Administration of chemically synthesized ghrelin (Ghr) peptide has been shown to increase food intake and body adiposity in most species. However, the biological role of endogenous Ghr in the molecular control of energy metabolism is far less understood. Mice deficient for either Ghr or its receptor (the growth hormone secretagogue receptor, GHS-R1a) seem to exhibit enhanced protection against high-fat diet-induced obesity but do not show a substantial metabolic phenotype on a standard diet. Here we present the first mouse mutant lacking both Ghr and the Ghr receptor. We demonstrate that simultaneous genetic disruption of both genes of the Ghr system leads to an enhanced energy metabolism phenotype. Ghr/Ghr receptor double knockout (dKO) mice exhibit decreased body weight, increased energy expenditure, and increased motor activity on a standard diet without exposure to a high caloric environment. Mice on the same genetic background lacking either the Ghr or the Ghr receptor gene did not exhibit such a phenotype on standard chow, thereby confirming earlier reports. No differences in food intake, meal pattern, or lean mass were observed between dKO, Ghr-deficient, Ghr receptor-deficient, and wild-type (WT) control mice. Only dKO showed a slight decrease in body length. In summary, simultaneous deletion of Ghr and its receptor enhances the metabolic phenotype of single gene-deficient mice compared with WT mice, possibly suggesting the existence of additional, as of yet unknown, molecular components of the endogenous Ghr system.
- growth hormone secretagogue receptor, constitutive receptor activity
after its discovery in 1999, a series of pharmacological and clinical studies suggested that the gastrointestinal hormone ghrelin (Ghr) was an endogenous regulator of energy balance. Administration of Ghr was shown to promote feeding (23), increase adiposity in rodents (14), and to increase food intake in humans (6). In obese individuals, Ghr levels were demonstrated to be lower than in healthy lean control subjects (16, 25). Careful monitoring of Ghr levels revealed that levels are highest before a meal (3) and subsequently drop postprandially (24). Ghr has therefore been proposed to be a meal initiation factor that induces hunger and subsequent food ingestion (3). New studies on the central action of Ghr on peripheral cell metabolism in rats (17, 22) suggest that Ghr might also be a meal preparation factor. As an integral component of a gut-brain axis that controls nutrient partitioning, it could act to prepare metabolic pathways to more efficiently store ingested calories.
Initial descriptions of Ghr-deficient mice resulting from targeted gene disruption, however, reported normal energy balance, food intake, and adiposity on a standard diet (19, 29, 30, 33). One interpretation of these early findings was that endogenous Ghr may not be of crucial relevance for physiological energy homeostasis. Others attributed such lack of a metabolic phenotype to compensatory processes during early developmental phases or the presumably redundant multiplicity of pathways controlling energy balance. Furthermore, a series of important studies by Birgitte Holst and Thue Schwartz (10, 12) provided solid evidence, suggesting that the (only known) Ghr receptor [growth hormone secretagogue receptor (GHS-R1a)], a 7-transmembrane receptor and member of the G protein-coupled receptor family 1, exhibits basal constitutive activity. GHS-R1a is expressed in numerous tissues that are known to control energy homeostasis and food intake, such as the arcuate and ventromedial nucleus of the hypothalamus (8). GHSR is also expressed in dopaminergic neurons, e.g., within the substantia nigra and ventral tegmental area (8, 32), in brain stem areas such as the nucleus of the solitary tract (32), but also in peripheral tissues such as brown adipose tissue (2) or pancreas (7, 28). The constitutive activity of the Ghr receptor therefore could still provide a possibly residual signaling in Ghr target cells, even in the absence of the primary ligand in mice with targeted disruption of the Ghr gene (11).
Finally, the putative existence of an additional receptor (or ligand) of the Ghr system might explain the lack of a major energy balance phenotype in existing Ghr- or Ghr receptor-deficient animals and can neither be proven nor discarded on the basis of the presently available data (9, 27). We tested the hypothesis that simultaneous deletion of Ghr and the Ghr receptor in mice would lead to a more significant energy metabolism phenotype than deletion of either the ligand or the receptor alone.
