- Email: [email protected]
Role of Ghrelin in the Relationship Between Hyperphagia and Accelerated Gastric Emptying in Diabetic Mice
GASTROENTEROLOGY 2008;135:1267–1276
Role of Ghrelin in the Relationship Between Hyperphagia and Accelerated Gastric Emptying in Diabetic Mice PIETER–JAN VERHULST,* BETTY DE SMET,* INGE SAELS,* THEO THIJS,* LUC VER DONCK,‡ DIEDER MOECHARS,‡ THEO L. PEETERS,* and INGE DEPOORTERE*
Background & Aims: Ghrelin is an orexigenic peptide with gastroprokinetic effects. Mice with streptozotocin (STZ)-induced diabetes exhibit hyperphagia, altered gastric emptying, and increased plasma ghrelin levels. We investigated the causative role of ghrelin herein by comparing changes in ghrelin receptor knockout (growth hormone secretagogue receptor [GHS-R]ⴚ/ⴚ) and wild-type (GHS-Rⴙ/ⴙ) mice with STZ-induced diabetes. Methods: Gastric emptying was measured with the [13C]octanoic acid breath test. The messenger RNA (mRNA) expression of neuropeptide Y (NPY), agoutirelated peptide (AgRP), and proopiomelanocortin was quantified by real-time reverse-transcription polymerase chain reaction. Neural contractions were elicited by electrical field stimulation in fundic smooth muscle strips. Results: Diabetes increased plasma ghrelin levels to a similar extent in both genotypes. Hyperphagia was more pronounced in GHS-Rⴙ/ⴙ than in GHS-Rⴚ/ⴚ mice between days 12 and 21. Increases in NPY and AgRP mRNA expression were less pronounced in diabetic GHS-Rⴚ/ⴚ than in GHS-Rⴙ/ⴙ mice from day 15 on, whereas decreases in proopiomelanocortin mRNA levels were similar in both genotypes. Gastric emptying was accelerated to a similar extent in both genotypes, starting on day 16. In fundic smooth muscle strips of diabetic GHS-Rⴙ/ⴙ and GHS-Rⴚ/ⴚ mice, neuronal relaxations were reduced, whereas contractions were increased; this increase was related to an increased affinity of muscarinic and tachykinergic receptors. Conclusions: Diabetic hyperphagia is regulated by central mechanisms in which the ghrelin-signaling pathway affects the expression of NPY and AgRP in the hypothalamus. The acceleration of gastric emptying, which is not affected by ghrelin signaling, is not the cause of diabetic hyperphagia and probably involves local contractility changes in the fundus.
I
n 1996, the growth hormone secretagogue receptor (GHS-R) was cloned and identified as the receptor for a family of synthetic growth hormone secretagogues.1 The endogenous ligand of this receptor, ghrelin, an acylated 28-amino acid peptide, was identified 3 years later2 and, as expected, was shown to stimulate growth hormone release. Because ghrelin is mainly produced by the
stomach, an organ well positioned to detect recently ingested food, it was soon demonstrated that ghrelin plays an important role in the regulation of the energy balance. Peripheral administration of ghrelin induces weight gain by decreasing fat utilization3 and by stimulating food intake in rodents4 as well as in humans.5 A role for ghrelin in meal initiation was also confirmed by the fluctuations in plasma ghrelin levels, which increase immediately before a meal and rapidly fall after food intake, following a 24-hour pattern reciprocal to that of insulin.6 This raises the question whether insulin negatively regulates ghrelin. Indeed, increased plasma ghrelin levels were observed during insulin deficiency in rodents with streptozotocin (STZ)-induced diabetes.7,8 Also in patients with type 1 diabetes, absolute insulin deficiency prevented prandial plasma ghrelin suppression until the insulin deficiency was corrected with an intravenous insulin bolus.9 In both conditions, the hyperghrelinemia is thought to play a role in the hyperphagia associated with uncontrolled type 1 diabetes. Important evidence for this hypothesis was provided in a study with diabetic ghrelin knockout mice.8 In the absence of its endogenous ligand, the ghrelin receptor still maintains an important level of signaling, and it is unclear to what extent this contributes to the hyperphagia.10 The physiologic importance of this constitutive activity in humans was demonstrated in 2 unrelated families harboring a missense mutation in the GHS-R, which leads to a syndrome characterized not only by short stature but also by obesity.11 Ghrelin stimulates food intake through activation of the ghrelin receptor present on neuropeptide Y (NPY)/ agouti-related peptide (AgRP) neurons in the arcuate nucleus either by crossing the blood/brain barrier or via activation of vagal nerve activity.5,12–14 Ghrelin also has important effects on gastrointestinal motility that may contribute to appetite signaling. It induces strong “hunAbbreviations used in this paper: ACh, acetylcholine; AgRP, agoutirelated peptide; EFS, electrical field stimulation; GHS-R, growth hormone secretagogue receptor; NPY, neuropeptide Y; POMC, proopiomelanocortin; STZ, streptozotocin. © 2008 by the AGA Institute 0016-5085/08/$34.00 doi:10.1053/j.gastro.2008.06.044
BASIC– ALIMENTARY TRACT
*Centre for Gastroenterological Research, Catholic University of Leuven, Leuven, Belgium; and ‡Johnson & Johnson Division of Pharmaceutical Research and Development, Janssen Pharmaceutica, Beerse, Belgium
1268
VERHULST ET AL
BASIC– ALIMENTARY TRACT
ger” contractions in the fasted state originating in the stomach and migrating distally15,16 and accelerates gastric emptying17–20 in rodents and man. Until now, it is unclear to what extent the prokinetic effects of ghrelin contribute to its effects on food intake. The aim of our study was therefore multiple. First, we determined the functional relevance of the ghrelinsignaling pathway in diabetic hyperphagia by eliminating the constitutive activity of the ghrelin receptor. To this end, we compared food intake in STZ-induced diabetic ghrelin receptor knockout (GHS-R⫺/⫺) and wild-type (GHS-R⫹/⫹) mice. Second, the contribution of a central pathway in the ghrelin-dependent regulation of hyperphagia was determined by comparing alterations in the expression of orexigenic and anorexigenic neuropeptides in the hypothalamus of both genotypes. Third, we investigated whether altered gastric emptying, known to occur in rats with STZ-induced diabetes,21,22 may contribute to hyperphagia, and we determined the role of ghrelin herein. Finally, we studied the effect of diabetes on local alterations in the in vitro contractility of smooth muscle strips of the fundus to elucidate whether peripheral mechanisms mediate the changes in gastric emptying. This multiparametric analysis should help us reveal the role of the ghrelin-signaling pathway in hyperphagia of mice with uncontrolled diabetes. It should also clarify whether central or peripheral pathways are involved.
GASTROENTEROLOGY Vol. 135, No. 4
Materials and Methods Animals Male (40 –50 weeks of age) GHS-R⫹/⫹ and GHSmice were housed in a temperature-controlled environment (20°C–22°C) under a 14-hour:10-hour lightdark cycle and had ad libitum access to food and drinking water. This research was approved by the Ethical Committee for Animal Experiments of the Catholic University of Leuven.
R⫺/⫺
Generation of GHS-Rⴚ/ⴚ Mice GHS-R⫺/⫺ mice were developed by Janssen Pharmaceutica (Beerse, Belgium) in collaboration with Lexicon Genetics, Inc (The Woodlands, TX). With a polymerase chain reaction (PCR) probe, genomic clones were isolated by screening of the 129SvEvBrd-derived lambda pKOS genomic library.23 A 9.5-kilobase genomic clone spanning exon 1 and exon 2 was used to generate the targeting vector via yeast-mediated homologous recombination. In this vector, a 1325-base pair genomic fragment, spanning exon 1 and exon 2, was replaced by a floxed version of exon 1 and exon 2 including a 1.7kilobase PGK-neo selection cassette flanked by 2 Frt sites (Figure 1). The NotI-linearized vector was electroporated into 129 Sv/Evbrd (LEX1) embryonic stem (ES) cells, and G418-fialuridine (FIAU)-resistant ES cell clones were isolated and analyzed for homologous recombination by
Figure 1. Targeted disruption of the GHS-R1 gene. (A) Structure of the targeting vector, wild-type locus, targeted locus, and Cre-excised locus. Boxes representing the exons in open and shaded boxes are the noncoding and coding regions, respectively. Arrows indicate the position of the PCR primers used for genotyping the wildtype and targeted allele. (B) Expression of the GHS-R1 transcript in brain and pituitary of both genotypes. Results are means ⫾ SEM (n ⫽ 6) relative expression levels after normalization to -actin.
October 2008
ROLE OF GHRELIN IN DIABETIC MICE
Experimental Design GHS-R⫹/⫹ (n ⫽ 26) and GHS-R⫺/⫺ mice (n ⫽ 22) were fasted overnight (14 hours) and injected intraperitoneally with 160 mg/kg STZ dissolved in 5 mmol/L sodium citrate buffer (pH 4.0). Body weight, 24-hour food intake, and tail-blood glucose levels (Accu-Chek Sensor Comfort; Roche Diagnostics, Mannheim, Germany) of 14 GHS-R⫹/⫹ and 11 GHS-R⫺/⫺ were monitored daily at 10 AM from 3 days before the injection of STZ until 27 days after. Mice were subjected to a 13C gastric emptying breath test before induction of STZ diabetes and 6, 10, 16, and 22 days thereafter. At day 27, mice (6 GHS-R⫹/⫹, 6 GHS-R⫺/⫺) were killed at 10 AM. Hypothalami were snap frozen in liquid nitrogen for real-time PCR
Table 1. Forward and Reverse Primer Sequences Gene
Primer sequences
A. Wild-type allele
5=-TgggggTgCgAACATTAgC-3= (forward) 5=-CTgAAggCATCTTTCACTACg-3=(reverse) 5=-ACATATTCTATgTgAggCACC-3=(forward) 5=-CTgAAggCATCTTTCACTACg-3=(reverse) 5=-CCgCCTCTggCAgTATCg-3= (forward) 5=-gCTgACAAACTggAAgAgTTTgC-3=(reverse) 5=-CCCTggAACTTCggCgACCTgC-3= [5=]FAM [3=]TAMRA (probe) 5=-CATCTTggCCTCACTgTCCAC-3=(forward) 5=-gggCCggACTCATCgTACT-3=(reverse) 5=-TgCTTgCTgATCCACATCTgCTggA-3= [5=]FAM [3=]TAMRA (probe) 5=-gCggAggTgCTAgATCCA-3= (forward) 5=-AggACTCgTgCAgCCTTA-3= (reverse) 5=-CCgCTCTgCgACACTACAT-3= (forward) 5=-TgTCTCAgggCTggATCTCT-3= (reverse) 5=-ACCTCACCACggAgAgCA-3= (forward) 5=-gCgAgAggTCgAgTTTgC-3= (reverse) 5=-TCAgggACCAgAACCACAAA-3= (forward) 5=-CCAgCAgAggATgAAAgCAA-3= (reverse) 5=-CCAgAggACAgAggACAAgC-3= (forward) 5=-ACATCgAAgggAgCATTgAA-3= (reverse) 5=-CCCCAATgTgTCCgTCgTg-3= (forward) 5=-gCCTgCTTCACCACCTTCT-3= (reverse)
Knockout allele B. GHS-R
-actin
C. AgRP NPY POMC GHS-R Ghrelin GAPDH
analysis, fat pads (epididymal, inguinal, and retroperitoneal fat) were collected and weighed, blood was collected by cardiac puncture for ghrelin measurements, and fundus was dissected for in vitro contractility measurements. Another group of diabetic GHS-R⫹/⫹ (n ⫽ 12) and GHSR⫺/⫺ (n ⫽ 11) mice was not used in the gastric emptying study but was killed for real-time PCR analysis and ghrelin measurements at day 7 (5 GHS-R⫹/⫹, 5 GHS-R⫺/⫺) and day 15 (7 GHS-R⫹/⫹, 6 GHS-R⫺/⫺). A group of age-matched nondiabetic ad libitum fed mice (7 GHSR⫹/⫹, 5 GHS-R⫺/⫺) was used as a control group for real-time PCR analysis, plasma ghrelin measurements, and contractility studies.
