Ghrelin Is Present in Pancreatic α-Cells of Humans and Rats and Stimulates Insulin Secretion

Male Wistar rats that weighed 300–350 g (Charles River Japan, Shiga, Japan) were used in all of the experiments. Rats were housed individually in plastic cages at constant room temperature in a 12-h light (0700–1900)/12-h dark cycle and were given standard laboratory diet and water ad libitum. All procedures were performed in accordance with the Japanese Physiological Society’s guidelines for animal care.

Two ghrelin-specific RIAs were developed as previously described (16). Two kinds of polyclonal antibodies were raised against the COOH-terminally Cys-extended rat ghrelin (position 1–11) with octanoylated Ser-3 and NH₂-terminally Cys-extended rat ghrelin (position 13–28) in New Zealand white rabbits. An anti–rat ghrelin (1–11) antiserum (#G606) specifically recognized ghrelin with n-octanoylated Ser 3 and did not recognize des-acyl ghrelin. An anti–rat ghrelin (13–28) antiserum (#G107) recognized n-octanoyl modified and des-acyl ghrelin equally. Both antisera had 100% cross-reactivity with human ghrelin (1–28).

Synthetic rat [Tyr²⁹]-ghrelin (1–28) and [Tyr⁰]-ghrelin (13–28) were radioiodinated by the lactoperoxidase method (16). Diluted samples or standard peptide solutions (100 μl) were incubated for 24 h with 100 μl of the antiserum diluent (final dilution of anti–ghrelin (1–11) antiserum, 1:620,000; final dilution of anti–ghrelin (13–28) antiserum, 1:20,000). A tracer solution (16,000 cpm in 100 μl) was added, and the mixture was incubated for another 24 h. The bound and free ligands were separated using a second antibody (200 μl). All of the procedures were performed at 4°C, and the samples were assayed in duplicate. Half-maximal inhibition by rat ghrelin (1–28) on the standard RIA curve with anti–ghrelin (1–11) antiserum was 3.8 fmol/tube and with anti-ghrelin (13–28) antiserum was 80 fmol. The intra- and interassay coefficients of variation for NH₂-terminus RIA were 3.5 and 3.2%, respectively, and those for COOH-terminus RIA were 3.7 and 3.3%, respectively.

Peptide extracts from three rat pancreata were loaded onto Sep-Pak C-18 cartridge (Waters, Milword, MA), and the bound peptides were eluted with 60% acetonitrile (CH₃CN) solution containing 0.1% trifluoroacetic acid (TFA). A fraction of the eluate was subjected to two RIAs for ghrelin, and the remainder was used for RP-HPLC under the conditions described in Fig. 1. All of the HPLC fractions were quantified by the ghrelin RIA. n-Octanoylated rat ghrelin (1–28) and des-acyl rat ghrelin (1–28) were chromatographed using the same HPLC system as standards.

Three Wistar rats were perfused transcardially with 0.1 mol/l phosphate buffer (pH 7.4) and then with 4% paraformaldehyde in a 0.1 mol/l phosphate buffer. Their pancreata were cut at −20°C with a cryostat into 12-μm-thick slices. Serial sections were incubated overnight at 4°C first with anti–ghrelin (1–11) antiserum (final dilution, 1:5,000), anti–ghrelin (13–28) antiserum (final dilution, 1:5,000), or anti-glucagon antiserum (NICHIREI, Tokyo, Japan; dilution 1:500), and then with Alexa Fluor 488 goat anti-rabbit IgG (Molecular Probes, Eugene, OR) or Alexa Fluor 568 goat anti-rabbit IgG. The slides were observed by fluorescence microscopy (BH2-RFC; Olympus, Tokyo, Japan).

