Ghrelin, the natural ligand of the growth hormone secretagogue receptor type 1a (GHS-R1a), is mainly secreted from the stomach and regulates food intake and energy homeostasis. p27 regulates cell cycle progression in many cell types. Here, we report that rats affected by the multiple endocrine neoplasia syndrome MENX, caused by a p27 mutation, develop pancreatic islet hyperplasia containing elevated numbers of ghrelin-producing ε-cells. The metabolic phenotype of MENX-affected rats featured high endogenous acylated and unacylated plasma ghrelin levels. Supporting increased ghrelin action, MENX rats show increased food intake, enhanced body fat mass, and elevated plasma levels of triglycerides and cholesterol. Ghrelin effect on food intake was confirmed by treating MENX rats with a GHS-R1a antagonist. At 7.5 months, MENX-affected rats show decreased mRNA levels of hypothalamic GHS-R1a, neuropeptide Y (NPY), and agouti-related protein (AgRP), suggesting that prolonged hyperghrelinemia may lead to decreased ghrelin efficacy. In line with ghrelin’s proposed role in glucose metabolism, we find decreased glucose-stimulated insulin secretion in MENX rats, while insulin sensitivity is improved. In summary, we provide a novel nontransgenic rat model with high endogenous ghrelin plasma levels and, interestingly, improved glucose tolerance. This model might aid in identifying new therapeutic approaches for obesity and obesity-related diseases, including type 2 diabetes.
Currently, increased food intake and reduced physical activity have resulted in a dramatic increase in the development of obesity-associated disorders, such as type 2 diabetes, hyperlipidemia, and cardiovascular diseases. The increased prevalence of obesity has led to a growing interest in the fundamental understanding of energy homeostasis and metabolism. In the last two decades, several clinical and experimental studies in the field of obesity and obesity-associated diseases identified ghrelin as a peripheral peptide hormone exerting orexigenic effects via the hypothalamus (1,2).
Ghrelin, the natural ligand of the growth hormone secretagogue receptor type 1a (GHS-R1a), was first discovered in 1999, and it is the only known circulating peripheral peptide hormone that stimulates food intake in humans and rodents (1,3,4). Ghrelin is a 28–amino acid peptide primarily released by X/A-like cells and P/D1 cells in the oxyntic glands of rat and human fundi and secreted to a lesser extent by the pancreas, intestine, kidneys, liver, pituitary, and hypothalamus (5–11). Pancreatic ghrelin-producing cells were described in 2002 (12) and later named ε-cells. They represent a distinct hormone-producing cell population, which is abundant in fetal and neonatal pancreata but poorly represented in adult pancreata in humans, mice, and rats (13). In mice, ghrelin has been found localized within the glucagon-positive α-cells at birth, whereas in rats this occurs only occasionally and in humans it has not been reported (13). Acylated or active ghrelin (AG) has a unique fatty acid modification on the third amino acid residue (Ser3). This chemical modification, achieved by the enzyme ghrelin O-acyltransferase, is essential for the binding of the hormone to the GHS-R1a receptor (3,14). However, the majority of circulating ghrelin lacks this acyl group and is called unacylated ghrelin (UAG).
The orexigenic properties of AG make it a promising target for the treatment of obesity and cachexia through either GHS-R1a activation or inhibition, respectively. Randomized, placebo-controlled human studies on cancer-associated cachexia have shown an increase in food intake after ghrelin infusions (15).
Aside from its function as an appetite-stimulating peptide, ghrelin exerts several other hormonal, metabolic, and cardiovascular activities (14,16). For instance, it enhances growth hormone, adrenocorticotropic hormone, prolactin, and cortisol release, and it suppresses testosterone and luteinizing hormone secretion. Ghrelin has been also shown to accelerate gastric emptying, increase gastrointestinal tract motility, increase antilipolytic effect on adipose tissue, and decrease colonic transit time, among other metabolic activities. Moreover, it may increase stroke volume, enhance systemic vasodilatation, and decrease blood pressure (14,16).
In recent years genetically engineered mouse models having loss- or gain-of-function mutations in the ghrelin axis were generated and have contributed to our understanding of the pleiotropic effects of ghrelin (17). However, background-dependent predisposition to lean or fat phenotypes needs to be taken into account while interpreting the data obtained from these animal models. Currently, there are only a few mouse models with high endogenous ghrelin levels, and studies of these animals have generated inconsistent results concerning food intake, body weight gain, and energy expenditure (17).