MATERIALS AND METHODS
Generation of Ghr- and GHSR-deficient mice has been described previously (1, 26, 29). Ghr knockout (KO) mice of both sexes were intercrossed with GHSR KO mice (F0 generation) to generate a double-heterozygous F1 generation. Double-heterozygous F1 mice were inbred to produce double KO mice deficient for both Ghr and GHSR (dKO), or single KO mice deficient for Ghr or GHSR, respectively. F2 mice with both functional alleles for Ghr and GHSR were used as controls (WT). All results from animals with an age of 12–15 mo were derived from F2 littermates after inbreeding of the double heterozygous F1 generation. Confirmatory data on body weights and body lengths were derived from an additional population of F3 animals at the date of weaning (3 wk) and after 13 wk on chow diet. These mice were produced by homozygous inbreeding of F2 WT, Ghr KO, GHSR KO, or dKO mice. To establish genotypes, DNA was extracted from tail snips, and PCRs were performed using the primers and conditions depicted in Table 1. In addition, all genotypes were confirmed by performing radioimmunoassays to measure Ghr in plasma or by measuring pituitary GHSR gene expression, respectively.
All animals were single housed on a 12:12-h light-dark cycle at 22°C and fed regular chow diet with free access to food and water. All studies were approved by and performed according to the guidelines of the Institutional Animal Care and Use Committee of the University of Cincinnati.
Real-time PCR for pituitary GHSR mRNA expression.
After finishing all in vivo studies, mice were decapitated and pituitaries collected and immediately frozen on dry ice. RNA was extracted using the RNEasy Mini Kit (Qiagen), according to the manufacturer's instructions. After subsequent DNase treatment, reverse transcriptions were performed using SuperScript III (Invitrogen) and oligo(dT20) primers (Invitrogen). Real-time PCRs to validate for GHSR expression and for the ribosomal housekeeping gene L-32 were performed on a Bio-Rad iCycler by using iQ SybrGreen Supermix (Bio-Rad) and two different GHSR primer pairs, 1) GHSR forward primer 5′-TCCGATCTGCTCATCTTCCT-3′ and reverse primer 5′-GGAAGCAGATGGCGAAGTAG-3′, 2) GHSR forward primer 5′-CTACTTCGCCATCTGCTTCC-3′ and reverse primer 5′-AAGACGCTCGACACCCATAC-3′, and L-32 forward primer 5′-GCCAGGAGACGACAAAAAT-3′ and L-32 reverse primer 5′-AATCCTCTTGCCCTGATCC-3′.
Anthropometry analysis and body temperature measurements.
Whole body composition (fat and lean mass) was measured using NMR technology (EchoMRI, Houston, TX). Body length was defined as the distance between nose and rectum and was measured by using a digital high-precision sliding caliper. Body core temperature was measured in the beginning of the light phase in nonanesthetized mice by using a rectal thermometer (Physitemp, Clifton, NJ). Simultaneously, peritoneal body surface temperature was measured by using an infrared thermometer (Fluke, Grossostheim, Germany).
Energy balance physiology measurements.
Energy intake and expenditure, as well as home-cage activity, were studied by using a combined indirect calorimetry system (TSE Systems, Bad Homburg, Germany). After adaptation for >12 h, O2 consumption and CO2 production were measured every 45 min for a total of 76 h to determine the respiratory quotient and energy expenditure. Simultaneously, food and water intake and meal patterns were determined continuously for 76 h by integration of scales into the sealed cage environment. Meals were defined as food intake events with a minimum duration of 60 s, and a break of 300 s between food intake events. In parallel, home-cage locomotor activity was determined using a multidimensional infrared light beam system with beams installed on cage bottom and cage top levels and activity being expressed as beam breaks. Stationary motor activity (fidgeting) was defined as consecutive breaks of one single light beam at cage bottom level, ambulatory movement as breaks of any two different light beams at cage bottom level, and rearing as simultaneous breaks of light beams on both cage bottom and top level.
