway enzymes that generate NADPH (19). However, our new data suggest instead that the overall direct effect of d-glucose on ghrelin cells is to induce oxidative stress rather than reduce it. Thus, a second question relates to glucose’s overall effect on the redox state of the ghrelin cell. Notably, such a proposed effect of high d-glucose to induce oxidative stress within ghrelin cells is not totally unexpected, as such has been shown in other cell types, such as pancreatic β-cells, neurons, and endothelial cells (4, 13, 17, 29, 41, 43, 44, 55). Yet, conversely, deprivation of d-glucose also induces oxidative stress in similar cell types (6, 21, 47, 57). Therefore, more studies are needed to both examine the pathways through which glucose is metabolized within ghrelin cells and confirm d-glucose’s overall effect on the redox state of ghrelin cells.

Regarding the effect of Nrf2 revealed here, it is also of interest that caloric restriction, which is a potent means of inducing ghrelin secretion in vivo, has been shown to activate Nrf2 in certain cell types (53). Therefore, although previous research has revealed a key role for activation of β₁-adrenergic receptors in mediating ghrelin secretion in response to caloric restriction (34, 64), it remains to be seen whether activation of the KEAP1-Nrf2 antioxidant pathway in ghrelin cells contributes to the increase in plasma ghrelin observed in caloric restricted rodent models and humans. Future studies will test the effect of glucose administration and caloric restriction on plasma ghrelin in mice that lack KEAP1 or Nrf2 selectively from ghrelin cells.

There are also some caveats to our study that should be kept in mind when interpreting the new data. First, it could be that the antioxidant effects of the pharmacological tools used in the present study have had such a strong effect that they have masked the actual physiological effects of glucose on the ghrelin cell machinery impacting ghrelin secretion. Second, although the new data presented indicate a marked effect of reducing oxidative stress on ghrelin secretion, we still do not know for certain how different metabolic challenges (such as rising glucose) or environmental stressors affect oxidative stress within ghrelin cells. The mixed population of primary gastric mucosal cells, in which ghrelin cells constitute only a small fraction of the total cells, and their poor adherence to plates or coverslips limit the ability to perform calorimetric plate reader experiments to measure reactive oxygen species within ghrelin cells or to perform histochemistry. The small fraction of ghrelin cells in the primary culture also limits the study of ghrelin synthesis and processing within the cell. Therefore, we would be better served by performing studies using a homogeneous population of ghrelin cells that can be grown in culture. However, although this has been achieved in the literature (31), our group has been unable to successfully grow purified ghrelin cells in primary culture after FACS separation. In an attempt to overcome this limitation, we previously generated SV40 large T-antigen-driven ghrelinoma cell lines derived from the stomach (SG1 cells) and pancreatic islets (PG1 cells) (64). Although these cell lines serve as very good model systems to study many aspects of ghrelin cell physiology (33, 64), they are not good models to study the effect of glucose on ghrelin secretion since they were found to be unresponsive to changes in d-glucose concentration within the culture media (35). This is likely related to the large T-antigen-mediated transformation, as such loss of glucose responsivity is also observed in other cell types transformed with the same process (50). Future studies should be directed at devising new strategies to investigate how alterations of redox state within ghrelin cells impact ghrelin secretion under different physiological settings.

GRANTS

This work was supported by the National Institutes of Health (Grant R01 DK-103884), the Diana and Richard C. Strauss Professorship in Biomedical Research, the Mr. and Mrs. Bruce G. Brookshire Professorship in Medicine, the Kent and Jodi Foster Distinguished Chair in Endocrinology, in Honor of Daniel Foster, M.D., and a gift from Mr. and Mrs. Robert E. Wegner.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

B.K.M. and J.M.Z. conceived and designed research; B.K.M., S.O.-L., and N.M. performed experiments; B.K.M. and S.O.-L. analyzed data; B.K.M. and J.M.Z. interpreted results of experiments; B.K.M. prepared figures; J.M.Z. edited and revised manuscript; B.K.M. drafted manuscript; B.K.M., S.O.-L., and J.M.Z. approved final version of manuscript.