MENX is a multiple endocrine neoplasia (MEN) syndrome in the rat caused by a spontaneous, homozygous loss-of-function mutation in the cyclin-dependent kinase inhibitor p27. This mutation leads to a rapid degradation of the encoded mutant p27 protein (18). p27 is critical in maintaining quiescence in adult tissues by blocking cell cycle progression at the G1/S boundary. During embryonic development, p27 promotes cell cycle withdrawal of progenitor cells in various neuronal and endocrine tissues, including the anterior pituitary gland, thereby allowing cell differentiation (19–21). Notably, we observed a persistent postnatal expression of embryonic progenitor cell markers in the pituitary gland of MENX-affected rats (22), suggesting altered development and differentiation of neuroendocrine cells in these animals due to p27 deficiency.
It has been reported that germline ablation of p27 in mice causes pancreatic islet hyperplasia and glucose-induced hyperinsulinemia (23). The aims of this study were to determine whether MENX rats also develop islet hyperplasia and to assess the associated metabolic phenotypes. We show here that MENX-affected rats develop pancreatic islet hyperplasia within their first year of life. Analysis of the endocrine pancreas revealed an unexpected large number of ghrelin-producing ε-cells in the pancreatic islets. MENX rats accordingly show high endogenous levels of UAG and AG in combination with hyperphagia, dyslipidemia, and an increased white adipose tissue (WAT) mass. Interestingly, obese hyperghrelinemic MENX-affected rats show decreased glucose-stimulated insulin secretion (GSIS) but improved insulin sensitivity.
Research Design and Methods
The phenotype and genotype of MENX rats have previously been described (24). In all studies, male MENX-affected rats (homozygous mut/mut) and their wild-type (wt/wt) littermates were group housed under controlled conditions (temperature 23°C, 12-h/12-h light/dark cycle). The rats had access to standard rodent chow (Fa. Altromin) and water ad libitum. All experiments and procedures were approved by local authorities (AZ 55.2-1-54-2532-84-11) and complied with German animal protection laws. The group size in each experiment was n ≥ 6.
Immunohistochemistry was performed on an automated immunostainer (Ventana Medical Systems) according to the manufacturer’s protocols. Staining of 2-μm-thick rat gastric corpus tissue sections was performed using antibodies against ghrelin (H-031-31; Phoenix Pharmaceuticals) (1:350) and GHS-R1a (H-001-62; Phoenix Pharmaceuticals) (1:500). Positively stained cells were visualized with DAB (Vector Laboratories). Images were recorded using an Olympus camera DP25 installed in an Olympus microscope BX43 with the Olympus LabSense software.
Sections of rat pancreas were incubated in xylene (2 × 10 min) to remove the paraffin. After treatment with isopropanol (10 min), rehydration was achieved by successive transfer of the tissue through the following graded series of ethanol: 100% (two times for 5 min), 96% (5 min), and 70% (5 min). The slides were washed with TBS with Tween (TBS-T) (5 min) and subsequently boiled in citrate buffer (Dako) (30 min) for epitope retrieval. After a washing step in TBS-T (5 min), the tissue was blocked with 0.1% Sudan Black B in 70% ethanol (45 min) and washed in TBS-T (three times for 5 min). Staining was performed using primary antibodies against ghrelin (H-031-31; Phoenix Pharmaceuticals) (1:350), glucagon (A0565; Dako) (1:1,500), insulin (A0564; Dako) (1:750), pancreatic polypeptide (PP) (ab77192; Abcam) (1:150), and somatostatin (sc7819; Santa Cruz Biotechnology) (1:300). Primary antibodies were applied overnight at 4°C. After the removal of unbound antibodies, the tissue was incubated with secondary antibodies (anti-goat IgG [H+L] fluorescein isothiocyanate–conjugated [905-095-180; Jackson ImmunoResearch] [1:200], anti-rabbit IgG [H+L] fluorescein isothiocyanate–conjugated [F-2765; Invitrogen] [1:200], and anti–guinea pig IgG [H+L] Alexa Fluor 555–conjugated [A-21453; Invitrogen] [1:200]) for 1 h at room temperature. The fluorescently labeled slides were washed with TBS-T (3 times for 1 min), followed by staining of nuclei with Hoechst 33342 bisBenzimide (Sigma-Aldrich). The tissue was subsequently washed in dH2O and covered with VECTASHIELD (Vector Laboratories). Cover glasses were mounted on the tissue and fixed with enamel. Images were recorded using an Axiolmager Fluorescence Microscope (Carl Zeiss).