Radioimmunoassay for Plasma Ghrelin Levels Acidified plasma samples from ad libitum fed mice (10 AM) were extracted on a Sep-Pak C18 cartridge (Waters Corporation, Milford, MA). The radioimmunoassay was performed with [125I] rabbit ghrelin as tracer (Bachem, Torrance, CA) and with a rabbit antibody raised against human ghrelin[14 –28] (Eurogentec, Seraing, Belgium) (final dilution 1:8000), which recognizes both octanoylated and desoctanoylated ghrelin. Intraassay coefficient of variation of the radioimmunoassay was 6.4% and 7.7%, respectively. The minimal detectable dose was 15.6 pg/mL.
Quantitative Real-Time PCR Total RNA was extracted from the hypothalamus by means of the TRIzol reagent and reverse transcribed to complementary DNA (cDNA) with Superscript II Reverse Transcriptase. The quantitative real-time PCR reaction was run on a Lightcycler 480 system (Roche Diagnostics, Mannheim, Germany) with LightCycler 480 SYBR Green I Master mix. Primer sequences are shown in Table 1, section C. An interrun calibrator was used, and a standard curve was created for each gene to obtain PCR efficiencies. Relative expression levels of all samples were calculated with the LightCycler 480 software and were expressed relative to GAPDH and corrected for interrun variability.
Breath Test for Gastric Emptying Gastric emptying was measured during the light phase (10 AM) with a noninvasive [13C]octanoic acid breath test in fasted (19 hours) mice according to the method of Kitazawa et al,19 except that the amount of 13CO in the exhaled air was determined by an Infrared 2 Isotope Analyser (IRIS, Wagner, Germany).
In Vitro Contractility Studies With Smooth Muscle Strips From the Fundus The fundus from control (5 GHS-R⫹/⫹, 3 GHSand diabetic (4 GHS-R⫹/⫹, 4 GHS-R⫺/⫺) mice was freed from mucosa, and strips were cut and suspended along their circular axis in a tissue bath filled with Krebs
R⫺/⫺)
BASIC– ALIMENTARY TRACT
Southern blot analysis. Targeted ES cell clones were injected into C57BL/6 (albino) blastocysts, and the resulting chimeras were mated to C57BL/6 (albino) females to generate animals heterozygote for the floxed GHS-R1 allele. These were subsequently crossed with ProtamineCre mice.24 Male descendants heterozygote for both the floxed GHS-R1 allele and the Protamine Cre transgene were crossed to C57Bl/6 females to obtain heterozygote GHS-R1 knockout animals. These were subsequently crossed to generate the genotypes used in the study. PCR was used to screen genotypes by using DNA isolated from mouse tail biopsy samples. Real-time reversetranscription (RT) PCR analysis was used to show expression or absence of the GHS-R1 transcript. Primers sequences are summarized in Table 1, sections A and B, respectively.
1269
1270
VERHULST ET AL
GASTROENTEROLOGY Vol. 135, No. 4
solution as described by De Smet et al.25 The strips’ response to increasing concentrations of ACh (10⫺4 to 10⫺9 mol/L) and substance P (10⫺5 to 10⫺9 mol/L) was measured isometrically. The negative logarithm of the median effective concentration (pEC50) value and the maximal contraction were calculated with the GraphPad Prism 4.00 software (San Diego, CA). Neural responses were elicited by electrical field stimulation (EFS) of strips as described previously.25 Responses were characterized pharmacologically by repeating the frequency spectrum in the presence of L-NAME (3 ⫻ 10⫺4 mol/L) in combination with atropine, the NK1-antagonist SR140333 (5 ⫻ 10⫺7 mol/L), the NK2antagonist SR48968 (5 ⫻ 10⫺7 mol/L), and tetrodotoxin (3 ⫻ 10⫺4 mol/L), respectively. Strips were pretreated with each antagonist for 25 minutes before EFS was applied. The on-relaxations were expressed relative to the response induced by 10⫺5 mol/L nitroglycerin. The offcontractions were expressed in grams/square millimeters.
Statistical Analysis
BASIC– ALIMENTARY TRACT
Results are presented as means ⫾ SEM. Data (body weight, glucose levels, food intake, gastric emptying, and in vitro contractility) obtained from repeated measurements performed at different time points in the same mice were analyzed using the Proc Mixed procedure with slice option from SAS (SAS Institute Inc, Cary, NC). All other data were analyzed with a factorial analysis of variance followed by Newmann–Keuls as a post hoc test (Statistica 6.0; StatSoft, Tulsa, OK). Significance was accepted at the 5% level.
Results Generation of the GHS-R Knockout Mice The strategy applied resulted in the deletion of the first 2 exons that encode GHS-R1 (ENSEMBL: ENSMUSG00000051136) (Figure 1A). Correct targeting in ES cells was confirmed by Southern analysis, and PCR analysis demonstrated the ablation of the wildtype GHS-R1 allele (results not shown). Loss of expression of the GHS-R1 transcript in the knockout mice was confirmed by quantitative real-time PCR. The GHS-R1 transcript was absent in the brain and pituitary derived from the homozygote GHS-R⫺/⫺ mice (Figure 1B).
Effect of STZ Diabetes on Blood Glucose Levels, Food Intake, and Body Weight Both genotypes displayed a significant (P ⬍ .001) increase in blood glucose levels within 24 hours after the injection of STZ (Figure 2A). Blood glucose levels were maximal from day 3 onward and did not differ significantly between both genotypes during the entire observation period. In both genotypes, body weight loss was apparent from day 2 after the induction of diabetes and
Figure 2. Changes in blood glucose levels (A), body weight (B), and daily (24 hours) food intake (C) in STZ-induced diabetic GHS-R⫹/⫹ and GHS-R⫺/⫺ mice as a function of time. Results are represented as the means ⫾ SEM (n ⫽ 11–14 animals/group). *P ⬍ .05; **P ⬍ .01; ***P ⬍ .001 (GHS-R⫹/⫹ vs GHS-R⫺/⫺ mice).
did not differ between GHS-R⫹/⫹ mice (⫺24.7% ⫾ 1.8%) and GHS-R⫺/⫺ mice (⫺25.6% ⫾ 3.8%) during the course of the experiment (Figure 2B). Total fat pad mass was similar in GHS-R⫹/⫹ and GHS⫺/⫺ R mice at the start of the experiment and amounted to 4.4% ⫾ 0.7% and to 4.3% ⫾ 0.6% of body weight, respectively. At day 27 after the induction of diabetes, no epididymal, inguinal, or retroperitoneal fat was detectable in either genotype (data not shown). Daily 24-hour food intake did not differ between both genotypes before the start of the experiment. After the induction of diabetes, food intake was significantly (P ⬍ .05) increased at day 5 in the wild-type mice. Mutant mice also became hyperphagic, but this did not reach statistical significance until day 9 (P ⬍ .01). This hyperphagic state remained during the further course of the experi-
October 2008
ment, but hyperphagia was significantly less (P ⬍ .01) in the GHS-R⫺/⫺ than in the GHS-R⫹/⫹ mice between days 12 and 21 (Figure 2C).
Effect of STZ Diabetes on Plasma Ghrelin Levels, Ghrelin Messenger RNA, and GHS-R Messenger RNA Expression in the Hypothalamus
1271
R⫺/⫺ mice was observed. As expected, GHS-R mRNA expression was absent in the mutant mice.
Effect of STZ Diabetes on NPY, AgRP, and Proopiomelanocortin mRNA Expression in the Hypothalamus In the wild-type mice, induction of diabetes led to a gradual and significant increase in NPY and AgRP mRNA expression from day 15 onward (P ⬍ .01) and a significant decrease in proopiomelanocortin (POMC) mRNA expression, which was already maximal from day 7 onward (P ⬍ .001) (Figure 3D–F). At day 27, a 35-fold increase in AgRP mRNA, a 9-fold increase in NPY mRNA, and a 7-fold decrease in POMC mRNA expression were observed. The increase in NPY and AgRP mRNA expression started to diverge between diabetic GHS-R⫹/⫹ and GHS-R⫺/⫺ mice from day 15 onward. No difference in POMC expression was observed between both genotypes.
Effect of STZ Diabetes on Gastric Emptying No difference in gastric half excretion time (Thalf) was observed between wild-type (Thalf: 109 ⫾ 5 minutes) and mutant mice (Thalf: 110 ⫾ 7 minutes) before the onset of diabetes. Gastric emptying of the solid meal was markedly accelerated in both genotypes at day 16 (Thalf: GHS-R⫹/⫹: BASIC– ALIMENTARY TRACT
Before the induction of diabetes, plasma ghrelin levels did not differ between both genotypes, neither in the fed (GHS-R⫹/⫹: 528, GHS-R⫺/⫺: 435 pg/mL) nor in the fasted state (GHS-R⫹/⫹: 1221, GHS-R⫺/⫺: 1193 pg/ mL). In diabetic GHS-R⫹/⫹ and GHS-R⫺/⫺ mice, plasma ghrelin levels were significantly (P ⬍ .05) increased at day 7 after the induction of diabetes and remained elevated during the further course of the experiment (Figure 3A). Changes in plasma ghrelin levels did not differ significantly between both genotypes. In the hypothalamus of diabetic GHS-R⫹/⫹ mice, the time course of the changes in ghrelin and GHS-R messenger RNA (mRNA) expression was bell-shaped with maximal increases (P ⬍ .01) at day 15 (Figure 3B and C). At day 27, ghrelin mRNA expression was normalized, whereas GHS-R mRNA expression was decreased (P ⬍ .01) compared with day 0. No difference in hypothalamic ghrelin mRNA expression between GHS-R⫹/⫹ and GHS-
ROLE OF GHRELIN IN DIABETIC MICE
Figure 3. Time-dependent effects of diabetes on plasma ghrelin levels (A) and hypothalamic mRNA expression of ghrelin (B); GHS-R (C); and the neuropeptides AgRP (D), NPY (E), and POMC (F) in ad libitum fed (10 AM) GHS-R⫹/⫹ and GHS-R⫺/⫺ mice before and 7, 15, and 27 days after the induction of diabetes. Results are represented as the means ⫾ SEM (n ⫽ 3–7 animals/group). *P ⬍ .05; **P ⬍ .01; ***P ⬍ .001 (GHS-R⫹/⫹ vs GHSR⫺/⫺ mice).
1272
VERHULST ET AL
GASTROENTEROLOGY Vol. 135, No. 4
Figure 4. Time-dependent changes in gastric half-excretion time (Thalf) (A) and lag time (Tlag) (B) as determined by the [13C]octanoic acid breath test in diabetic GHS-R⫹/⫹ and GHS-R⫺/⫺ mice. Results are means ⫾ SEM (n ⫽ 11–14 animals/group). †P ⬍ .01 (GHS-R⫹/⫹, compared with day 0); ‡P ⬍ .05; ‡‡P ⬍ .01; ‡‡‡P ⬍ .001 (GHS-R⫺/⫺, compared with day 0).
56 ⫾ 10 minutes, P ⬍ .01, and GHS-R⫺/⫺: 55 ⫾ 9 minutes, P ⬍ .001) and day 22 (GHS-R⫹/⫹: 56 ⫾ 18 minutes, P ⬍ .01, and GHS-R⫺/⫺: 48 ⫾ 5 minutes, P ⬍ .001) after induction of diabetes (Figure 4A). Similar effects were observed on lag time (Figure 4B). Gastric emptying correlated with food intake in diabetic wild-type (r ⫽ 0.48, P ⬍ .01) and mutant mice (r ⫽ 0.65, P ⬍ .01).