Human pancreata obtained at autopsy from three patients who had died of cardiovascular disease were immersed in the same fixative for 24 h at 4°C. They were cut at −20°C with a cryostat into 12-μm-thick slices, then incubated overnight at 4°C with anti–ghrelin (1–11) antiserum (final dilution, 1:5,000), anti–ghrelin (13–28) antiserum (final dilution, 1:5,000), or anti-glucagon antiserum (dilution, 1:500). The slides were stained by the avidin-biotin-peroxidase complex method (Vectastain Elite ABC kit; Vector Laboratories, Burlingame, CA) as described previously (11).

The rat pancreatic sections were incubated overnight at 4°C with sheep anti-glucagon antiserum (Biogenesis, England, U.K.; dilution, 1:1,000) and then with Alexa Fluor 594 donkey anti-sheep IgG (Molecular Probes). Next, the sections were incubated with anti–ghrelin (13–28) antiserum (final dilution, 1:3,000) and then with Alexa Fluor 488 goat anti-rabbit IgG (Molecular Probes). The slides were observed by fluorescence microscopy (BH2-RFC; Olympus).

Total RNA was individually extracted from the pancreata and the isolated islets of Langerhans of Wistar rats by the acid guanidinium thiocyanate-phenol-chloroform method (17). First-strand cDNA was synthesized from 2.5 μg of RNA and oligo (dT)₁₈ primer using ReverTra Ace-α (Toyobo, Osaka, Japan) and then subjected to PCR amplification using primers specific for ghrelin or the ghrelin receptor as described elsewhere (11). The PCR products were electrophoresed on a 2% agarose gel (FMC BioProducts, Rockland, ME). Some portions of the PCR products of ghrelin and the ghrelin receptor were digested with PstI and HinfI, respectively, then electrophoresed. The other portions of the PCR products were purified with Magic PCR DNA Purification System (Promega, Madison, WI) and then sequenced by BigDye Terminator Cycle Sequencing System (Applied Biosystems, Foster City, CA).

Wistar rats aged 10–12 weeks were anesthetized with an intraperitoneal injection of sodium pentobarbital (80 mg/kg body wt). After the abdomen was opened, the common bile duct proximal to the pancreas was first ligated, and then collagenase (3 mg/ml) dissolved in 5 mmol/l Ca²⁺-containing Krebs-Ringer bicarbonate buffer (KRB) was injected into the common bile duct at the distal end. The pancreas was resected and incubated at 37°C for 17 min. Islets were hand-collected under a microscope and either used for the study of insulin release or dispersed into single cells by treatment with Ca²⁺-free KRB + 0.1 mmol/l EGTA. The single cells were plated on coverslips and maintained in a short-term culture for up to 2 days in Eagle’s minimum essential medium containing 5.6 mmol/l glucose supplemented with 10% fetal bovine serum, 100 mg/ml streptomycin, and 100 mU/ml penicillin at 37°C in a 95% air and 5% CO₂ atmosphere. The cells during this culture period responded to test agents in a consistent manner. β-Cells were selected by immunostaining with anti-insulin antiserum or by morphological and physiological criteria, as reported previously (18,19).

Insulin release from rat pancreatic islets was examined as described previously (19,20). Briefly, groups of seven isolated islets were first incubated for 30 min in KRB containing 2.8 mmol/l glucose for stabilization. The islets then were incubated at 37°C for 30 min in KRB containing 2.8 or 8.3 mmol/l glucose, either with or without the addition of 10⁻¹² mol/l ghrelin. At the end of the incubation period, aliquots of the medium were collected for the measurement of insulin concentration with an enzyme immunoassay kit (Morinaga, Yokohama, Japan).

[Ca²⁺]_i was measured by dual-wavelength fura-2 microfluorometry combined with digital imaging as reported previously (20). Briefly, cells on coverslips were incubated with 2 mmol/l fura-2 acetoxymethyl ester in KRB containing 2.8 mmol/l glucose for 30 min at 37°C. They were then mounted in a chamber and superfused with KRB at a rate of 1 ml/min at 37°C. The cells were excited alternately at 340 and 380 nm every 2.5 s, emission signals at 510 nm were detected with an intensified charge-coupled device camera, and ratio images were produced by an Argus-50 system (Hamamatsu Photonics, Hamamatsu, Japan). Ratio values were converted to [Ca²⁺]_i according to calibration curves (20).