ELISA of Acylated Ghrelin and Total Ghrelin
Fasting blood was collected in EDTA tubes, and plasma was isolated by centrifugation. The plasma was acidified with HCl and treated with cOmplete protease inhibitor cocktail (Roche, Mannheim, Germany). Active (EZRGRA-90K; Millipore, Billerica, MA) and total (EZRGRT-91K; Millipore) ghrelin in rat plasma was measured with the indicated ELISA assays according to the manufacturer’s protocols.
RNA Isolation and Real-Time Quantitative RT-PCR
Rat tissues were snap frozen in liquid nitrogen and stored at −80°C until use. For total RNA extraction from snap-frozen rat tissues, serial 20-μm tissue sections were cut and resuspended in RLT buffer (RNeasy Mini Kit, Qiagen). Total RNA used for quantitative RT-PCR was purified by using the RNeasy Mini Kit. We synthesized the first-strand cDNA from total mRNA using random hexamers and SuperScript II (Invitrogen). Quantitative RT-PCR for rat mRNA was performed with TaqMan primers and probes specific for rat Agrp, Npy, Ghs-r1a, Lep, and Ghrl genes, as well as for β2-microglobulin as an internal control (Applied Biosystems). The assays were performed as previously reported (18).
Food Intake Experiments
Daily food consumption of mut/mut and age-matched wt/wt rats was measured on five consecutive days (8:00 a.m.) at different ages, and the mean consumption of each individual rat was calculated.
Mut/mut and wt/wt rats at 3.5 months of age were food deprived for 24 h with water available ad libitum. Experiments were started at 8:00 a.m. Cumulative food intake was measured for each individual rat at 0.5 h, 1 h, 2 h, 4 h, 8 h, and 24 h after refeeding.
GHS-R1a Antagonist Treatment
Mut/mut and wt/wt rats at 3.5 months of age were food deprived for 24 h with water available ad libitum. They were injected intraperitoneally with ghrelin antagonist [D-Lys3]-GHRP-6 (12 mg/kg of body weight) (H-3108; Bachem, Switzerland) or vehicle (saline) at 8:00 a.m. Cumulative food intake was measured for each individual rat at 0.5 h, 1 h, 2 h, 4 h, 8 h, and 24 h after injection.
Preparation of Gastric Corpus Tissue for Ghrelin ELISA
The gastric corpus of 7.5-month-old mut/mut and wt/wt rats was quickly removed from killed animals and snap frozen. Serial 10 μm sections of each tissue were obtained using a cryotome and were lysed in cold lysis solution (radioimmunoprecipitation assay buffer [R0278; Sigma-Aldrich], 10 mmol/L HCl, and cOmplete protease inhibitor cocktail). The lysates were centrifuged at 16,000 g 4°C for 15 min, and supernatants were collected and used for total ghrelin measurements with an ELISA kit (EZRGRT-91K; Millipore) and for total protein content measurement to normalize the ELISA results.
Oral and Intraperitoneal Glucose Tolerance Tests
For the intraperitoneal glucose tolerance test (i.p.GTT) and oral glucose tolerance test (oGTT), a tail-tip blood sample was obtained from fasting rats (overnight 16 h [t0]). The blood glucose concentration was immediately measured using a glucometer (Contour; Bayer Health Care) after D-glucose solution (40%) was injected intraperitoneally (i.p.GTT) or administered by oral gavage (2 g/kg [oGTT]). Additional blood samples for glucose and insulin measurements were obtained 15, 30, 60, 90, and 120 min postinjection.
Intraperitoneal Insulin Tolerance Test
For the intraperitoneal insulin tolerance test (i.p.ITT), a tail-tip blood sample was obtained from fasting rats (6 h [t0]). The blood glucose concentration was immediately measured using a glucometer (Contour) after insulin solution (Berlinsulin H; Berlin-Chemie AG) was injected (0.75 IU/kg [Fig. 8A and B] or 0.375 IU/kg [Fig. 8D] i.p.). Additional blood samples for glucose measurements were obtained 15, 30, 60, 90, and 120 min postinjection.
Data are presented as the mean ± SEM. Statistical significance between two series of data were determined by Student t test. P < 0.05 was considered statistically significant.