Effect of STZ Diabetes on the In Vitro Contractile Responses in the Mouse Fundus Neuronal responses under nondiabetic conditions. The responses of fundic smooth muscle strips, BASIC– ALIMENTARY TRACT
subjected to EFS, were frequency dependent and consisted mainly of on-relaxations. Off-contractions were observed at higher frequencies (8 –32 Hz). EFS-induced on-relaxations were more pronounced (P ⬍ .05) between 2 and 16 Hz in GHS-R⫺/⫺ than in GHS-R⫹/⫹ mice (Figure 5A). Within the same frequency range, EFS-induced offcontractions were significantly (P ⬍ .05) increased in GHS-R⫺/⫺ compared with GHS-R⫹/⫹ mice (Figure 5B). Effect of STZ diabetes on neuronal responses. A representative tracing of the neuronal response of fundic strips from GHS-R⫹/⫹ and GHS-R⫺/⫺ mice 27 days after the induction of diabetes is shown in Figure 6A and B. Changes in tension are summarized in the same Figure (Figure 6C–F). In both genotypes, inhibitory responses were reduced at low frequencies and abolished at high (16 –32 Hz) frequencies (Figure 6C and D). In addition, the poststimulus off-contractions were strongly enhanced after the induction of diabetes at all investigated
Figure 5. Electrical field stimulation of strips from nondiabetic GHS-R⫹/⫹ and GHS-R⫺/⫺ mice resulted in onrelaxations (A) and off-contractions (B). Results are means ⫾ SEM (n ⫽ 3 animals/group; n ⫽ 3 strips/animal). *P ⬍ .05 (GHS-R⫹/⫹ vs GHS-R⫺/⫺ mice). On-relaxations are expressed as a percentage of the maximal relaxation obtained after stimulation with 10⫺5 mol/L nitroglycerin; offcontractions are expressed in grams/square millimeters.
frequencies in wild-type (P ⬍ .001) as well as in mutant (P ⬍ .05) mice (Figure 6E and F). No significant differences in the contractile responses were observed between diabetic GHS-R⫹/⫹ and GHS-R⫺/⫺ mice. Effect of STZ diabetes on the contractile response to ACh and substance P. Neural responses of
fundic strips from nondiabetic mice were characterized pharmacologically (Figure 7A and B). Pretreatment with the NO-synthase inhibitor L-NAME reversed the on-relaxations into strong contractions over the entire frequency spectrum. These contractions were abolished by addition of atropine at low frequencies. The atropineresistant contractions were further reduced by successive addition of the NK1 and the NK2 receptor antagonists, SR140333 and SR48968, respectively. All responses were abolished in the presence of tetrodotoxin. Because ACh and substance P are the 2 main neurotransmitters mediating excitatory responses in the fundus, the effect of diabetes on the susceptibility of the smooth muscle strips to ACh and substance P was investigated by establishing dose-response curves to both agents (Figure 7C and D). In both genotypes, the doseresponse curve to ACh was shifted to the left at day 27 (GHS-R⫹/⫹: pEC50 from 6.13 ⫾ 0.05 to 6.38 ⫾ 0.09, P ⬍ .01, and GHS-R⫺/⫺: pEC50 from 6.09 ⫾ 0.03 to 6.42 ⫾ 0.05, P ⬍ .01). A significant increase of the pEC50 value for substance P was also observed in diabetic wild-type (from 6.37 ⫾ 0.10 to 7.24 ⫾ 0.20, P ⬍ .001) and mutant mice (from 6.27 ⫾ 0.11 to 7.31 ⫾ 0.23, P ⬍ .001). Neither
ROLE OF GHRELIN IN DIABETIC MICE
1273
Figure 6. Effect of STZ-induced diabetes on the neural responses of fundic strips from wild-type and mutant mice. Representative tracings of EFSinduced responses in nondiabetic and diabetic GHS-R⫹/⫹ (A) and GHS-R⫺/⫺ (B) mice are shown. The graphs summarize the change in tension of the on-relaxations (C and D) and off-contractions (E and F) in GHS-R⫹/⫹ (left) and GHS-R⫺/⫺ (right) mice. Results are represented as the means ⫾ SEM (n ⫽ 3–5 animals/group; n ⫽ 3 strips/ animal). *P ⬍ .05; **P ⬍ . 01; ***P ⬍ .001 (nondiabetic vs diabetic mice).
the genotype nor the treatment affected the maximal contraction to ACh or substance P.
Discussion In the present study, we showed that the ghrelinsignaling pathway plays an important role in the hyperphagia associated with STZ-induced diabetes. This hyperphagia is mainly driven by central pathways stimulated by the high circulating plasma ghrelin levels. In contrast, the acceleration of gastric emptying during the diabetic process occurs independently of ghrelin and is not the cause of the hyperphagia. Impaired accommodation of the stomach because of local ghrelin-independent changes in contractility may accelerate gastric emptying. In the pathogenesis of hyperphagia associated with STZ-induced diabetes, a model of insulin-deficient diabetes, reduced signaling by insulin and leptin plays a key role.26 –28 Recent studies suggest that STZ-induced diabetes in rats is characterized by hyperghrelinemia,7,29 a finding that was confirmed in the present study. Evidence for a causative role of endogenous ghrelin in hyperphagia was further provided in diabetic ghrelin knockout mice, which showed an attenuation of hyperphagia.8 Likewise, treatment of diabetic rodents with a ghrelin receptor antagonist was able to partially inhibit diabetic hyperphagia.7,8
In diabetic ghrelin knockout mice the ligand-independent signaling activity of the ghrelin receptor may still play an important role in hyperphagia. In the present study, we showed that ablation of the ghrelin receptor did not cause a further reduction of hyperphagia but rather prolonged the period of reduced hyperphagia. In the ghrelin knockout mice, the initial attenuation of hyperphagia had already normalized at day 10 after the induction of diabetes, whereas, in the ghrelin receptor knockout mice, the reduced food intake was apparent at a later stage between days 12 and 21. Thus, it appears that an increased constitutive activity of the ghrelin receptor may compensate for the loss of ghrelin in the ghrelin knockout mice and lead to an early normalization of hyperphagia in these mice. The role of hypothalamic neuropeptides in the divergence of hyperphagic responses between GHS-R⫹/⫹ and GHS-R⫺/⫺ mice was further investigated. In accordance with earlier studies,30 –32 we observed an increased mRNA expression of NPY and AgRP and a reduced expression of POMC after the induction of STZ diabetes in wild-type mice. However, in diabetic GHS-R⫺/⫺ mice, the timedependent increase in the expression of NPY and AgRP mRNA was less pronounced. These changes coincided with the divergence in food intake between both genotypes. Because the AgRP/NPY neurons express the GHSR33 and are the primary targets of ghrelin’s orexigenic
BASIC– ALIMENTARY TRACT
October 2008
1274
VERHULST ET AL
GASTROENTEROLOGY Vol. 135, No. 4
BASIC– ALIMENTARY TRACT
Figure 7. Pharmacologic characterization of neural responses of nondiabetic wild-type mice (A and B). (A) Representative tracings obtained in the presence of L-NAME (3 ⫻ 10⫺4 mol/L), and after cumulative addition of atropine (5 ⫻ 10⫺6 mol/L), SR140333 (5 ⫻ 10⫺7 mol/L), SR48968 (5 ⫻ 10⫺7 mol/L), and tetrodotoxin (TTX) (3 ⫻ 10⫺4 mol/L), respectively. (B) Tension of the off-contractions obtained with the various pharmacologic treatments. Results are mean ⫾ SEM (n ⫽ 3 animals/ group; n⫽ 3 strips/animal). *P ⬍ .05; ** P ⬍ .001. Effect of diabetes on the contractile response to ACh and substance P in fundic strips (C and D). Dose-response curves to ACh (C) and substance P (D) in strips from nondiabetic (open symbols) and diabetic (solid symbols) GHS-R⫹/⫹ and GHSR⫺/⫺ mice. Results are represented as the means ⫾ SEM (n ⫽ 3–5 animals/ group; n ⫽ 3 strips/animal).
actions in the hypothalamus,12,13 this observation implies a role for the NPY/AgRP neuron as a downstream effector of ghrelin in diabetic hyperphagia. The time-dependent changes in hypothalamic ghrelin and GHS-R mRNA expression were bell-shaped. The decrease of GHS-R mRNA expression at the end of the observation period in diabetic wild-type mice may point to a negative feedback mechanism initiated by the sustained increase in plasma ghrelin levels and by the higher mRNA expression of AgRP and NPY to prevent overeating. Nevertheless, it is not sufficient to prevent hyperphagia. During the ingestion of food, the brain also receives important information from mechanoreceptors, quantitating stretch and chemoreceptors, activated by nutrients present in the gastrointestinal tract. In addition, the wellcontrolled process of gastric emptying also plays a crucial role in the regulation of satiety. Rapid emptying would reduce the negative feedback satiety signal produced by the presence of nutrients inside the stomach and, thus, precipitate a feeling of hunger. However, a recent study showed an increase of postprandial symptoms and decreased caloric intake following acceleration of gastric emptying.34 The rapid downloading of nutrients in the duodenum may release intestinal peptides such as CCK, PYY, and GLP-1 that may enhance satiety signaling. In the present study, we found a good correlation between gastric emptying and hyperphagia in both genotypes. Nevertheless, our data in-
dicate that the accelerated gastric emptying was not the cause of hyperphagia because it occurred in a later phase of the diabetic process. In human, diabetes is often accompanied by delayed gastric emptying. A dysfunction of the vagal nerve resulting from autonomic neuropathy is believed to be the major cause.35 However, accelerated gastric emptying can be seen in subgroups of patients in the early stage of type 136 or type 2 diabetes.37,38 Likewise, there is evidence for an accelerated gastric emptying in several rodent models of insulin-dependent21,39,40 and noninsulin-dependent diabetes mellitus,41,42 but only few studies investigated the mechanisms involved. Our study showed that the increased plasma ghrelin levels do not contribute to the accelerated gastric emptying in diabetic mice because no difference in gastric emptying rate could be established between GHS-R⫹/⫹ and GHS-R⫺/⫺ mice. It is also unlikely that acute hyperglycemia contributed to accelerated gastric emptying because earlier studies showed that high glucose induces a reduction of gastric motility43,44 and markedly delays gastric emptying.45 Other studies in STZ-induced diabetic rats46 and patients with type 1 diabetes47 also report a slowing effect of acute hyperglycemia on gastric emptying. Gastric emptying results from the coordinated activity of the proximal stomach (fundus), antrum, pylorus, and duodenum. Loss of fundic relaxation is expected to accelerate
gastric emptying, whereas loss of pyloric or duodenal relaxation may delay gastric emptying. We focused on the regional alterations in fundic smooth muscle contractility in control and diabetic mice of both genotypes. In both genotypes, induction of diabetes abolished or reduced nitrergic inhibitory responses in fundic strips. It has been well established that diabetes is associated with low NO levels because of a decreased neuronal NO synthase (nNOS) expression in the gastric myenteric plexus.48 –50 The molecular mechanisms regulating changes in expression of nNOS are unclear. Insulin-induced reversal of nNOS loss in diabetic mice suggests that insulin or insulin-like growth factors may regulate the nNOS expression in these systems.50 Alternatively, continuous high glucose levels may directly or indirectly influence nNOS expression, especially because the enteric nervous system contains glucoseresponsive neurons.51 Our results therefore suggest that gastric emptying is accelerated because of a major loss of nitrergic inhibitory responses in the fundus. In addition, in both genotypes, the excitatory rebound responses were markedly enhanced compared with controls. The dose-response curves to ACh and substance P were shifted to the left in fundic smooth muscle strips, pointing to an increased affinity of the muscarinic and tachykinergic receptors for their substrates. An increased sensitivity to ACh was also reported by Kamata et al52 in longitudinal smooth muscle strips of the fundus of STZinduced diabetic rats. Because the changes in contractility were similar between GHS-R⫹/⫹ and GHS-R⫺/⫺ mice, these data confirm our finding that the accelerated gastric emptying in diabetic mice occurs independently from the ghrelin-signaling pathway. Reduced relaxations and increased rebound contractions in the diabetic mice may lead to impaired fundic accommodation and accelerated gastric emptying. Because impaired gastric accommodation normally results in increased satiety,53 it cannot play an important role in diabetic hyperphagia. This reinforces our hypothesis that changes in the central mechanisms associated with diabetic hyperphagia overrule the symptoms induced by peripheral changes in gastric contractility. In conclusion, diabetic hyperphagia is predominantly regulated by central mechanisms in which peripheral ghrelin plays a predominant role by affecting the expression of NPY and AgRP in the hypothalamus. Accelerated gastric emptying, which occurs independently of ghrelin, is not the cause of increased food intake and involves local changes in the enteric nervous system and in smooth muscle contractility. References 1. Howard AD, Feighner SD, Cully DF, et al. A receptor in pituitary and hypothalamus that functions in growth hormone release. Science 1996;273:974 –977. 2. Kojima M, Hosoda H, Date Y, et al. Ghrelin is a growthhormone-releasing acylated peptide from stomach. Nature 1999;402:656 – 660.