Blood was collected after an overnight fast from tail veins using EDTA-coated Microvette tubes (Sarstedt, Nuremberg, Germany) and immediately chilled on ice. After 15 min of centrifugation at 3,000 g and 4°C, plasma was stored at −80°C. For quantification of plasma Ghr levels, a commercially available radioimmunoassay [Ghrelin (Rat/Mouse) RIA; Phoenixpeptide, Burlingame, CA] was used. This assay is validated for the measurement of both des-octanoylated and octanoylated Ghr in nonacidified and acidified plasma samples and does not require protein extraction. For quantification of plasma insulin levels, a radioimmunoassay from Linco (Sensitive Rat Insulin RIA; Linco Research, St. Charles, MO) was used. Plasma leptin levels were measured by using an ELISA kit (Murine Leptin; Diagnostic Systems Laboratories, Webster, TX). Plasma glucose and nonesterified fatty acid (NEFA) were measured by using commercially available enzymatic assay kits from Wako (Autokit Glucose and NEFA C; Wako, Neuss, Germany). Plasma cholesterol levels were determined with the cholesterol oxidase method by using Infinity Cholesterol reagent, and plasma triglycerides were quantified by using the Infinity Triglyceride reagent (Thermo Electron, Pittsburgh, PA). All assays were performed according to the assay manufacturer's instructions.
Glucose tolerance test and insulin tolerance test.
For the measurements of glucose tolerance and insulin sensitivity, mice were subjected to 6 h of fasting and injected intraperitoneally with 2 g glucose/kg body wt (50% D-glucose (Sigma) in 0.9% saline) for the glucose tolerance test (GTT), and 1 U insulin/kg body wt (0.1 U/ml; Humolog Pen, Lily, Indianapolis, IN) for the insulin tolerance test (ITT). Tail blood glucose levels (mg/dl) were measured by using a handheld glucometer (TheraSense Freestyle) before (0 min) and at 15, 30, 60, 90, and 120 min after injection.
After 20 h of fasting, mice were refed with standard diet and food intake, and meal patterns were measured continuously for 24 h by using an automated food-monitoring system (TSE Systems, Bad Homburg, Germany).
Unless indicated otherwise, all statistical analyses were performed on the basis of planned comparisons and/or using Fishers least significant difference (LSD) test. These analyses of statistical validity were selected because in each case simple comparisons were made with strong a-priori predictions on the basis of the genetic deletion of both Ghr and the Ghr receptor. Differences in the refeeding between groups, as well as in glucose levels during GTT or ITT were examined by two-way ANOVAs, and Fishers LSD post hoc tests. All statistical analyses were performed with Statistica 6.0 (StatSoft, Tulsa, OK). Experiment-wise α error was set at P < 0.05. All results are presented as means ± SE.
Generation of dKO mice.
To produce mice deficient for both Ghr and GHSR, homozygous KO mice for either Ghr or GHSR were intercrossed. The generation of Ghr-deficient mice has been described previously (1, 29). In a similar fashion as Ghr-deficient mice, GHSR-deficient mice were generated by using the high-throughput VelociGene gene-targeting system (26). The resulting double-heterozygous F1 generation was inbred to produce F2 animals. All animals with an age of 12–15 mo during the times of studies were littermates of this F2 generation. In addition, F2 animals of each genotype were bred true to generate F3 WT, Ghr KO, GHSR KO, and dKO mice to study the effect of Ghr deficiency, GHSR deficiency, or both in young animals (See supplemental information available online at the American Journal of Physiology Gastrointestinal and Liver Physiology website.). Genotypes were initially established by using PCR (Fig. 1A) from genomic DNA. In addition, Ghr and GHSR deficiency was verified in F2 mice by measuring total plasma Ghr levels and pituitary GHSR gene expression, respectively. Ghr KO and dKO mice showed no detectable levels of total Ghr, whereas GHSR KO and WT showed similar levels corresponding with normal mouse Ghr levels reported previously (Fig. 1B) (18, 23). GHSR KO and dKO mice did not express GHSR mRNA in their pituitaries, in contrast to dWT and Ghr KO mice (Fig. 1C). Fertility was normal in all mice, and no differences in mortality rate or overall health could be observed between groups.
Reduced body weight of dKO mice.