Enhanced Ghrelin Immunoreactivity in Hyperplastic Pancreatic Islets of MENX Rats
Previous studies in transgenic mice showed that deletion of the gene encoding p27 (Cdkn1b) leads to pancreatic β-cell hyperplasia (23,25). In the current study, we observed that 7.5-month-old MENX-affected (mut/mut) rats displayed a greater than twofold increase in islet cell mass compared with age-matched unaffected wt/wt rats (Fig. 1A and B). To determine whether this phenotype represents an early or a late event, we also studied the pancreatic islets of 14-day-old mut/mut and wt/wt rats. Already at this early age, mut/mut rats have an increased islet cell mass compared with wt/wt rats (Fig. 1B). For analysis of the cellular composition of the hyperplastic islets in mut/mut rats, glucagon-producing α-cells, insulin-producing β-cells, somatostatin-producing δ-cells, PP-producing cells, and ghrelin-producing ε-cells were examined by immunohistochemistry or immunofluorescence using antibodies against the respective hormones. The results show not only β-cell hyperplasia but also an increase in the number of α-cells, δ-cells and PP-producing cells in the endocrine pancreas of mut/mut rats compared with wt/wt littermates at both 14 days and 7.5 months (Fig. 1 C–J and Supplementary Fig. 1). This result suggests an altered development and differentiation of endocrine pancreatic cells due to defective p27 function. Further, we found abnormally high numbers of ghrelin-positive ε-cells in the islets of adult mut/mut rats (Fig. 1K). In 14-day-old mut/mut and wt/wt rats, ghrelin was not found localized with glucagon (Supplementary Fig. 1).
Previous studies in rats described large numbers of ghrelin-producing cells at birth, occasional ghrelin-positive cells at up to 1 month of age, and virtually no positive cells in the endocrine pancreas of adult animals (13,26). The overrepresentation of ghrelin-positive cells in mut/mut rat pancreata together with the fact that they are usually not found in adult wt/wt rats lead us to hypothesize that pancreatic islet-derived ghrelin may play a physiologically meaningful role in these animals.
Plasma Ghrelin Levels, Food Intake, and Body Weight in MENX Rats
Based on the observation that mut/mut rats contain high numbers of ghrelin-positive ε-cells in their islets, we assessed possible differences in fasting plasma AG and UAG levels between mut/mut and wt/wt rats at various ages (2.5, 5.5, and 7.5 months). We found that at 5.5 months of age, the mut/mut rats had increased ghrelin levels compared with the age-matched wt/wt rats (AG P = 0.023; UAG P = 0.002). The amount of ghrelin was even higher in the 7.5-month-old rats and resulted in endogenous hyperghrelinemia (Fig. 2A and B). To assess the physiological relevance of this endogenous hyperghrelinemia, we determined the daily food intake and body weight changes in both the mut/mut and wt/wt rats over time. AG levels were slightly elevated in 2.5-month-old mut/mut rats compared with wt/wt littermates (not significant) and highly elevated at the subsequent time points (5.5 and 7.5 months; P < 0.01 and P < 0.001, respectively). In parallel with the increment in AG levels in mut/mut rats, we found a significant increase in food intake combined with a significant body weight gain in these rats compared with the wt/wt animals, especially between the ages of 3 and 6 months (Fig. 2C and D).
To better understand the effect of the high ghrelin levels in mut/mut rats on their feeding behavior, we also performed refeeding experiments where rats were food deprived and then fed, and then the food intake was measured at various time points thereafter (30 min, 1 h, 2 h, 4 h, 8 h, and 24 h). In parallel, AG and UAG levels were measured in all rats after fasting and 2 h postrefeeding. We used 3.5-month-old animals because at this age the mut/mut rats show a significant increase in body weight compared with wt/wt rats (Fig. 2D). The results show that up to 4 h postrefeeding both mut/mut and wt/wt rats eat similar amounts of food but after 8 h the mut/mut animals start eating significantly more than the unaffected ones and the difference in food intake further increases at the 24-h time point (Supplementary Fig. 2A). We also observed that after fasting, UAG and AG levels are similarly elevated in both rat groups (mut/mut and wt/wt), in agreement with the physiological role of ghrelin (Supplementary Fig. 2B and C). After refeeding, the levels of both UAG and AG decreased in all rats, as expected. However, at 2-h postrefeeding, we could already measure a significantly higher level of UAG and a slightly elevated level of AG (not significant) in mut/mut rats compared with the wt/wt ones (Supplementary Fig. 2B and C). These results show that ghrelin levels fluctuate in mut/mut rats with feeding, as in wt/wt rats, but in the former group the hormone levels are increased.