ROLE OF GHRELIN IN DIABETIC MICE
1275
3. Tschop M, Smiley DL, Heiman ML. Ghrelin induces adiposity in rodents. Nature 2000;407:908 –913. 4. Nakazato M, Murakami N, Date Y, et al. A role for ghrelin in the central regulation of feeding. Nature 2001;409:194 –198. 5. Wren AM, Seal LJ, Cohen MA, et al. Ghrelin enhances appetite and increases food intake in humans. J Clin Endocrinol Metab 2001;86:5992. 6. Cummings DE, Purnell JQ, Frayo RS, et al. A preprandial rise in plasma ghrelin levels suggests a role in meal initiation in humans. Diabetes 2001;50:1714 –1719. 7. Ishii S, Kamegai J, Tamura H, et al. Role of ghrelin in streptozotocin-induced diabetic hyperphagia. Endocrinology 2002;143: 4934 – 4937. 8. Dong J, Peeters TL, De Smet B, et al. Role of endogenous ghrelin in the hyperphagia of mice with streptozotocin-induced diabetes. Endocrinology 2006;147:2634 –2642. 9. Murdolo G, Lucidi P, Di Loreto C, et al. Insulin is required for prandial ghrelin suppression in humans. Diabetes 2003;52: 2923–2927. 10. Holst B, Cygankiewicz A, Jensen TH, et al. High constitutive signaling of the ghrelin receptor—identification of a potent inverse agonist. Mol Endocrinol 2003;17:2201–2210. 11. Pantel J, Legendre M, Cabrol S, et al. Loss of constitutive activity of the growth hormone secretagogue receptor in familial short stature. J Clin Invest 2006;116:760 –768. 12. Seoane LM, Lopez M, Tovar S, et al. Agouti-related peptide, neuropeptide Y, and somatostatin-producing neurons are targets for ghrelin actions in the rat hypothalamus. Endocrinology 2003; 144:544 –551. 13. Chen HY, Trumbauer ME, Chen AS, et al. Orexigenic action of peripheral ghrelin is mediated by neuropeptide Y and agoutirelated protein. Endocrinology 2004;145:2607–2612. 14. Date Y, Murakami N, Toshinai K, et al. The role of the gastric afferent vagal nerve in ghrelin-induced feeding and growth hormone secretion in rats. Gastroenterology 2002;123: 1120 –1128. 15. Fujino K, Inui A, Asakawa A, et al. Ghrelin induces fasted motor activity of the gastrointestinal tract in conscious fed rats. J Physiol 2003;550:227–240. 16. Tack J, Depoortere I, Bisschops R, et al. Influence of ghrelin on interdigestive gastrointestinal motility in humans. Gut 2006;55: 327–333. 17. Masuda Y, Tanaka T, Inomata N, et al. Ghrelin stimulates gastric acid secretion and motility in rats. Biochem Biophys Res Commun 2000;276:905–908. 18. Tack J, Depoortere I, Bisschops R, et al. Influence of ghrelin on gastric emptying and meal-related symptoms in idiopathic gastroparesis. Aliment Pharmacol Ther 2005;22:847– 853. 19. Kitazawa T, De Smet B, Verbeke K, et al. Gastric motor effects of peptide and non-peptide ghrelin agonists in mice in vivo and in vitro. Gut 2005;54:1078 –1084. 20. Depoortere I, De Winter B, Thijs T, et al. Comparison of the gastroprokinetic effects of ghrelin, GHRP-6 and motilin in rats in vivo and in vitro. Eur J Pharmacol 2005;515:160 –168. 21. Granneman JG, Stricker EM. Food intake and gastric emptying in rats with streptozotocin-induced diabetes. Am J Physiol 1984; 247:R1054 –R1061. 22. Miyamoto Y, Yoneda M, Morikawa A, et al. Gastric neuropeptides and gastric motor abnormality in streptozotocin-induced diabetic rats: observation for four weeks after streptozotocin. Dig Dis Sci 2001;46:1596 –1603. 23. Wattler S, Kelly M, Nehls M. Construction of gene targeting vectors from lambda KOS genomic libraries. Biotechniques 1999;26:1150 –1160. 24. O’Gorman S, Dagenais NA, Qian M, et al. Protamine-Cre recombinase transgenes efficiently recombine target sequences in the
BASIC– ALIMENTARY TRACT
October 2008
1276
25.
26.
27.
28.
29.
30.
31.
32. 33.
BASIC– ALIMENTARY TRACT
34.
35.
36.
37.
38.
39.
40.
VERHULST ET AL
male germ line of mice, but not in embryonic stem cells. Proc Natl Acad Sci U S A 1997;94:14602–14607. De Smet B, Thijs T, Peeters TL, et al. Effect of peripheral obestatin on gastric emptying and intestinal contractility in rodents. Neurogastroenterol Motil 2007;19:211–217. Booth DA. Some characteristics of feeding during streptozotocininduced diabetes in the rat. J Comp Physiol Psychol 1972;80: 238 –249. Sipols AJ, Baskin DG, Schwartz MW. Effect of intracerebroventricular insulin infusion on diabetic hyperphagia and hypothalamic neuropeptide gene expression. Diabetes 1995;44:147– 151. Sindelar DK, Havel PJ, Seeley RJ, et al. Low plasma leptin levels contribute to diabetic hyperphagia in rats. Diabetes 1999;48: 1275–1280. Masaoka T, Suzuki H, Hosoda H, et al. Enhanced plasma ghrelin levels in rats with streptozotocin-induced diabetes. FEBS Lett 2003;541:64 – 68. Sindelar DK, Mystkowski P, Marsh DJ, et al. Attenuation of diabetic hyperphagia in neuropeptide Y-deficient mice. Diabetes 2002;51:778 –783. Havel PJ, Hahn TM, Sindelar DK, et al. Effects of streptozotocininduced diabetes and insulin treatment on the hypothalamic melanocortin system and muscle uncoupling protein 3 expression in rats. Diabetes 2000;49:244 –252. Qu SY, Yang YK, Li JY, et al. Agouti-related protein is a mediator of diabetic hyperphagia. Regul Pept 2001;98:69 –75. Willesen MG, Kristensen P, Romer J. Co-localization of growth hormone secretagogue receptor and NPY mRNA in the arcuate nucleus of the rat. Neuroendocrinology 1999;70:306 –316. Alsina ST, Bruix SA, Sunyé P, et al. Pharmacological acceleration of gastric emptying decreases caloric intake and increases postprandial symptoms in obese subjects. Gastroenterology 2008; 134:A-315. Feldman M, Corbett DB, Ramsey EJ, et al. Abnormal gastric function in longstanding, insulin-dependent diabetic patients. Gastroenterology 1979;77:12–17. Nowak TV, Johnson CP, Kalbfleisch JH, et al. Highly variable gastric emptying in patients with insulin dependent diabetes mellitus. Gut 1995;37:23–29. Phillips WT, Schwartz JG, McMahan CA. Rapid gastric emptying in patients with early non-insulin-dependent diabetes mellitus. N Engl J Med 1991;324:130 –131. Bertin E, Schneider N, Abdelli N, et al. Gastric emptying is accelerated in obese type 2 diabetic patients without autonomic neuropathy. Diabetes Metab 2001;27:357–364. Nowak TV, Roza AM, Weisbruch JP, et al. Accelerated gastric emptying in diabetic rodents: effect of insulin treatment and pancreas transplantation. J Lab Clin Med 1994;123:110 –116. Young AA, Gedulin B, Vine W, et al. Gastric emptying is accelerated in diabetic BB rats and is slowed by subcutaneous injections of amylin. Diabetologia 1995;38:642– 648.
GASTROENTEROLOGY Vol. 135, No. 4
41. Itoh H, Yoneda M, Tamori K, et al. Rapid gastric emptying and pathological changes of vagus nerve in the spontaneously diabetic Chinese hamster. Diabetes Res Clin Pract 1995;28: 89 –95. 42. Green GM, Guan D, Schwartz JG, et al. Accelerated gastric emptying of glucose in Zucker type 2 diabetic rats: role in postprandial hyperglycaemia. Diabetologia 1997;40:136 –142. 43. Sakaguchi T, Shimojo E. Inhibition of gastric motility induced by hepatic portal injections of D-glucose and its anomers. J Physiol 1984;351:573–581. 44. Raybould HE, Zittel TT. Inhibition of gastric motility induced by intestinal glucose in awake rats: role of Na(⫹)-glucose co-transporter. Neurogastroenterol Motil 1995;7:9 –14. 45. Ishiguchi T, Tada H, Nakagawa K, et al. Hyperglycemia impairs antro-pyloric coordination and delays gastric emptying in conscious rats. Auton Neurosci 2002;95:112–120. 46. Chang FY, Lee SD, Yeh GH, et al. Hyperglycaemia is responsible for the inhibited gastrointestinal transit in the early diabetic rat. Acta Physiol Scand 1995;155:457– 462. 47. Fraser RJ, Horowitz M, Maddox AF, et al. Hyperglycaemia slows gastric emptying in type 1 (insulin-dependent) diabetes mellitus. Diabetologia 1990;33:675– 680. 48. Takahashi T, Nakamura K, Itoh H, et al. Impaired expression of nitric oxide synthase in the gastric myenteric plexus of spontaneously diabetic rats. Gastroenterology 1997;113:1535–1544. 49. Wrzos HF, Cruz A, Polavarapu R, et al. Nitric oxide synthase (NOS) expression in the myenteric plexus of streptozotocin-diabetic rats. Dig Dis Sci 1997;42:2106 –2110. 50. Watkins CC, Sawa A, Jaffrey S, et al. Insulin restores neuronal nitric oxide synthase expression and function that is lost in diabetic gastropathy. J Clin Invest 2000;106:373–384. 51. Liu M, Seino S, Kirchgessner AL, et al. Identification and characterization of glucoresponsive neurons in the enteric nervous system. J Neurosci 1999;19:10305–10317. 52. Kamata K, Sakamoto A, Kasuya Y. Changes in sensitivity of the rat stomach fundus to various drugs in streptozotocin-induced diabetic rats. Jpn J Pharmacol 1988;47:99 –102. 53. Tack J, Piessevaux H, Coulie B, et al. Role of impaired gastric accommodation to a meal in functional dyspepsia. Gastroenterology 1998;115:1346 –1352. Received November 20, 2007. Accepted June 19, 2008. Address requests for reprints to: Inge Depoortere, PhD, Centre for Gastroenterological Research, Gasthuisberg O&N1, bus 701, 3000 Leuven, Belgium. e-mail: [email protected]; fax: (32) 16-345-939. Supported by grants from the Flemish Foundation for Scientific Research (contract FWO G.0144.04 and 1.5.125.05) and the Belgian Ministry of Science (contract GOA 03/11). Conflicts of interest: D. Moechars and L. Ver Donck are employees of Johnson & Johnson Pharmaceutical Research and Development. Other authors have no conflicts of interest. P.J.V. and B.D.S. equally contributed to this paper.