Body weight was found to be significantly decreased in both old dKO mice (13.4 mo) (P = 0.032, Fig. 2A), as well as in mice of young age (13 wk) (P = 0.036, Supplemental Fig. S1). Body weight at the age of weaning (3 wk after birth) was comparable between groups (Supplemental Fig. S1). The decrease in body weight of old mice was predominantly due to a decrease in fat mass (P = 0.047, Fig. 2C) but not lean mass (P = 0.18, Fig. 2D). Young mice at 13 wk of age also showed a slightly but significantly reduced lean mass (P = 0.011, Supplemental Fig. S1). However, the percentage of both fat and lean mass in young mice was comparable between groups (Supplemental Fig. S1). Old dKO mice showed reduced body length (P = 0.013, Fig. 2B), whereas young dKO mice at 13 wk did not differ from WT controls (P = 0.78, Supplemental Fig. S1). Leptin levels in old mice and IGF-I plasma levels in young and old mice did not differ between genotypes and did not accompany the observed changes in fat mass or body length, respectively (Table 2).
Altered lipid profiles in dKO mice.
Free fatty acid and triglyceride levels did not show differences between groups, but cholesterol levels were significantly lower only in old dKO mice compared with WT mice following an overnight fast (P = 0.015, Table 2).
Normal feeding behavior in dKO mice.
No differences in overall food intake (Fig. 3A, left) or water intake (data not shown) were found between aged mice with deficiencies for Ghr, GHSR, or both genes compared with WT controls. Twenty-four-hour food intake did not differ between genotypes during the light or dark phase, respectively (Fig. 3A, right, top). In addition, normalization of 24-h food intake per body weight (Fig. 3A, right, middle) or lean mass (Fig. 3A, right, bottom) did not reveal any differences in relative food intake between genotypes. Meal patterns were also not affected by deficiencies for Ghr, GHSR, or both genes. The number of meals per 24 h (Fig. 3B, left, top), the average duration of the meals (Fig. 3B, right, top), the average size of a meal (Fig. 3B, left, bottom), and the average meal rate (Fig. 3B, right, bottom) did not differ between the four genotypes.
Increased energy expenditure and body core temperature in dKO mice.
Only dKO mice, but not mice deficient for either Ghr or the Ghr receptor, showed significantly higher energy expenditure (Fig. 4A, left) in both the light (P = 0.0006) and dark phase (P = 0.0007; Fig. 4A, right, top). Normalization of the average energy expenditure to the “metabolically active” lean tissue corroborated this finding (light phase P = 0.0027, dark phase P = 0.006 vs. WT, respectively; Fig. 4A, right, bottom). This increase in energy expenditure was consistent with higher body core temperature of dKO mice (P = 0.043; Fig. 4C, top, +0.5°C), whereas peritoneal surface temperature (Fig. 4C, bottom) was not found to be significantly higher (P = 0.73). As depicted in Fig. 4B, we were unable to find any changes in the respiratory quotient between groups.
Increased spontaneous physical activity in dKO mice.
In parallel to the indirect calorimetry measurements described above, locomotor activity was monitored by using infrared beam break systems (Fig. 5A). Total locomotor activity (Fig. 5A) was further dissected into ambulatory movements (Fig. 5B), stationary movements (Fig. 5C), and vertical movements (rearing, Fig. 5D). dKO mice showed significantly more stationary movements (P = 0.033 dark phase; P = 0.036 light phase; Fig. 5C) and exhibited strong tendencies toward higher ambulatory and rearing activity. Interestingly, GHSR KO mice showed no significant differences in locomotor activity, and Ghr KO mice only showed a minor increase in total locomotor activity during the light phase, compared with WT control mice (P = 0.004; Fig. 5, A, B, and D, right).
Normal insulin sensitivity in dKO mice.