To verify whether the high levels of endogenous ghrelin are responsible for the feeding response of mut/mut rats, we treated these animals with a GHS-R1a antagonist able to block ghrelin signaling, i.e., [D-Lys3]-GHRP-6. Eight mut/mut rats were food deprived for 24 h and then injected with [D-Lys3]-GHRP-6 or saline solution (control group). Food intake was measured at 30 min, 1 h, 2 h, 4 h, 8 h, and 24 h after injection. We found that, up to the 8-h time point, [D-Lys3]-GHRP-6 significantly decreased food intake in the control wt/wt rats, attesting that this compound efficiently blocks ghrelin signaling (Fig. 5A). Similarly, mut/mut rats treated with the GHS-R1 antagonist showed a significant reduction in food intake up to 8 h posttreatment (Fig. 5B). Twenty-four hours after antagonist administration, no changes in food intake could be detected in either rat group, as the compound is probably no longer active. The reduction in food intake for up to 8 h posttreatment was significantly higher in wt/wt versus mut/mut rats, probably because of a higher amount of circulating ghrelin in the latter. Altogether, these data confirm that the elevated ghrelin levels in mut/mut rats are responsible for their feeding response.
Ghrelin-Producing Cells in the Gastric Corpus Mucosa of MENX Rats
Ghrelin is predominantly secreted from X/A-like oxyntic gland cells in the gastric corpus mucosa of the rat stomach (9). To determine whether the number of ghrelin-producing cells changes in affected rats, we stained the gastric corpus tissue of mut/mut and wt/wt rats using an anti-ghrelin antibody (Fig. 3A). Interestingly, the mut/mut rats with high circulating ghrelin levels showed a significant decrease in the number of ghrelin-positive cells in the gastric corpus mucosa compared with the wt/wt rats (Fig. 3B). Concordantly, the ghrelin content of the gastric corpus of wt/wt rats was higher than that of mut/mut rats (Fig. 3C). Ghrelin gene expression in the gastric corpus was not significantly different between the mut/mut and wt/wt animals. These findings support the hypothesis that the high levels of circulating ghrelin in mut/mut rats originate from the endocrine pancreas.
GHS-R1a Expression in the Gastric Corpus Mucosa of MENX Rats
Previous studies showed that GHS-R1a is expressed in both rat and human gastric tissues, among other tissues. Similar to other studies (27), we also demonstrated a high expression of GHS-R1a in the gastric oxyntic glands in the wt/wt rats, particularly in the glands distributed along the middle and basal regions of the stomach corpus region (Fig. 3C). Interestingly, immunostaining against GHS-R1a consistently showed a significant decrease in receptor abundance in the gastric corpus mucosa of 7.5-month-old mut/mut rats compared with that of the age-matched wt/wt rats (Fig. 3D).
Gene expression profiling of gastric corpus mucosa tissue also revealed a decrease in GHS-R1a in 7.5-month-old mut/mut rats compared with wt/wt rats (Fig. 4A), potentially as a consequence of high endogenous ghrelin levels in mut/mut rats.
Ghs-r1a, Npy, and Agrp Expression in the Hypothalamus of MENX Rats
To assess whether the elevated plasma ghrelin levels in mut/mut rats coincide with enhanced activation of orexigenic hypothalamic neurocircuits, we investigated the expression of Ghs-r1a and the ghrelin target genes Npy and Agrp in the hypothalamus of these animals over time. At the age of 5.5 months, MENX-affected rats showed a significant increase in Ghs-r1a and Npy expression compared with age-matched wt/wt animals, whereas no difference was observed for Agrp (Fig. 5A and C). Interestingly, at 7.5 months of age, Ghs-r1a, Npy, and Agrp transcripts were significantly decreased in the arcuate nucleus of mut/mut rats, suggesting desensitization of the ghrelin system (Fig. 5B, D, and F) due to persistently high plasma ghrelin levels.
Body Composition, Leptin Expression, and Plasma Cholesterol and Triglyceride Levels in MENX Rats
Earlier reports showed that chronic ghrelin administration increases body fat mass in rodents and humans (1). Furthermore, UAG and AG were found to stimulate lipid accumulation in adipose tissue (28). The analysis of WAT revealed that epididymal, peritoneal, and inguinal fat pad mass was significantly increased in hyperghrelinemic mut/mut rats compared with age-matched wt/wt rats (Fig. 6A), but this fat accumulation was independent of the common hypothalamic orexigenic pathways (Fig. 5D and F). This observation is in line with previous reports indicating that ghrelin’s orexigenic action in the hypothalamus is independent from its effect in promoting lipogenesis in the adipose tissue (29).