Role of Ghrelin in the Relationship Between Hyperphagia and Accelerated Gastric Emptying in Diabetic Mice PIETER–JAN VERHULST,* BETTY DE SMET,* INGE SAELS,* THEO THIJS,* LUC VER DONCK,‡ DIEDER MOECHARS,‡ THEO L. PEETERS,* and INGE DEPOORTERE*
Background & Aims: Ghrelin is an orexigenic peptide with gastroprokinetic effects. Mice with streptozotocin (STZ)-induced diabetes exhibit hyperphagia, altered gastric emptying, and increased plasma ghrelin levels. We investigated the causative role of ghrelin herein by comparing changes in ghrelin receptor knockout (growth hormone secretagogue receptor [GHS-R]ⴚ/ⴚ) and wild-type (GHS-Rⴙ/ⴙ) mice with STZ-induced diabetes. Methods: Gastric emptying was measured with the [13C]octanoic acid breath test. The messenger RNA (mRNA) expression of neuropeptide Y (NPY), agoutirelated peptide (AgRP), and proopiomelanocortin was quantified by real-time reverse-transcription polymerase chain reaction. Neural contractions were elicited by electrical field stimulation in fundic smooth muscle strips. Results: Diabetes increased plasma ghrelin levels to a similar extent in both genotypes. Hyperphagia was more pronounced in GHS-Rⴙ/ⴙ than in GHS-Rⴚ/ⴚ mice between days 12 and 21. Increases in NPY and AgRP mRNA expression were less pronounced in diabetic GHS-Rⴚ/ⴚ than in GHS-Rⴙ/ⴙ mice from day 15 on, whereas decreases in proopiomelanocortin mRNA levels were similar in both genotypes. Gastric emptying was accelerated to a similar extent in both genotypes, starting on day 16. In fundic smooth muscle strips of diabetic GHS-Rⴙ/ⴙ and GHS-Rⴚ/ⴚ mice, neuronal relaxations were reduced, whereas contractions were increased; this increase was related to an increased affinity of muscarinic and tachykinergic receptors. Conclusions: Diabetic hyperphagia is regulated by central mechanisms in which the ghrelin-signaling pathway affects the expression of NPY and AgRP in the hypothalamus. The acceleration of gastric emptying, which is not affected by ghrelin signaling, is not the cause of diabetic hyperphagia and probably involves local contractility changes in the fundus.
I
n 1996, the growth hormone secretagogue receptor (GHS-R) was cloned and identified as the receptor for a family of synthetic growth hormone secretagogues.1 The endogenous ligand of this receptor, ghrelin, an acylated 28-amino acid peptide, was identified 3 years later2 and, as expected, was shown to stimulate growth hormone release. Because ghrelin is mainly produced by the
stomach, an organ well positioned to detect recently ingested food, it was soon demonstrated that ghrelin plays an important role in the regulation of the energy balance. Peripheral administration of ghrelin induces weight gain by decreasing fat utilization3 and by stimulating food intake in rodents4 as well as in humans.5 A role for ghrelin in meal initiation was also confirmed by the fluctuations in plasma ghrelin levels, which increase immediately before a meal and rapidly fall after food intake, following a 24-hour pattern reciprocal to that of insulin.6 This raises the question whether insulin negatively regulates ghrelin. Indeed, increased plasma ghrelin levels were observed during insulin deficiency in rodents with streptozotocin (STZ)-induced diabetes.7,8 Also in patients with type 1 diabetes, absolute insulin deficiency prevented prandial plasma ghrelin suppression until the insulin deficiency was corrected with an intravenous insulin bolus.9 In both conditions, the hyperghrelinemia is thought to play a role in the hyperphagia associated with uncontrolled type 1 diabetes. Important evidence for this hypothesis was provided in a study with diabetic ghrelin knockout mice.8 In the absence of its endogenous ligand, the ghrelin receptor still maintains an important level of signaling, and it is unclear to what extent this contributes to the hyperphagia.10 The physiologic importance of this constitutive activity in humans was demonstrated in 2 unrelated families harboring a missense mutation in the GHS-R, which leads to a syndrome characterized not only by short stature but also by obesity.11 Ghrelin stimulates food intake through activation of the ghrelin receptor present on neuropeptide Y (NPY)/ agouti-related peptide (AgRP) neurons in the arcuate nucleus either by crossing the blood/brain barrier or via activation of vagal nerve activity.5,12–14 Ghrelin also has important effects on gastrointestinal motility that may contribute to appetite signaling. It induces strong “hunAbbreviations used in this paper: ACh, acetylcholine; AgRP, agoutirelated peptide; EFS, electrical field stimulation; GHS-R, growth hormone secretagogue receptor; NPY, neuropeptide Y; POMC, proopiomelanocortin; STZ, streptozotocin. © 2008 by the AGA Institute 0016-5085/08/$34.00 doi:10.1053/j.gastro.2008.06.044
BASIC– ALIMENTARY TRACT
*Centre for Gastroenterological Research, Catholic University of Leuven, Leuven, Belgium; and ‡Johnson & Johnson Division of Pharmaceutical Research and Development, Janssen Pharmaceutica, Beerse, Belgium
1268
VERHULST ET AL
BASIC– ALIMENTARY TRACT
ger” contractions in the fasted state originating in the stomach and migrating distally15,16 and accelerates gastric emptying17–20 in rodents and man. Until now, it is unclear to what extent the prokinetic effects of ghrelin contribute to its effects on food intake. The aim of our study was therefore multiple. First, we determined the functional relevance of the ghrelinsignaling pathway in diabetic hyperphagia by eliminating the constitutive activity of the ghrelin receptor. To this end, we compared food intake in STZ-induced diabetic ghrelin receptor knockout (GHS-R⫺/⫺) and wild-type (GHS-R⫹/⫹) mice. Second, the contribution of a central pathway in the ghrelin-dependent regulation of hyperphagia was determined by comparing alterations in the expression of orexigenic and anorexigenic neuropeptides in the hypothalamus of both genotypes. Third, we investigated whether altered gastric emptying, known to occur in rats with STZ-induced diabetes,21,22 may contribute to hyperphagia, and we determined the role of ghrelin herein. Finally, we studied the effect of diabetes on local alterations in the in vitro contractility of smooth muscle strips of the fundus to elucidate whether peripheral mechanisms mediate the changes in gastric emptying. This multiparametric analysis should help us reveal the role of the ghrelin-signaling pathway in hyperphagia of mice with uncontrolled diabetes. It should also clarify whether central or peripheral pathways are involved.
GASTROENTEROLOGY Vol. 135, No. 4
Materials and Methods Animals Male (40 –50 weeks of age) GHS-R⫹/⫹ and GHSmice were housed in a temperature-controlled environment (20°C–22°C) under a 14-hour:10-hour lightdark cycle and had ad libitum access to food and drinking water. This research was approved by the Ethical Committee for Animal Experiments of the Catholic University of Leuven.
R⫺/⫺
Generation of GHS-Rⴚ/ⴚ Mice GHS-R⫺/⫺ mice were developed by Janssen Pharmaceutica (Beerse, Belgium) in collaboration with Lexicon Genetics, Inc (The Woodlands, TX). With a polymerase chain reaction (PCR) probe, genomic clones were isolated by screening of the 129SvEvBrd-derived lambda pKOS genomic library.23 A 9.5-kilobase genomic clone spanning exon 1 and exon 2 was used to generate the targeting vector via yeast-mediated homologous recombination. In this vector, a 1325-base pair genomic fragment, spanning exon 1 and exon 2, was replaced by a floxed version of exon 1 and exon 2 including a 1.7kilobase PGK-neo selection cassette flanked by 2 Frt sites (Figure 1). The NotI-linearized vector was electroporated into 129 Sv/Evbrd (LEX1) embryonic stem (ES) cells, and G418-fialuridine (FIAU)-resistant ES cell clones were isolated and analyzed for homologous recombination by
Figure 1. Targeted disruption of the GHS-R1 gene. (A) Structure of the targeting vector, wild-type locus, targeted locus, and Cre-excised locus. Boxes representing the exons in open and shaded boxes are the noncoding and coding regions, respectively. Arrows indicate the position of the PCR primers used for genotyping the wildtype and targeted allele. (B) Expression of the GHS-R1 transcript in brain and pituitary of both genotypes. Results are means ⫾ SEM (n ⫽ 6) relative expression levels after normalization to -actin.
October 2008
ROLE OF GHRELIN IN DIABETIC MICE
Experimental Design GHS-R⫹/⫹ (n ⫽ 26) and GHS-R⫺/⫺ mice (n ⫽ 22) were fasted overnight (14 hours) and injected intraperitoneally with 160 mg/kg STZ dissolved in 5 mmol/L sodium citrate buffer (pH 4.0). Body weight, 24-hour food intake, and tail-blood glucose levels (Accu-Chek Sensor Comfort; Roche Diagnostics, Mannheim, Germany) of 14 GHS-R⫹/⫹ and 11 GHS-R⫺/⫺ were monitored daily at 10 AM from 3 days before the injection of STZ until 27 days after. Mice were subjected to a 13C gastric emptying breath test before induction of STZ diabetes and 6, 10, 16, and 22 days thereafter. At day 27, mice (6 GHS-R⫹/⫹, 6 GHS-R⫺/⫺) were killed at 10 AM. Hypothalami were snap frozen in liquid nitrogen for real-time PCR
Table 1. Forward and Reverse Primer Sequences Gene
Primer sequences
A. Wild-type allele
5=-TgggggTgCgAACATTAgC-3= (forward) 5=-CTgAAggCATCTTTCACTACg-3=(reverse) 5=-ACATATTCTATgTgAggCACC-3=(forward) 5=-CTgAAggCATCTTTCACTACg-3=(reverse) 5=-CCgCCTCTggCAgTATCg-3= (forward) 5=-gCTgACAAACTggAAgAgTTTgC-3=(reverse) 5=-CCCTggAACTTCggCgACCTgC-3= [5=]FAM [3=]TAMRA (probe) 5=-CATCTTggCCTCACTgTCCAC-3=(forward) 5=-gggCCggACTCATCgTACT-3=(reverse) 5=-TgCTTgCTgATCCACATCTgCTggA-3= [5=]FAM [3=]TAMRA (probe) 5=-gCggAggTgCTAgATCCA-3= (forward) 5=-AggACTCgTgCAgCCTTA-3= (reverse) 5=-CCgCTCTgCgACACTACAT-3= (forward) 5=-TgTCTCAgggCTggATCTCT-3= (reverse) 5=-ACCTCACCACggAgAgCA-3= (forward) 5=-gCgAgAggTCgAgTTTgC-3= (reverse) 5=-TCAgggACCAgAACCACAAA-3= (forward) 5=-CCAgCAgAggATgAAAgCAA-3= (reverse) 5=-CCAgAggACAgAggACAAgC-3= (forward) 5=-ACATCgAAgggAgCATTgAA-3= (reverse) 5=-CCCCAATgTgTCCgTCgTg-3= (forward) 5=-gCCTgCTTCACCACCTTCT-3= (reverse)
Knockout allele B. GHS-R
-actin
C. AgRP NPY POMC GHS-R Ghrelin GAPDH
analysis, fat pads (epididymal, inguinal, and retroperitoneal fat) were collected and weighed, blood was collected by cardiac puncture for ghrelin measurements, and fundus was dissected for in vitro contractility measurements. Another group of diabetic GHS-R⫹/⫹ (n ⫽ 12) and GHSR⫺/⫺ (n ⫽ 11) mice was not used in the gastric emptying study but was killed for real-time PCR analysis and ghrelin measurements at day 7 (5 GHS-R⫹/⫹, 5 GHS-R⫺/⫺) and day 15 (7 GHS-R⫹/⫹, 6 GHS-R⫺/⫺). A group of age-matched nondiabetic ad libitum fed mice (7 GHSR⫹/⫹, 5 GHS-R⫺/⫺) was used as a control group for real-time PCR analysis, plasma ghrelin measurements, and contractility studies.