After an overnight fast, plasma glucose (P = 0.86, Fig. 6A) and plasma insulin levels (P = 0.12, Fig. 6B) did not differ between aged WT and dKO mice. A GTT (intraperitoneal glucose bolus: 2 g glucose/kg body wt) overall failed to reveal significant differences in glucose tolerance between genotypes (Fig. 6, C and D). Mice deficient in either Ghr, GHSR, or both showed lower glucose peak levels at a single time point (15 min after the injection), possibly suggesting a slightly faster release of insulin (Fig. 6C). To further investigate the endogenous role of the Ghr system in the control of glucose metabolism, mice were subjected to an ITT (1 U insulin/kg body wt) (Fig. 6, E and F). In Ghr KO mice, glucose levels were similar to WT mice. In dKO and GHSR KO mice, glucose levels, however, dropped more rapidly. After 15 min, glucose values in dKO mice were 25 mg/dl (P = 0.08), or 11% lower than in WT mice. In GHSR KO mice, glucose levels after 15 min were 30 mg/dl (P = 0.0061), or 15.7% lower than glucose levels of WT mice after 15 min. In general, glucose levels of dKO and GHSR mice tended to remain lower throughout the 120 min of the ITT, compared with WT mice (Fig. 6E). However, although integrated (area under the curve) glucose levels in both GHSR KO and dKO mice tended to be lower compared with WT control mice (dKO: P = 0.105, and GHSR KO: P = 0.065; Fig. 6F), the deficiency of Ghr, its receptor, or both did not seem to have a major impact on overall insulin sensitivity or the overall regulation of glucose homeostasis.
Normal meal pattern following food deprivation in dKO mice.
Old mice of all four genotypes were fasted for 20 h and then given free access to standard chow diet. Subsequently, food intake and meal patterns were analyzed for 24 h using an automated food intake monitoring system. No changes were observed in overall food consumption between genotypes (including normalization of food intake per body weight or lean mass). Although a meal pattern analysis of the first 4 h of refeeding suggested a slightly lower food ingestion rate only in dKO mice compared with WT mice (P = 0.11; Fig. 7B, right, bottom), none of the meal pattern tests revealed significant differences. No significant changes where found regarding number of meals (P = 0.60 WT vs. dKO; Fig. 7B, left, top), duration of meals (P = 0.29 WT vs. dKO; Fig. 7B, left, bottom), or size of meals (P = 0.29 WT vs. dKO; Fig. 7B, left, bottom). When meal patterns of the first 8 h of refeeding were analyzed, the trend toward a lower food ingestion rate became smaller and completely disappeared over 24 h (data not shown).
Here we demonstrate that simultaneous deficiency for Ghr and the Ghr receptor leads to changes in energy balance, which is not observed in mice deficient for either the ligand or the receptor alone. We observed significant differences in energy expenditure, stationary locomotor activity, body weight, body fat, body length, and plasma cholesterol. We also observed substantial but mostly nonsignificant trends in glucose tolerance and insulin sensitivity. Importantly, all data presented here resulted from studies of dKO mice maintained on normal standard chow diet, whereas previous energy balance or metabolism phenotypes reported from mouse mutants with deficient Ghr signaling were observed mostly after chronic exposure to high-fat diet (29, 30, 33). Our results on the biological function of the endogenous Ghr system using three mouse mutant strains (Ghr deficient, GHSR deficient, Ghr-GHSR double deficient), indicate an additive “gene-dosage”-like effect. For most parameters in our study, Ghr deficiency had less impact on energy metabolism components than GHSR deficiency, but only simultaneous deficiency for both genes led to significant differences from wild-type (WT) control mice. These findings are consistent with both the functional relevance of constitutive GHSR activity and the putative existence of other Ghr system components, such as an additional Ghr receptor or another Ghr-like ligand.
This is the first report on an animal model with combined deletion of Ghr and the Ghr receptor genes. The most important observation may be the substantially increased energy expenditure in these mice compared with WT control mice. We were unable to verify the previously reported increase of energy expenditure in Ghr-deficient mice on chow diet (4) but rather found the additional absence of the Ghr receptor GHSR to be a prerequisite for increased energy expenditure compared with WT controls. The lack of multiple Ghr system components therefore had an additive impact on thermogenetic processes. The exact cause for the lower body weight of dKO mice cannot be easily identified on the basis of the present data. On one hand, the old mice also exhibit a modestly, but significantly, shorter body length. However, the similar body weights after 3 wk, as well as the comparable body lengths in 13-wk-old animals, do not suggest an impaired embryonic or pubertal development. Furthermore, IGF-I levels did not differ between any groups in young or old mice. In fact, the lower body weights in old and young dKO mice are unlikely to be a consequence of decreased body length alone because both changes in lean mass and fat mass contributed to the observed changes. No significant differences in lean mass were detected in old mice. Young dKO mice, albeit having the same body length, have a significantly lower lean mass compared with control mice. However, the percentage of lean tissue towards whole body mass does not differ between groups. In old mice, the lower fat mass certainly contributes to their significantly lower body weight. Supporting that notion, the lower plasma cholesterol, as well as the increased energy expenditure and locomotor activity levels, which we have observed in the dKO mice, would be consistent with a negative energy balance and decreased body fat. However, young mice, in contrast, show no changes in fat mass, at least at 13 wk of age. Further research will therefore be necessary to dissect the impact of aging on the role of Ghr signaling in the control of energy balance and body composition.