Additionally, fasting plasma cholesterol and triglyceride levels showed a highly significant elevation in mut/mut rats compared with wt/wt rats (Fig. 6B and C), suggesting that endogenous hyperghrelinemia is associated with increased fat mass and dyslipidemia in mut/mut rats. Leptin is primarily produced by the adipocytes in WAT, usually in direct proportion to the amount of WAT. Gene expression profiling of epididymal and peritoneal fat tissue revealed significantly lower levels of leptin in mut/mut rats (Fig. 6D and E). These results are in agreement with plasma leptin measurements in fasting rats, which showed a trend toward a decrease in circulating leptin in mut/mut rats compared with wt/wt rats.
Suppressed GSIS in MENX Rats
As illustrated above, mut/mut rats showed elevated AG and UAG levels at 5.5 months, and these levels further increased at 7.5 months. Previous studies have reported that AG can act as a negative regulator of GSIS (30,31).
Consistent with these publications, mut/mut rats showed a significant decrease in GSIS during i.p.GTT compared with age-matched wt/wt rats (Fig. 7A–F). This phenotype was already observed at 5.5 months of age, when AG and UAG levels were already significantly elevated (Fig. 7A and B).
For determination of whether the route of glucose administration may influence the outcome of the GTT, oGTTs on 7.5-month-old mut/mut and wt/wt rats were performed (Fig. 7G–J). The mut/mut rats also showed suppressed insulin secretion during the oGTT compared with the age-matched wt/wt animals (Fig. 7G–J). Together, these tests suggest that the high plasma levels of AG in mut/mut rats are able to inhibit GSIS in vivo.
Enhanced Insulin Sensitivity in MENX Rats
As described above, mut/mut rats exhibit low insulin secretion with simultaneous unchanged blood glucose levels during GTTs. These data prompted us to perform i.p.ITTs to determine whether the mut/mut animals with high circulating AG and UAG levels display a change in insulin sensitivity in comparison with wt/wt animals (Fig. 8A–E).
In contrast to ghrelin’s proposed effect on impairing glucose metabolism (11), we observed a significant increase in insulin sensitivity in the 7.5-month-old mut/mut rats compared with the age-matched wt/wt animals (Fig. 8B and C). As depicted in Fig. 8A and C, the mut/mut animals showed increased insulin sensitivity already at 5.5 months of age. For gathering additional evidence supporting different insulin sensitivity in mut/mut and wt/wt rats, i.p.ITTs were performed in 5.5-month-old rats with half the dose of insulin we used in the previous experiments. This low amount of insulin caused a strong reduction in plasma glucose in the mut/mut rats but not in wt/wt rats (Fig. 8D and E). Taken together, these data indicate that obese mut/mut rats with high circulating AG and UAG levels show a suppressed GSIS but, interestingly, improved insulin sensitivity.
Here, we present a new nontransgenic animal model with high endogenous levels of AG and UAG in combination with hyperphagia and dyslipidemia and an increased WAT mass.
We observed that rats affected by the MENX syndrome develop pancreatic islet hyperplasia. This observation is consistent with recent reports describing β-cell hyperplasia in p27−/− mice (25,32). In our model, MENX-affected rats show an increase not only in the number of β-cells but also in that of α-cells, δ-cells, and PP-producing cells, leading to an increase in total islet mass.