Radioimmunoassay for Plasma Ghrelin Levels Acidified plasma samples from ad libitum fed mice (10 AM) were extracted on a Sep-Pak C18 cartridge (Waters Corporation, Milford, MA). The radioimmunoassay was performed with [125I] rabbit ghrelin as tracer (Bachem, Torrance, CA) and with a rabbit antibody raised against human ghrelin[14 –28] (Eurogentec, Seraing, Belgium) (final dilution 1:8000), which recognizes both octanoylated and desoctanoylated ghrelin. Intraassay coefficient of variation of the radioimmunoassay was 6.4% and 7.7%, respectively. The minimal detectable dose was 15.6 pg/mL.
Quantitative Real-Time PCR Total RNA was extracted from the hypothalamus by means of the TRIzol reagent and reverse transcribed to complementary DNA (cDNA) with Superscript II Reverse Transcriptase. The quantitative real-time PCR reaction was run on a Lightcycler 480 system (Roche Diagnostics, Mannheim, Germany) with LightCycler 480 SYBR Green I Master mix. Primer sequences are shown in Table 1, section C. An interrun calibrator was used, and a standard curve was created for each gene to obtain PCR efficiencies. Relative expression levels of all samples were calculated with the LightCycler 480 software and were expressed relative to GAPDH and corrected for interrun variability.
Breath Test for Gastric Emptying Gastric emptying was measured during the light phase (10 AM) with a noninvasive [13C]octanoic acid breath test in fasted (19 hours) mice according to the method of Kitazawa et al,19 except that the amount of 13CO in the exhaled air was determined by an Infrared 2 Isotope Analyser (IRIS, Wagner, Germany).
In Vitro Contractility Studies With Smooth Muscle Strips From the Fundus The fundus from control (5 GHS-R⫹/⫹, 3 GHSand diabetic (4 GHS-R⫹/⫹, 4 GHS-R⫺/⫺) mice was freed from mucosa, and strips were cut and suspended along their circular axis in a tissue bath filled with Krebs
R⫺/⫺)
BASIC– ALIMENTARY TRACT
Southern blot analysis. Targeted ES cell clones were injected into C57BL/6 (albino) blastocysts, and the resulting chimeras were mated to C57BL/6 (albino) females to generate animals heterozygote for the floxed GHS-R1 allele. These were subsequently crossed with ProtamineCre mice.24 Male descendants heterozygote for both the floxed GHS-R1 allele and the Protamine Cre transgene were crossed to C57Bl/6 females to obtain heterozygote GHS-R1 knockout animals. These were subsequently crossed to generate the genotypes used in the study. PCR was used to screen genotypes by using DNA isolated from mouse tail biopsy samples. Real-time reversetranscription (RT) PCR analysis was used to show expression or absence of the GHS-R1 transcript. Primers sequences are summarized in Table 1, sections A and B, respectively.
1269
1270
VERHULST ET AL
GASTROENTEROLOGY Vol. 135, No. 4
solution as described by De Smet et al.25 The strips’ response to increasing concentrations of ACh (10⫺4 to 10⫺9 mol/L) and substance P (10⫺5 to 10⫺9 mol/L) was measured isometrically. The negative logarithm of the median effective concentration (pEC50) value and the maximal contraction were calculated with the GraphPad Prism 4.00 software (San Diego, CA). Neural responses were elicited by electrical field stimulation (EFS) of strips as described previously.25 Responses were characterized pharmacologically by repeating the frequency spectrum in the presence of L-NAME (3 ⫻ 10⫺4 mol/L) in combination with atropine, the NK1-antagonist SR140333 (5 ⫻ 10⫺7 mol/L), the NK2antagonist SR48968 (5 ⫻ 10⫺7 mol/L), and tetrodotoxin (3 ⫻ 10⫺4 mol/L), respectively. Strips were pretreated with each antagonist for 25 minutes before EFS was applied. The on-relaxations were expressed relative to the response induced by 10⫺5 mol/L nitroglycerin. The offcontractions were expressed in grams/square millimeters.
Statistical Analysis
BASIC– ALIMENTARY TRACT
Results are presented as means ⫾ SEM. Data (body weight, glucose levels, food intake, gastric emptying, and in vitro contractility) obtained from repeated measurements performed at different time points in the same mice were analyzed using the Proc Mixed procedure with slice option from SAS (SAS Institute Inc, Cary, NC). All other data were analyzed with a factorial analysis of variance followed by Newmann–Keuls as a post hoc test (Statistica 6.0; StatSoft, Tulsa, OK). Significance was accepted at the 5% level.
Results Generation of the GHS-R Knockout Mice The strategy applied resulted in the deletion of the first 2 exons that encode GHS-R1 (ENSEMBL: ENSMUSG00000051136) (Figure 1A). Correct targeting in ES cells was confirmed by Southern analysis, and PCR analysis demonstrated the ablation of the wildtype GHS-R1 allele (results not shown). Loss of expression of the GHS-R1 transcript in the knockout mice was confirmed by quantitative real-time PCR. The GHS-R1 transcript was absent in the brain and pituitary derived from the homozygote GHS-R⫺/⫺ mice (Figure 1B).
Effect of STZ Diabetes on Blood Glucose Levels, Food Intake, and Body Weight Both genotypes displayed a significant (P ⬍ .001) increase in blood glucose levels within 24 hours after the injection of STZ (Figure 2A). Blood glucose levels were maximal from day 3 onward and did not differ significantly between both genotypes during the entire observation period. In both genotypes, body weight loss was apparent from day 2 after the induction of diabetes and
Figure 2. Changes in blood glucose levels (A), body weight (B), and daily (24 hours) food intake (C) in STZ-induced diabetic GHS-R⫹/⫹ and GHS-R⫺/⫺ mice as a function of time. Results are represented as the means ⫾ SEM (n ⫽ 11–14 animals/group). *P ⬍ .05; **P ⬍ .01; ***P ⬍ .001 (GHS-R⫹/⫹ vs GHS-R⫺/⫺ mice).
did not differ between GHS-R⫹/⫹ mice (⫺24.7% ⫾ 1.8%) and GHS-R⫺/⫺ mice (⫺25.6% ⫾ 3.8%) during the course of the experiment (Figure 2B). Total fat pad mass was similar in GHS-R⫹/⫹ and GHS⫺/⫺ R mice at the start of the experiment and amounted to 4.4% ⫾ 0.7% and to 4.3% ⫾ 0.6% of body weight, respectively. At day 27 after the induction of diabetes, no epididymal, inguinal, or retroperitoneal fat was detectable in either genotype (data not shown). Daily 24-hour food intake did not differ between both genotypes before the start of the experiment. After the induction of diabetes, food intake was significantly (P ⬍ .05) increased at day 5 in the wild-type mice. Mutant mice also became hyperphagic, but this did not reach statistical significance until day 9 (P ⬍ .01). This hyperphagic state remained during the further course of the experi-
October 2008
ment, but hyperphagia was significantly less (P ⬍ .01) in the GHS-R⫺/⫺ than in the GHS-R⫹/⫹ mice between days 12 and 21 (Figure 2C).
Effect of STZ Diabetes on Plasma Ghrelin Levels, Ghrelin Messenger RNA, and GHS-R Messenger RNA Expression in the Hypothalamus
1271
R⫺/⫺ mice was observed. As expected, GHS-R mRNA expression was absent in the mutant mice.
Effect of STZ Diabetes on NPY, AgRP, and Proopiomelanocortin mRNA Expression in the Hypothalamus In the wild-type mice, induction of diabetes led to a gradual and significant increase in NPY and AgRP mRNA expression from day 15 onward (P ⬍ .01) and a significant decrease in proopiomelanocortin (POMC) mRNA expression, which was already maximal from day 7 onward (P ⬍ .001) (Figure 3D–F). At day 27, a 35-fold increase in AgRP mRNA, a 9-fold increase in NPY mRNA, and a 7-fold decrease in POMC mRNA expression were observed. The increase in NPY and AgRP mRNA expression started to diverge between diabetic GHS-R⫹/⫹ and GHS-R⫺/⫺ mice from day 15 onward. No difference in POMC expression was observed between both genotypes.
Effect of STZ Diabetes on Gastric Emptying No difference in gastric half excretion time (Thalf) was observed between wild-type (Thalf: 109 ⫾ 5 minutes) and mutant mice (Thalf: 110 ⫾ 7 minutes) before the onset of diabetes. Gastric emptying of the solid meal was markedly accelerated in both genotypes at day 16 (Thalf: GHS-R⫹/⫹: BASIC– ALIMENTARY TRACT
Before the induction of diabetes, plasma ghrelin levels did not differ between both genotypes, neither in the fed (GHS-R⫹/⫹: 528, GHS-R⫺/⫺: 435 pg/mL) nor in the fasted state (GHS-R⫹/⫹: 1221, GHS-R⫺/⫺: 1193 pg/ mL). In diabetic GHS-R⫹/⫹ and GHS-R⫺/⫺ mice, plasma ghrelin levels were significantly (P ⬍ .05) increased at day 7 after the induction of diabetes and remained elevated during the further course of the experiment (Figure 3A). Changes in plasma ghrelin levels did not differ significantly between both genotypes. In the hypothalamus of diabetic GHS-R⫹/⫹ mice, the time course of the changes in ghrelin and GHS-R messenger RNA (mRNA) expression was bell-shaped with maximal increases (P ⬍ .01) at day 15 (Figure 3B and C). At day 27, ghrelin mRNA expression was normalized, whereas GHS-R mRNA expression was decreased (P ⬍ .01) compared with day 0. No difference in hypothalamic ghrelin mRNA expression between GHS-R⫹/⫹ and GHS-
ROLE OF GHRELIN IN DIABETIC MICE
Figure 3. Time-dependent effects of diabetes on plasma ghrelin levels (A) and hypothalamic mRNA expression of ghrelin (B); GHS-R (C); and the neuropeptides AgRP (D), NPY (E), and POMC (F) in ad libitum fed (10 AM) GHS-R⫹/⫹ and GHS-R⫺/⫺ mice before and 7, 15, and 27 days after the induction of diabetes. Results are represented as the means ⫾ SEM (n ⫽ 3–7 animals/group). *P ⬍ .05; **P ⬍ .01; ***P ⬍ .001 (GHS-R⫹/⫹ vs GHSR⫺/⫺ mice).
1272
VERHULST ET AL
GASTROENTEROLOGY Vol. 135, No. 4
Figure 4. Time-dependent changes in gastric half-excretion time (Thalf) (A) and lag time (Tlag) (B) as determined by the [13C]octanoic acid breath test in diabetic GHS-R⫹/⫹ and GHS-R⫺/⫺ mice. Results are means ⫾ SEM (n ⫽ 11–14 animals/group). †P ⬍ .01 (GHS-R⫹/⫹, compared with day 0); ‡P ⬍ .05; ‡‡P ⬍ .01; ‡‡‡P ⬍ .001 (GHS-R⫺/⫺, compared with day 0).
56 ⫾ 10 minutes, P ⬍ .01, and GHS-R⫺/⫺: 55 ⫾ 9 minutes, P ⬍ .001) and day 22 (GHS-R⫹/⫹: 56 ⫾ 18 minutes, P ⬍ .01, and GHS-R⫺/⫺: 48 ⫾ 5 minutes, P ⬍ .001) after induction of diabetes (Figure 4A). Similar effects were observed on lag time (Figure 4B). Gastric emptying correlated with food intake in diabetic wild-type (r ⫽ 0.48, P ⬍ .01) and mutant mice (r ⫽ 0.65, P ⬍ .01).