All the data are presented as means ± SE, and comparisons between groups were performed by unpaired t test. The level of significance was set at P < 0.05.

RP-HPLC coupled with two separate RIAs was used to analyze ghrelin molecules in rat pancreas. In the RIA for the ghrelin COOH-terminus, two major peaks were observed, eluting at positions consistent with des-acyl (1–28) and n-octanoylated rat ghrelin (1–28) (Fig. 1). RIA for the ghrelin NH₂-terminus revealed only one major ghrelin-immunoreactive peak eluting at a position consistent with n-octanoylated rat ghrelin (1–28). The amount of ghrelin in the pancreas as measured by COOH-terminus RIA was 9.43 ± 0.76 pg/mg (mean ± SE), and that measured by NH₂-terminus RIA was 2.74 ± 0.71 pg/mg.

Ghrelin-immunoreactive cells were found in the mantle of the pancreatic islets, just like glucagon-producing α-cells (Figs. 2A–C and E–H). In rat islets, ghrelin-immunoreactive cells co-localized with glucagon in immunofluorescence double-staining (Fig. 2D). In humans, ghrelin-immunoreactive cells also had the same distribution as α-cells (Figs. 2E–H).

Ghrelin and ghrelin receptor transcripts corresponding to predicted sizes of 347 and 313 bp, respectively, were found in the pancreas and the isolated rat islet (Fig. 3). The PCR product of ghrelin was cleaved into 281- and 66-bp fragments on PstI digestion, the sizes expected (Fig. 3B). The product of the ghrelin receptor was cleaved into 176- and 137-bp fragments on HinfI digestion, the sizes expected (Fig. 3B). The PCR products of 347 and 313 bp were also confirmed to correspond to ghrelin and the ghrelin receptor, respectively, by direct sequencing (data not shown).

Insulin release from isolated rat islets under static incubation condition was stimulated by 8.3 mmol/l glucose. The glucose-stimulated insulin release was significantly (P < 0.05) increased by the addition of 10⁻¹² mol/l ghrelin. In contrast, 10⁻¹² mol/l ghrelin had no effect on the basal insulin release in the presence of 2.8 mmol/l glucose (Fig. 4).

Ghrelin at 10⁻¹² mol/l increased [Ca²⁺]_i in single rat pancreatic β-cells in the presence of 8.3 mmol/l glucose (Fig. 5A), whereas it had no effect on [Ca²⁺]_i in the presence of 2.8 mmol/l glucose (Fig. 5B). In the β-cells that exhibited glucose-induced [Ca²⁺]_i oscillations, the peak of the [Ca²⁺]_i increase in response to ghrelin administration was much higher than the peaks of [Ca²⁺]_i oscillations (Fig. 5A). This effect could be considered the potentiation of glucose-induced [Ca²⁺]_i oscillations. Ghrelin at 10⁻¹² mol/l either elicited [Ca²⁺]_i increases or potentiated [Ca²⁺]_i oscillations in 20 of 59 (34%) β-cells at 8.3 mmol/l glucose. The time lag until the onset of [Ca²⁺]_i responses ranged from 1 to 4 min.

