Serum Ghrelin Levels Are Inversely Correlated with Body Mass Index, Age, and Insulin Concentrations in Normal Children and Are Markedly Increased in Prader-Willi Syndrome

Ghrelin, an endogenous ligand of the GH secretagogue receptor, stimulates appetite and causes obesity in animal models and in humans when given in pharmacologic doses. Prader-Willi Syndrome (PWS) is a genetic obesity syndrome characterized by GH deficiency and the onset of a voracious appetite and obesity in childhood. We, therefore, hypothesized that ghrelin levels may play a role in the expression of obesity in this syndrome. We measured fasting serum ghrelin levels in 13 PWS children with an average age of 9.5 yr (range, 5–15) and body mass index (BMI) of 31.3 kg/m² (range, 22–46). The PWS group was compared with 4 control groups: 20 normal weight controls matched for age and sex, 17 obese children (OC), and 14 children with melanocortin-4 receptor mutations (MC4) matched for age, sex, and BMI, and a group of 3 children with leptin deficiency (OB). In non-PWS subjects, ghrelin levels were inversely correlated with age (r = 0.36, P = 0.007), insulin (r = 0.55, P < 0.001), and BMI (r = 0.62, P < 0.001), but not leptin. In children with PWS, fasting ghrelin concentrations were not significantly different compared with normal weight controls (mean ± sd; 429 ± 374 vs. 270 ± 102 pmol/liter; P = 0.14). However, children with PWS did demonstrate higher fasting ghrelin concentrations (3- to 4-fold elevation) compared with all obese groups (OC, MC4, OB) (mean ± sd; 429 ± 374 vs. 139 ± 70 pmol/liter; P < 0.001). In conclusion, ghrelin levels in children with PWS are significantly elevated (3- to 4-fold) compared with BMI-matched obese controls (OC, MC4, OB). Elevation of serum ghrelin levels to the degree documented in this study may play a role as an orexigenic factor driving the insatiable appetite and obesity found in PWS.

GHRELIN IS AN endogenous ligand of the GH secretagogue receptor, a hypothalamic G protein-coupled receptor (1). Enteroendocrine cells (X/A-like cells) of the stomach are the major site of ghrelin synthesis, although a minor proportion of ghrelin synthesis occurs in other sites such as the hypothalamus, pituitary, duodenum, jejunum, and lung (2).

Studies in rodents support the premise that ghrelin is involved in energy balance. In rats fed ad libitum, both intracerebroventricular and ip administration of ghrelin potently stimulates food intake (3). In mice, peripheral injection of ghrelin for 2 wk resulted in a significant increase in body weight that was attributed to an increase in fat mass (4). These mice also displayed an increased respiratory quotient, reflecting a reduced utilization of fat. However, neither energy expenditure nor locomotor activity significantly changed in these mice (4).

In humans, ghrelin was incidentally noted to increase appetite during a clinical study designed to evaluate the GH-releasing activity of pharmacological ghrelin injection in seven normal individuals. Three out of the seven subjects reported a subjective increase in hunger after ghrelin injection (5). An effect of more physiological levels of ghrelin on appetite and food intake in humans was studied recently. In a randomized, double-blind, cross-over study that examined the effects of iv ghrelin on appetite and food intake, energy consumption increased 28 ± 3.9% (P < 0.001) during ghrelin infusion (resulting in ghrelin levels approximately two times higher than fasting) compared with saline infusion (6). Although ghrelin levels are inversely related to body weight in humans (7, 8), ghrelin concentrations are higher during starvation (7, 8) and increase with weight loss (9). Therefore, ghrelin may signal conservation of energy to prevent further weight loss and restore usual body weight.

Prader-Willi Syndrome (PWS) is one of the most common genetic obesity syndromes and is caused by a lack of expression of paternal genes on the long arm of chromosome 15 (15q11-q13) (10). Hypothalamic dysfunction is thought to be the basis for many features of PWS including insatiable hunger and obesity, deficient GH secretion, hypogonadism, aberrant body temperature control, and sleep disturbances, although the underlying mechanisms remain unknown (11). A recent report by Cummings et al. (12) has shown increased plasma ghrelin levels in adults with PWS compared with lean and obese control groups. Therefore, we sought to determine if ghrelin may play a role in the pathogenesis of obesity in children with PWS. In addition to studying children with PWS, we measured endogenous fasting serum ghrelin concentrations in children who were healthy but obese (OC), children with other monogenic causes of obesity [melanocortin-4 receptor mutations (MC4), leptin deficiency (OB)] and children who were healthy and of normal weight (NC). We hypothesized that if circulating ghrelin concentrations were higher in children with PWS than the two control groups, then the orexigenic properties of this hormone may contribute to the obesity phenotype of PWS.

Subjects and Methods

Subjects

Thirteen children with PWS, 20 children of normal weight (NC) matched for sex and age, 17 healthy but obese children (OC), and 14 children with MC4 receptor (MC4-R) mutations (MC4) matched for age, sex, and body mass index (BMI) and 3 children with leptin deficiency (OB) were studied. This study was approved of by the Institutional Review Board of Oregon Health and Science University. A parent of each child gave written informed consent and when appropriate, each child provided assent before entry into the study.

Blood sample analysis

Blood samples were collected after an overnight fast between 0800 and 1000 into red-top vacutainer tubes. Following 30 min to allow clotting on ice, each sample was centrifuged and serum was removed and stored at −70 C until the time of the assay. All samples were measured in duplicate.