Effect of STZ Diabetes on the In Vitro Contractile Responses in the Mouse Fundus Neuronal responses under nondiabetic conditions. The responses of fundic smooth muscle strips, BASIC– ALIMENTARY TRACT
subjected to EFS, were frequency dependent and consisted mainly of on-relaxations. Off-contractions were observed at higher frequencies (8 –32 Hz). EFS-induced on-relaxations were more pronounced (P ⬍ .05) between 2 and 16 Hz in GHS-R⫺/⫺ than in GHS-R⫹/⫹ mice (Figure 5A). Within the same frequency range, EFS-induced offcontractions were significantly (P ⬍ .05) increased in GHS-R⫺/⫺ compared with GHS-R⫹/⫹ mice (Figure 5B). Effect of STZ diabetes on neuronal responses. A representative tracing of the neuronal response of fundic strips from GHS-R⫹/⫹ and GHS-R⫺/⫺ mice 27 days after the induction of diabetes is shown in Figure 6A and B. Changes in tension are summarized in the same Figure (Figure 6C–F). In both genotypes, inhibitory responses were reduced at low frequencies and abolished at high (16 –32 Hz) frequencies (Figure 6C and D). In addition, the poststimulus off-contractions were strongly enhanced after the induction of diabetes at all investigated
Figure 5. Electrical field stimulation of strips from nondiabetic GHS-R⫹/⫹ and GHS-R⫺/⫺ mice resulted in onrelaxations (A) and off-contractions (B). Results are means ⫾ SEM (n ⫽ 3 animals/group; n ⫽ 3 strips/animal). *P ⬍ .05 (GHS-R⫹/⫹ vs GHS-R⫺/⫺ mice). On-relaxations are expressed as a percentage of the maximal relaxation obtained after stimulation with 10⫺5 mol/L nitroglycerin; offcontractions are expressed in grams/square millimeters.
frequencies in wild-type (P ⬍ .001) as well as in mutant (P ⬍ .05) mice (Figure 6E and F). No significant differences in the contractile responses were observed between diabetic GHS-R⫹/⫹ and GHS-R⫺/⫺ mice. Effect of STZ diabetes on the contractile response to ACh and substance P. Neural responses of
fundic strips from nondiabetic mice were characterized pharmacologically (Figure 7A and B). Pretreatment with the NO-synthase inhibitor L-NAME reversed the on-relaxations into strong contractions over the entire frequency spectrum. These contractions were abolished by addition of atropine at low frequencies. The atropineresistant contractions were further reduced by successive addition of the NK1 and the NK2 receptor antagonists, SR140333 and SR48968, respectively. All responses were abolished in the presence of tetrodotoxin. Because ACh and substance P are the 2 main neurotransmitters mediating excitatory responses in the fundus, the effect of diabetes on the susceptibility of the smooth muscle strips to ACh and substance P was investigated by establishing dose-response curves to both agents (Figure 7C and D). In both genotypes, the doseresponse curve to ACh was shifted to the left at day 27 (GHS-R⫹/⫹: pEC50 from 6.13 ⫾ 0.05 to 6.38 ⫾ 0.09, P ⬍ .01, and GHS-R⫺/⫺: pEC50 from 6.09 ⫾ 0.03 to 6.42 ⫾ 0.05, P ⬍ .01). A significant increase of the pEC50 value for substance P was also observed in diabetic wild-type (from 6.37 ⫾ 0.10 to 7.24 ⫾ 0.20, P ⬍ .001) and mutant mice (from 6.27 ⫾ 0.11 to 7.31 ⫾ 0.23, P ⬍ .001). Neither
ROLE OF GHRELIN IN DIABETIC MICE
1273
Figure 6. Effect of STZ-induced diabetes on the neural responses of fundic strips from wild-type and mutant mice. Representative tracings of EFSinduced responses in nondiabetic and diabetic GHS-R⫹/⫹ (A) and GHS-R⫺/⫺ (B) mice are shown. The graphs summarize the change in tension of the on-relaxations (C and D) and off-contractions (E and F) in GHS-R⫹/⫹ (left) and GHS-R⫺/⫺ (right) mice. Results are represented as the means ⫾ SEM (n ⫽ 3–5 animals/group; n ⫽ 3 strips/ animal). *P ⬍ .05; **P ⬍ . 01; ***P ⬍ .001 (nondiabetic vs diabetic mice).
the genotype nor the treatment affected the maximal contraction to ACh or substance P.
Discussion In the present study, we showed that the ghrelinsignaling pathway plays an important role in the hyperphagia associated with STZ-induced diabetes. This hyperphagia is mainly driven by central pathways stimulated by the high circulating plasma ghrelin levels. In contrast, the acceleration of gastric emptying during the diabetic process occurs independently of ghrelin and is not the cause of the hyperphagia. Impaired accommodation of the stomach because of local ghrelin-independent changes in contractility may accelerate gastric emptying. In the pathogenesis of hyperphagia associated with STZ-induced diabetes, a model of insulin-deficient diabetes, reduced signaling by insulin and leptin plays a key role.26 –28 Recent studies suggest that STZ-induced diabetes in rats is characterized by hyperghrelinemia,7,29 a finding that was confirmed in the present study. Evidence for a causative role of endogenous ghrelin in hyperphagia was further provided in diabetic ghrelin knockout mice, which showed an attenuation of hyperphagia.8 Likewise, treatment of diabetic rodents with a ghrelin receptor antagonist was able to partially inhibit diabetic hyperphagia.7,8
In diabetic ghrelin knockout mice the ligand-independent signaling activity of the ghrelin receptor may still play an important role in hyperphagia. In the present study, we showed that ablation of the ghrelin receptor did not cause a further reduction of hyperphagia but rather prolonged the period of reduced hyperphagia. In the ghrelin knockout mice, the initial attenuation of hyperphagia had already normalized at day 10 after the induction of diabetes, whereas, in the ghrelin receptor knockout mice, the reduced food intake was apparent at a later stage between days 12 and 21. Thus, it appears that an increased constitutive activity of the ghrelin receptor may compensate for the loss of ghrelin in the ghrelin knockout mice and lead to an early normalization of hyperphagia in these mice. The role of hypothalamic neuropeptides in the divergence of hyperphagic responses between GHS-R⫹/⫹ and GHS-R⫺/⫺ mice was further investigated. In accordance with earlier studies,30 –32 we observed an increased mRNA expression of NPY and AgRP and a reduced expression of POMC after the induction of STZ diabetes in wild-type mice. However, in diabetic GHS-R⫺/⫺ mice, the timedependent increase in the expression of NPY and AgRP mRNA was less pronounced. These changes coincided with the divergence in food intake between both genotypes. Because the AgRP/NPY neurons express the GHSR33 and are the primary targets of ghrelin’s orexigenic
BASIC– ALIMENTARY TRACT
October 2008
1274
VERHULST ET AL
GASTROENTEROLOGY Vol. 135, No. 4
BASIC– ALIMENTARY TRACT
Figure 7. Pharmacologic characterization of neural responses of nondiabetic wild-type mice (A and B). (A) Representative tracings obtained in the presence of L-NAME (3 ⫻ 10⫺4 mol/L), and after cumulative addition of atropine (5 ⫻ 10⫺6 mol/L), SR140333 (5 ⫻ 10⫺7 mol/L), SR48968 (5 ⫻ 10⫺7 mol/L), and tetrodotoxin (TTX) (3 ⫻ 10⫺4 mol/L), respectively. (B) Tension of the off-contractions obtained with the various pharmacologic treatments. Results are mean ⫾ SEM (n ⫽ 3 animals/ group; n⫽ 3 strips/animal). *P ⬍ .05; ** P ⬍ .001. Effect of diabetes on the contractile response to ACh and substance P in fundic strips (C and D). Dose-response curves to ACh (C) and substance P (D) in strips from nondiabetic (open symbols) and diabetic (solid symbols) GHS-R⫹/⫹ and GHSR⫺/⫺ mice. Results are represented as the means ⫾ SEM (n ⫽ 3–5 animals/ group; n ⫽ 3 strips/animal).
actions in the hypothalamus,12,13 this observation implies a role for the NPY/AgRP neuron as a downstream effector of ghrelin in diabetic hyperphagia. The time-dependent changes in hypothalamic ghrelin and GHS-R mRNA expression were bell-shaped. The decrease of GHS-R mRNA expression at the end of the observation period in diabetic wild-type mice may point to a negative feedback mechanism initiated by the sustained increase in plasma ghrelin levels and by the higher mRNA expression of AgRP and NPY to prevent overeating. Nevertheless, it is not sufficient to prevent hyperphagia. During the ingestion of food, the brain also receives important information from mechanoreceptors, quantitating stretch and chemoreceptors, activated by nutrients present in the gastrointestinal tract. In addition, the wellcontrolled process of gastric emptying also plays a crucial role in the regulation of satiety. Rapid emptying would reduce the negative feedback satiety signal produced by the presence of nutrients inside the stomach and, thus, precipitate a feeling of hunger. However, a recent study showed an increase of postprandial symptoms and decreased caloric intake following acceleration of gastric emptying.34 The rapid downloading of nutrients in the duodenum may release intestinal peptides such as CCK, PYY, and GLP-1 that may enhance satiety signaling. In the present study, we found a good correlation between gastric emptying and hyperphagia in both genotypes. Nevertheless, our data in-
dicate that the accelerated gastric emptying was not the cause of hyperphagia because it occurred in a later phase of the diabetic process. In human, diabetes is often accompanied by delayed gastric emptying. A dysfunction of the vagal nerve resulting from autonomic neuropathy is believed to be the major cause.35 However, accelerated gastric emptying can be seen in subgroups of patients in the early stage of type 136 or type 2 diabetes.37,38 Likewise, there is evidence for an accelerated gastric emptying in several rodent models of insulin-dependent21,39,40 and noninsulin-dependent diabetes mellitus,41,42 but only few studies investigated the mechanisms involved. Our study showed that the increased plasma ghrelin levels do not contribute to the accelerated gastric emptying in diabetic mice because no difference in gastric emptying rate could be established between GHS-R⫹/⫹ and GHS-R⫺/⫺ mice. It is also unlikely that acute hyperglycemia contributed to accelerated gastric emptying because earlier studies showed that high glucose induces a reduction of gastric motility43,44 and markedly delays gastric emptying.45 Other studies in STZ-induced diabetic rats46 and patients with type 1 diabetes47 also report a slowing effect of acute hyperglycemia on gastric emptying. Gastric emptying results from the coordinated activity of the proximal stomach (fundus), antrum, pylorus, and duodenum. Loss of fundic relaxation is expected to accelerate
gastric emptying, whereas loss of pyloric or duodenal relaxation may delay gastric emptying. We focused on the regional alterations in fundic smooth muscle contractility in control and diabetic mice of both genotypes. In both genotypes, induction of diabetes abolished or reduced nitrergic inhibitory responses in fundic strips. It has been well established that diabetes is associated with low NO levels because of a decreased neuronal NO synthase (nNOS) expression in the gastric myenteric plexus.48 –50 The molecular mechanisms regulating changes in expression of nNOS are unclear. Insulin-induced reversal of nNOS loss in diabetic mice suggests that insulin or insulin-like growth factors may regulate the nNOS expression in these systems.50 Alternatively, continuous high glucose levels may directly or indirectly influence nNOS expression, especially because the enteric nervous system contains glucoseresponsive neurons.51 Our results therefore suggest that gastric emptying is accelerated because of a major loss of nitrergic inhibitory responses in the fundus. In addition, in both genotypes, the excitatory rebound responses were markedly enhanced compared with controls. The dose-response curves to ACh and substance P were shifted to the left in fundic smooth muscle strips, pointing to an increased affinity of the muscarinic and tachykinergic receptors for their substrates. An increased sensitivity to ACh was also reported by Kamata et al52 in longitudinal smooth muscle strips of the fundus of STZinduced diabetic rats. Because the changes in contractility were similar between GHS-R⫹/⫹ and GHS-R⫺/⫺ mice, these data confirm our finding that the accelerated gastric emptying in diabetic mice occurs independently from the ghrelin-signaling pathway. Reduced relaxations and increased rebound contractions in the diabetic mice may lead to impaired fundic accommodation and accelerated gastric emptying. Because impaired gastric accommodation normally results in increased satiety,53 it cannot play an important role in diabetic hyperphagia. This reinforces our hypothesis that changes in the central mechanisms associated with diabetic hyperphagia overrule the symptoms induced by peripheral changes in gastric contractility. In conclusion, diabetic hyperphagia is predominantly regulated by central mechanisms in which peripheral ghrelin plays a predominant role by affecting the expression of NPY and AgRP in the hypothalamus. Accelerated gastric emptying, which occurs independently of ghrelin, is not the cause of increased food intake and involves local changes in the enteric nervous system and in smooth muscle contractility. References 1. Howard AD, Feighner SD, Cully DF, et al. A receptor in pituitary and hypothalamus that functions in growth hormone release. Science 1996;273:974 –977. 2. Kojima M, Hosoda H, Date Y, et al. Ghrelin is a growthhormone-releasing acylated peptide from stomach. Nature 1999;402:656 – 660.