The pancreas and liver develop from a common precursor (the hepatopancreatic ring) that is derived from the gastrointestinal tract (the anterior primitive intestine). In the most highly evolved vertebrates, a close functional connection persists between the gastrointestinal endocrine cells and the pancreatic islets via the “enteroinsular” axis (21). We identified the cellular origin and distribution of ghrelin in the gastrointestinal tracts of rats and humans by in situ hybridization histochemistry, immunohistochemistry, electron microscopy, RIA, and RT-PCR (11). Ghrelin production in the rat stomach has been localized to “previously described” X cells, which are round to ovoid, with round, compact, electron-dense granules. Although these cells, which account for ∼20% of all endocrine cells in the oxyntic mucosae of both rats and humans, represent a major endocrine cell population in the oxyntic gland (22,23), their hormonal product and physiological functions have not previously been clarified. The present study shows that ghrelin is present in the α-cells of rat and human pancreas. X cells also have been called A(α)-like cells; although they do not exhibit glucagon-immunoreactivity, they do share some morphological features with pancreatic α-cells, including the presence of compact and dense secretory granules. Although α-cells are lacking or scarce in the gastric fundus of mammals except for dogs and cats, they are fairly abundant in the fundic mucosa of human fetuses (24,25). These findings suggest that pancreatic α-cells and gastric ghrelin cells may originate from the same endodermal progenitor cells.

The pancreatic islets vary in size and cellularity but generally are composed of four main cell types: α (20% of the total), β (68%), δ (10%), and PP (2%) (26). Insulin-producing β-cells are located in the center of the islet, and glucagon-producing α-cells are found at the periphery, with somatostatin-producing δ-cells interposed between the two. One widely accepted model of the vascular supply in the islets proposes that arterial blood is first carried to the β-cell–rich center before passing on to the α- and δ-cells in the periphery (27). Both ghrelin and its receptor are present in the pancreatic islets. Therefore, ghrelin in α-cells may affect islet functions via the systemic circulation. There is another possibility that ghrelin may interact with insulin, somatostatin, or both in a paracrine manner independent of blood flow, because α-, β-, and δ-cells are often in intimate contact within the islets and gap junctions exist between these cells.

Ghrelin at 10⁻¹² mol/l stimulated insulin release and increased [Ca²⁺]_i in rat islet β-cells in the presence of a stimulatory (8.3 mmol/l) but not basal (2.8 mmol/l) glucose concentration. Ghrelin at a higher concentration of 10⁻⁸ mol/l showed lesser effects on both insulin release and [Ca²⁺]_i increase, and at 10⁻¹⁴ mol/l it had no effects (data not shown). Ghrelin promotes calcium release after it binds to the ghrelin receptor (4). These findings suggest that ghrelin is a novel potentiator of insulin release and that Ca²⁺ may be a messenger signal for ghrelin in β-cells. Considering that the plasma level of n-octanoylated ghrelin in rats is 4 × 10⁻¹² mol/l (16), the effects of ghrelin at 10⁻¹² mol/l on insulin release and [Ca²⁺]_i increase are thought to be physiological. Ghrelin may serve as a physiological regulator of insulin release. Because increases and oscillations of [Ca²⁺]_i have been implicated not only in insulin release but also in insulin synthesis and gene expression (13,14,15), ghrelin might serve as a trophic factor for islet β-cells, although additional studies are definitely needed.

Ghrelin stimulates insulin secretion, whereas ghrelin secretion from α-cells may be regulated by the plasma concentration of glucose and/or insulin. We recently showed that ghrelin mRNA expression in the rat stomach is upregulated upon fasting and insulin-induced hypoglycemia (28). These findings suggest that ghrelin may function as an anabolic signal molecule during energy depletion. The presence of ghrelin and its receptor in the pancreatic islets provides a new clue for our understanding of the regulation of energy homeostasis.

This study was supported in part by Special Coordination Funds for Promoting Science and Technology from the Science and Technology Agency (Encouragement System of COE) (H.M.); a grant-in-aid for the Promotion of Fundamental Studies in Health Science from the Organization for Pharmaceutical Safety and Research of Japan (K.K.); grants-in-aid from the Ministry of Education, Culture, Sports, Science and Technology, Japan (M.N. and T.Y.); the Ministry of Health, Labor and Welfare, Japan (M.N.); the Novartis Foundation (Japan) for the Promotion of Science (M.N.); the Society of Molecular Mechanism of the Digestive Tract (M.N.); and the Japan Diabetes Foundation (T.Y.).