Increased locomotor activity in the absence of Ghr signaling is consistent with earlier observations, demonstrating decreased spontaneous physical activity after intracerebroventricular Ghr administration in rats (21). Furthermore, increased locomotor activity and energy expenditure was previously described in Ghr-deficient (30) and GHSR-deficient mice (33) when chronically challenged with a high caloric diet.
We here also observed tendencies toward enhanced glucose tolerance and insulin sensitivity in dKO mice. However, those differences barely reached statistical significance, and basal blood glucose and insulin levels did not differ between the four mouse strains. Our results therefore fail to further support recent studies showing that genetic deletion of Ghr in leptin-deficient (ob/ob), mice can partially rescue their diabetic phenotype (20). Since the obesity phenotype of Ghr-deficient ob/ob mice was still normal, that report had very plausibly suggested that endogenous Ghr may be more important for glucose homeostasis than for energy metabolism (20).
Finally, we speculate that the presently described mouse mutants still may exhibit some level of Ghr signaling, although Ghr (15), its putative Ghr associated peptide (31), Ghr splice variants (14), and the constitutively active Ghr receptor GHSR (13) have by definition all been genetically deleted. The existence of both an additional ligand and an additional receptor coded for by genes other than the Ghr and the GHSR gene could explain why the dKO mouse shows a phenotype that still has to be categorized as very mild (Supplemental Fig. S2).
Indirect indication for additional molecular components of the endogenous Ghr system can be derived from the studies of single gene deletion studies targeting either the Ghr ligand or its receptor (19, 20, 29, 30, 33). Although both phenotypes overall show some level of protection against high-fat diet-induced obesity on the basis of a more negative energy balance, a number of clear differences can be found when comparing published data on these two animal models, which resulted from chronic exposure to high-fat diet (30, 33). In one of the first studies on Ghr-deficient mice, they only differed from WT controls if exposed to a high-fat diet immediately after weaning and for at least 4 mo (30). Fat mass gain of Ghr receptor-deficient mice differs from WT control mice even when exposed to high-fat diet at adult age. In the presently presented studies on Ghr-deficient and Ghr receptor-deficient single gene KO mice, we partially confirm the observations of Zigman et al. (33) and Wortley et al. (29, 30), but we cannot confirm the findings reported by De Smet and colleagues (4) indicating an energy expenditure and body weight phenotype of Ghr single gene KO mice on chow diet. Increased motor activity was found in Ghr-deficient mice on a high-fat diet but not in GHSR-deficient mice on any diet (4, 30). As a matter of fact one report suggested decreased locomotor activity in mice deficient for the Ghr receptor GHSR (33). We speculate that either Ghr may act also through an additional, as of yet unknown, receptor, or that another Ghr-like ligand exists. On the basis of presently published data, this ligand is unlikely to be identical with the recently described Ghr-associated peptide, obestatin.
In summary, we conclude that Ghr has a physiological role in the regulation of energy expenditure and body weight under unchallenged environmental conditions. Future studies will have to address 1) whether the additive phenotype resulting from simultaneous deletion of Ghr and its receptor will also be apparent for other biological functions of Ghr such as learning and memory (5), 2) whether high-fat diet exposure enhances the phenotype of dKO similar to observations in mice deficient for either the Ghr ligand or the Ghr receptor, and 3) whether additional, as of yet unknown, ligands or receptors play a functionally relevant role for the endogenous Ghr system.
This research was supported by National Institutes of Health/National Institute of Diabetes and Digestive and Kidney Diseases Grant R01-DK069987-02 (M. H. Tschöp).
We are grateful to Kimberly Brown, Erin Grant, Nickki Ottaway, Hilary Wilson, and Rebecca Greenwald for skilled assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society