One interesting observation derived from studies of islet cell hyperplasia is the highly increased number of ghrelin-producing ε-cells present in the endocrine pancreas of adult mut/mut rats with a simultaneous decrease in X/A-like cells in the gastric corpus mucosa, which is normally the major source of ghrelin. An inverse relationship between the number of ghrelin-producing cells in the pancreas and in the stomach has been documented in rats during development (13), with the ghrelin cells in the pancreas being more abundant in late prenatal and early postnatal development when the density of the ghrelin cells in the stomach is still low. In our rat model, the density of the pancreatic and gastric ghrelin cells is also reversed, and we speculate that the elevated circulating ghrelin levels in mut/mut rats are due to the increase in pancreatic ghrelin-producing cells. Although most of the circulating ghrelin comes from the stomach, pancreatic islet-derived ghrelin has been shown to be secreted and to play a role in controlling glucose levels in rodents (30). In vitro, pancreatic islets isolated from wild-type rats were found to produce and secrete AG, which downregulates insulin release after glucose stimulation (31). The same research group later demonstrated that islet-originated ghrelin suppresses insulin secretion in perfused rat pancreata, where islet circulation is well preserved, and that the concentration of ghrelin is higher in pancreatic veins than in arteries of wild-type rats in vivo (30). Additionally, the findings that gastrectomy in human reduces circulating ghrelin to a level of 35% (33) and that rats subjected to sleeve gastrectomy with duodenal-jejunal bypass also show a moderate decrease in fasting ghrelin levels (34) suggest that a minor but substantial amount of the hormone originates from organs other than the stomach, including the pancreas. Altogether, pancreas-derived ghrelin is secreted into circulation and, in addition to its role as a physiological regulator of GSIS, may also affect other metabolic functions. The elevated number of ε-cells in the pancreas of adult mut/mut rats is contradictory to previous publications in which few such cells were present in the endocrine pancreas of adult rats and humans (13) and suggests an altered development and differentiation of neuroendocrine cells in the pancreas of mut/mut animals due to the lack of functional p27. In agreement with previous studies reporting that in rats glucagon and ghrelin only occasionally colocalize around birth (13), in the pancreatic islets of 2-week-old mut/mut rats the ghrelin-positive cells do not coexpress glucagon.
After measuring plasma ghrelin levels, we found that mut/mut rats had increased amounts of AG and UAG compared with the wt/wt animals. The elevated ghrelin levels in the mut/mut rats were associated with hyperphagia and an increased body weight in 5.5-month-old mut/mut rats. At the age of 7.5 months, the mut/mut rats exhibited a high level of dyslipidemia associated with increased WAT. The ghrelin receptor GHS-R1a was initially characterized in the pituitary and hypothalamus, where it is highly expressed and mediates growth hormone release and appetite regulation (35,36). GHS-R1a is also expressed in numerous peripheral tissues, including the gastric enteric nervous plexus in rats and humans (27). Given that the effects of ghrelin are mediated by binding to its receptor, we assessed GHS-R1a expression in the stomach of affected and unaffected rats. We found a significant downregulation of GHS-R1a in the gastric corpus mucosa of the 7.5-month-old hyperghrelinemic mut/mut rats compared with the wt/wt rats. These data suggest the existence of a possible link between chronically elevated ghrelin levels and decreased GHS-R1a expression in the gastric mucosa of mut/mut rats. The fact that stomach cells express both ghrelin and its receptor suggests that ghrelin may potentially act in a paracrine manner to self-regulate its own secretion.
GHS-R1a is mainly present on the NPY/AgRP-expressing neurons in the arcuate nucleus of the hypothalamus, where it mediates ghrelin’s orexigenic action (37). Consistent with data obtained by injecting ghrelin into wild-type Sprague-Dawley rats (38), 5.5-month-old MENX-affected rats showed a significant increase in Npy expression, probably due to the high circulating ghrelin levels. Interestingly, at the age of 7.5 months, MENX-affected rats showed a significant decrease in Ghs-r1a, Npy, and Agrp expression. This reduction could potentially be an indicator of ghrelin resistance. This phenomenon was already described in mice with diet-induced obesity (DIO), which show decreased hypothalamic Ghs-r1a expression. In these mice, DIO is accompanied with downregulation of both Npy and Agrp, and central (intracerebroventricular) ghrelin injection was unable to effect Npy or Agrp expression (39). However, other mechanisms, such as compensatory effects or a negative feedback, could also explain the time-dependent decrease in ghrelin-regulated gene expression in the hypothalamus of our rats.
If the elevated ghrelin is responsible for the feeding behavior of mut/mut rats, we would expect that a blockade of ghrelin signaling with a GHS-R1a antagonist would reduce food intake. This is exactly what we observed after administration of the antagonist [D-Lys3]-GHRP-6 to mut/mut rats. The control group of wt/wt rats also showed a similar response to the antagonist, but the reduction of food intake was more pronounced in wt/wt than in mut/mut rats, likely because of the higher ghrelin levels in the latter.