ROLE OF GHRELIN IN DIABETIC MICE
1275
3. Tschop M, Smiley DL, Heiman ML. Ghrelin induces adiposity in rodents. Nature 2000;407:908 –913. 4. Nakazato M, Murakami N, Date Y, et al. A role for ghrelin in the central regulation of feeding. Nature 2001;409:194 –198. 5. Wren AM, Seal LJ, Cohen MA, et al. Ghrelin enhances appetite and increases food intake in humans. J Clin Endocrinol Metab 2001;86:5992. 6. Cummings DE, Purnell JQ, Frayo RS, et al. A preprandial rise in plasma ghrelin levels suggests a role in meal initiation in humans. Diabetes 2001;50:1714 –1719. 7. Ishii S, Kamegai J, Tamura H, et al. Role of ghrelin in streptozotocin-induced diabetic hyperphagia. Endocrinology 2002;143: 4934 – 4937. 8. Dong J, Peeters TL, De Smet B, et al. Role of endogenous ghrelin in the hyperphagia of mice with streptozotocin-induced diabetes. Endocrinology 2006;147:2634 –2642. 9. Murdolo G, Lucidi P, Di Loreto C, et al. Insulin is required for prandial ghrelin suppression in humans. Diabetes 2003;52: 2923–2927. 10. Holst B, Cygankiewicz A, Jensen TH, et al. High constitutive signaling of the ghrelin receptor—identification of a potent inverse agonist. Mol Endocrinol 2003;17:2201–2210. 11. Pantel J, Legendre M, Cabrol S, et al. Loss of constitutive activity of the growth hormone secretagogue receptor in familial short stature. J Clin Invest 2006;116:760 –768. 12. Seoane LM, Lopez M, Tovar S, et al. Agouti-related peptide, neuropeptide Y, and somatostatin-producing neurons are targets for ghrelin actions in the rat hypothalamus. Endocrinology 2003; 144:544 –551. 13. Chen HY, Trumbauer ME, Chen AS, et al. Orexigenic action of peripheral ghrelin is mediated by neuropeptide Y and agoutirelated protein. Endocrinology 2004;145:2607–2612. 14. Date Y, Murakami N, Toshinai K, et al. The role of the gastric afferent vagal nerve in ghrelin-induced feeding and growth hormone secretion in rats. Gastroenterology 2002;123: 1120 –1128. 15. Fujino K, Inui A, Asakawa A, et al. Ghrelin induces fasted motor activity of the gastrointestinal tract in conscious fed rats. J Physiol 2003;550:227–240. 16. Tack J, Depoortere I, Bisschops R, et al. Influence of ghrelin on interdigestive gastrointestinal motility in humans. Gut 2006;55: 327–333. 17. Masuda Y, Tanaka T, Inomata N, et al. Ghrelin stimulates gastric acid secretion and motility in rats. Biochem Biophys Res Commun 2000;276:905–908. 18. Tack J, Depoortere I, Bisschops R, et al. Influence of ghrelin on gastric emptying and meal-related symptoms in idiopathic gastroparesis. Aliment Pharmacol Ther 2005;22:847– 853. 19. Kitazawa T, De Smet B, Verbeke K, et al. Gastric motor effects of peptide and non-peptide ghrelin agonists in mice in vivo and in vitro. Gut 2005;54:1078 –1084. 20. Depoortere I, De Winter B, Thijs T, et al. Comparison of the gastroprokinetic effects of ghrelin, GHRP-6 and motilin in rats in vivo and in vitro. Eur J Pharmacol 2005;515:160 –168. 21. Granneman JG, Stricker EM. Food intake and gastric emptying in rats with streptozotocin-induced diabetes. Am J Physiol 1984; 247:R1054 –R1061. 22. Miyamoto Y, Yoneda M, Morikawa A, et al. Gastric neuropeptides and gastric motor abnormality in streptozotocin-induced diabetic rats: observation for four weeks after streptozotocin. Dig Dis Sci 2001;46:1596 –1603. 23. Wattler S, Kelly M, Nehls M. Construction of gene targeting vectors from lambda KOS genomic libraries. Biotechniques 1999;26:1150 –1160. 24. O’Gorman S, Dagenais NA, Qian M, et al. Protamine-Cre recombinase transgenes efficiently recombine target sequences in the
BASIC– ALIMENTARY TRACT
October 2008
1276
25.
26.
27.
28.
29.
30.
31.
32. 33.
BASIC– ALIMENTARY TRACT
34.
35.
36.
37.
38.
39.
40.
VERHULST ET AL
male germ line of mice, but not in embryonic stem cells. Proc Natl Acad Sci U S A 1997;94:14602–14607. De Smet B, Thijs T, Peeters TL, et al. Effect of peripheral obestatin on gastric emptying and intestinal contractility in rodents. Neurogastroenterol Motil 2007;19:211–217. Booth DA. Some characteristics of feeding during streptozotocininduced diabetes in the rat. J Comp Physiol Psychol 1972;80: 238 –249. Sipols AJ, Baskin DG, Schwartz MW. Effect of intracerebroventricular insulin infusion on diabetic hyperphagia and hypothalamic neuropeptide gene expression. Diabetes 1995;44:147– 151. Sindelar DK, Havel PJ, Seeley RJ, et al. Low plasma leptin levels contribute to diabetic hyperphagia in rats. Diabetes 1999;48: 1275–1280. Masaoka T, Suzuki H, Hosoda H, et al. Enhanced plasma ghrelin levels in rats with streptozotocin-induced diabetes. FEBS Lett 2003;541:64 – 68. Sindelar DK, Mystkowski P, Marsh DJ, et al. Attenuation of diabetic hyperphagia in neuropeptide Y-deficient mice. Diabetes 2002;51:778 –783. Havel PJ, Hahn TM, Sindelar DK, et al. Effects of streptozotocininduced diabetes and insulin treatment on the hypothalamic melanocortin system and muscle uncoupling protein 3 expression in rats. Diabetes 2000;49:244 –252. Qu SY, Yang YK, Li JY, et al. Agouti-related protein is a mediator of diabetic hyperphagia. Regul Pept 2001;98:69 –75. Willesen MG, Kristensen P, Romer J. Co-localization of growth hormone secretagogue receptor and NPY mRNA in the arcuate nucleus of the rat. Neuroendocrinology 1999;70:306 –316. Alsina ST, Bruix SA, Sunyé P, et al. Pharmacological acceleration of gastric emptying decreases caloric intake and increases postprandial symptoms in obese subjects. Gastroenterology 2008; 134:A-315. Feldman M, Corbett DB, Ramsey EJ, et al. Abnormal gastric function in longstanding, insulin-dependent diabetic patients. Gastroenterology 1979;77:12–17. Nowak TV, Johnson CP, Kalbfleisch JH, et al. Highly variable gastric emptying in patients with insulin dependent diabetes mellitus. Gut 1995;37:23–29. Phillips WT, Schwartz JG, McMahan CA. Rapid gastric emptying in patients with early non-insulin-dependent diabetes mellitus. N Engl J Med 1991;324:130 –131. Bertin E, Schneider N, Abdelli N, et al. Gastric emptying is accelerated in obese type 2 diabetic patients without autonomic neuropathy. Diabetes Metab 2001;27:357–364. Nowak TV, Roza AM, Weisbruch JP, et al. Accelerated gastric emptying in diabetic rodents: effect of insulin treatment and pancreas transplantation. J Lab Clin Med 1994;123:110 –116. Young AA, Gedulin B, Vine W, et al. Gastric emptying is accelerated in diabetic BB rats and is slowed by subcutaneous injections of amylin. Diabetologia 1995;38:642– 648.
GASTROENTEROLOGY Vol. 135, No. 4
41. Itoh H, Yoneda M, Tamori K, et al. Rapid gastric emptying and pathological changes of vagus nerve in the spontaneously diabetic Chinese hamster. Diabetes Res Clin Pract 1995;28: 89 –95. 42. Green GM, Guan D, Schwartz JG, et al. Accelerated gastric emptying of glucose in Zucker type 2 diabetic rats: role in postprandial hyperglycaemia. Diabetologia 1997;40:136 –142. 43. Sakaguchi T, Shimojo E. Inhibition of gastric motility induced by hepatic portal injections of D-glucose and its anomers. J Physiol 1984;351:573–581. 44. Raybould HE, Zittel TT. Inhibition of gastric motility induced by intestinal glucose in awake rats: role of Na(⫹)-glucose co-transporter. Neurogastroenterol Motil 1995;7:9 –14. 45. Ishiguchi T, Tada H, Nakagawa K, et al. Hyperglycemia impairs antro-pyloric coordination and delays gastric emptying in conscious rats. Auton Neurosci 2002;95:112–120. 46. Chang FY, Lee SD, Yeh GH, et al. Hyperglycaemia is responsible for the inhibited gastrointestinal transit in the early diabetic rat. Acta Physiol Scand 1995;155:457– 462. 47. Fraser RJ, Horowitz M, Maddox AF, et al. Hyperglycaemia slows gastric emptying in type 1 (insulin-dependent) diabetes mellitus. Diabetologia 1990;33:675– 680. 48. Takahashi T, Nakamura K, Itoh H, et al. Impaired expression of nitric oxide synthase in the gastric myenteric plexus of spontaneously diabetic rats. Gastroenterology 1997;113:1535–1544. 49. Wrzos HF, Cruz A, Polavarapu R, et al. Nitric oxide synthase (NOS) expression in the myenteric plexus of streptozotocin-diabetic rats. Dig Dis Sci 1997;42:2106 –2110. 50. Watkins CC, Sawa A, Jaffrey S, et al. Insulin restores neuronal nitric oxide synthase expression and function that is lost in diabetic gastropathy. J Clin Invest 2000;106:373–384. 51. Liu M, Seino S, Kirchgessner AL, et al. Identification and characterization of glucoresponsive neurons in the enteric nervous system. J Neurosci 1999;19:10305–10317. 52. Kamata K, Sakamoto A, Kasuya Y. Changes in sensitivity of the rat stomach fundus to various drugs in streptozotocin-induced diabetic rats. Jpn J Pharmacol 1988;47:99 –102. 53. Tack J, Piessevaux H, Coulie B, et al. Role of impaired gastric accommodation to a meal in functional dyspepsia. Gastroenterology 1998;115:1346 –1352. Received November 20, 2007. Accepted June 19, 2008. Address requests for reprints to: Inge Depoortere, PhD, Centre for Gastroenterological Research, Gasthuisberg O&N1, bus 701, 3000 Leuven, Belgium. e-mail: [email protected]; fax: (32) 16-345-939. Supported by grants from the Flemish Foundation for Scientific Research (contract FWO G.0144.04 and 1.5.125.05) and the Belgian Ministry of Science (contract GOA 03/11). Conflicts of interest: D. Moechars and L. Ver Donck are employees of Johnson & Johnson Pharmaceutical Research and Development. Other authors have no conflicts of interest. P.J.V. and B.D.S. equally contributed to this paper.