Previous studies on type 2 diabetic db/db mice showed that deletion of the gene encoding p27 ameliorated hyperglycemia by increasing pancreatic islet mass and maintaining compensatory hyperinsulinemia, effects that were attributable to the stimulation of pancreatic β-cell proliferation (23,25). In the current study, we demonstrated a very similar hyperplastic phenotype of the pancreatic islets in the mut/mut rats. For this reason, we would have expected high insulin levels after the glucose challenge during the oGTTs and i.p.GTTs in the mut/mut rats, but the opposite was the case; the mut/mut animals exhibited a severe inhibition of GSIS, probably due to the high levels of circulating AG. Interestingly, the i.p.ITT performed on obese mut/mut rats showing decreased GSIS revealed significantly improved insulin sensitivity. Data concerning the relationship between ghrelin levels and insulin sensitivity are still controversial. In contrast to our data, it has been demonstrated that ghrelin administration to humans and rodents is accompanied by glucose intolerance (31,40–45). Vice versa, ablation of ghrelin was found to improve glucose disposal and insulin sensitivity (30,46). Tong et al. (43) have shown that in healthy humans a continuous ghrelin infusion suppresses GSIS but also that glucose disappearance is then reduced. However, in agreement with our findings, it has been shown that continuous ghrelin administration enhances insulin sensitivity mainly by acting on peripheral tissues, thereby increasing glucose disposal. For example, Barazzoni et al. (47) have shown that sustained ghrelin administration to rats had an insulin-sensitizing effect on skeletal muscles by enhancing muscle AKT activation. Furthermore, Heijboer et al. (48) have demonstrated that muscle and adipose tissue–specific glucose uptake were significantly higher (by a factor of 3–4) in mice treated with ghrelin. Similarly, Patel et al. (49) found that ghrelin potentiates insulin-stimulated glucose uptake in isolated white adipocytes. The later published data are in agreement with data from our mutant rats displaying hyperghrelinemia and lower GSIS but unchanged glucose concentrations compared with wild-type controls. We therefore speculate that hyperghrelinemic mut/mut rats display improved insulin sensitivity and lower GSIS due to improved peripheral tissue insulin sensitivity. The functional relationship between high ghrelin levels and GSIS in mut/mut rats warrants further investigation.
Another possible explanation for our findings is that the improved insulin sensitivity might be caused by the elevated UAG plasma concentration in mut/mut rats. This effect was described by Gauna et al. (50) in humans, where AG injection caused a decrease in insulin sensitivity, while the coinjection of AG and UAG led to an improvement in insulin sensitivity. Alternatively, we could speculate that in MENX rats the genetic mutation in p27 affects ghrelin-independent signaling mechanisms able to override the negative effect of ghrelin on glucose metabolism. Further studies are required to clarify this issue.
It is well-known that dyslipidemia associated with obesity is a strong risk factor for the development of cardiovascular diseases and type 2 diabetes (51). Because cardiovascular disease has been the leading cause of death worldwide since the 1970s and its frequency is likely to increase, it is imperative to fully understand the mechanisms leading to this metabolic lipid phenotype. As previously mentioned, MENX rats develop dyslipidemia associated with an increased WAT mass and, therefore, represent a model with which to elucidate open questions concerning dyslipidemia development. It has been reported that receptor-mediated uptake and accumulation of plasma cholesterol into the endocrine pancreas reduces GSIS and therefore may represent a link between obesity and type 2 diabetes (52,53). These findings are in accordance with our MENX rat model, in which mut/mut rats showed decreased GSIS that was probably due to high plasma AG levels and further reinforced by high plasma triglyceride and cholesterol levels.
Altogether, the MENX rat provides a meaningful animal model for ghrelin and obesity research. A substantial benefit of this experimental model is the endogenous origin of the elevated plasma ghrelin levels without the need of external administration of the hormone at supraphysiological doses or of genetic modifications of the animals (17,54,55).
The main focus of future studies will be to use this new animal model to shed light on the role of UAG, especially concerning insulin secretion and sensitivity. This research could help to identify new therapeutic approaches for insulin resistance and type 2 diabetes.
Funding. Funding of this research was provided by the Helmholtz Center Munich.
Duality of Interest. No potential conflicts of interest relevant to this article were reported.
Author Contributions. T.W. planned and performed the experiments, wrote the manuscript, and researched data. M.Bie. performed the experiments and contributed to discussion. T.D.M. reviewed and edited the manuscript and contributed to discussion. M.Bid. researched data and contributed to discussion. N.S.P. planned the experiments and reviewed and edited the manuscript. N.S.P. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.