Ghrelin analysis

Serum samples were assayed for immunoreactive ghrelin concentration using a commercial RIA (Phoenix Pharmaceuticals, Belmont, CA). Three serum controls from common laboratory stocks were included in each assay to measure intraassay and interassay variability. Based on these controls, the intraassay and interassay coefficients of variation (CV) were 7.16% and 9.86%, respectively. The lower and upper limits of detection for this assay were 24 and 1516 pmol/liter.

Insulin analysis

Serum insulin concentrations were measured in all groups (PWS, NC, OC, MC4, and OB) by using a two-site immunoradiometric method (13). The intraassay and interassay CV were 3–5% and 5–7%. The normal range was 0–69.5 pmol/liter.

Leptin analysis

Serum leptin levels were determined in all groups (PWS, NC, OC, MC4, and OB) using a commercially available RIA (Linco, Inc., St. Charles, MO) with a detection limit of 0.5 ng/ml and intraassay and interassay CV of 2 and 5%, respectively.

Statistical analysis

Outcome variables were compared among the subgroups (PWS, OC, MC4, OB, and NL) using t tests. The correlation between fasting serum ghrelin concentrations and sex, age, BMI, and leptin and insulin concentrations were examined by linear regression and Pearson product moment correlation analyses. All statistical analyses were performed using SigmaStat software (SPSS, Inc., Chicago, IL).

Results

Thirteen PWS children (PWS) (7 girls, 6 boys) with mean age 9.5 ± 3.2 yr and mean BMI of 31.3 ± 8.4 kg/m² were compared with 20 healthy children of normal body weight (NC) (8 girls, 12 boys) with a mean age of 8.7 ± 3.9 yr and BMI of 17.5 ± 2.9 kg/m² and 17 healthy but obese children (OC) (11 girls, 6 boys) with mean age of 10.6 ± 3.5 and BMI of 35.4 ± 6.5 kg/m² and 14 children with MC4-R mutations (MC4) (7 girls, 7 boys) with a mean age of 8.9 ± 2.6 yr and BMI of 29.8 ± 6.0 kg/m² and a group of 3 children with leptin deficiency (OB) (1 girl, 2 boys) with an mean age of 5.0 ± 3.5 yr and BMI of 41.0 ± 6.4 kg/m² (Table 1).

Table 1.

Demographic and laboratory parameters of the participating groups

	Normal controls (NC)	Obese controls (OC)	MC4-R mutants (MC4)	Leptin mutants (OB)	PWS subjects (PWS)
n	20	17	14	3	13
Female	8	11	7	1	7
Male	12	6	7	2	6
Age (yr)	8.7 ± 3.9	10.6 ± 3.5	8.9 ± 2.6	5.0 ± 3.5	9.5 ± 3.2
	(4.4–17.0)	(5.1–17.4)	(5.0–13.4)	(3.0–9.0)	(4.5–14.6)
BMI (kg/m2)	17.5 ± 2.9	35.4 ± 6.5	29.8 ± 6.0	41.0 ± 6.4	31.3 ± 8.4
	(14.1–24.8)	(28.4–54.9)	(22.1–45.0)	(36.0–48.2)	(21.9–46.1)
Serum insulin (pmol/liter)	26.4 ± 37.2	91.2 ± 48	147 ± 72	218 ± 66	40.8 ± 40.8
	(6–166)	(31.2–166)	(79.2–317)	(162–291)	(6.0–140)
Serum leptin (ng/ml)	3.4 ± 3.1	54.0 ± 26.4	36.9 ± 21.1	Not detectable	76.7 ± 56.6
	(1.0–10.5)	(14.4–105.6)	(9.1–90.1)		(6.6–208.8)
Serum ghrelin (pmol/liter)	270 ± 102	146 ± 75	119 ± 40	135 ± 15.5	429 ± 374
	(103–459)	(79.3–313)	(70–221)	(126–153)	(72.5–1570)

Data are means ± sd unless otherwise indicated. Parentheses indicate the range of values. NC and PWS were matched for age and sex; OC and PWS group were matched for age, sex, and BMI.

Fasting ghrelin concentration trended higher (∼1.5-fold) in PWS compared with NC (mean ± sd; 429 ± 374 vs. 270 ± 102 pmol/liter; P = 0.18), although it did not reach statistical significance. PWS demonstrated significantly higher fasting ghrelin concentrations compared with the OC group (mean ± sd; 429 ± 374 vs. 146 ± 75 pmol/liter; P = 0.003), MC4 group (mean ± sd; 429 ± 374 vs. 119 ± 40 pmol/liter; P < 0.001) or OB group (mean ± sd; 429 ± 374 vs. 135 ± 15.5 pmol/liter; P = 0.03). Removal of the PWS with the outlying ghrelin (see Fig. 4) did not change the statistical significance of these results. In addition, we found no significant difference in ghrelin concentrations for those PWS subjects who had maternal disomy vs. a deletion of chromosome 15 as the etiology for their syndrome. No difference in ghrelin concentrations was seen in subjects with MC4-R or leptin deficiency compared with obese controls (Fig. 1).

Figure 4.

Negative relationship between serum ghrelin and BMI is shown in all control groups (NC, OC, MC4, and OB) (r = 0.62, P < 0.001), but not PWS group. The regression line shown is for all control groups only.

Figure 1.

Serum ghrelin levels in 20 normal body weight children (NC) matched for age and sex, 17 obese children (OC), and 14 MC-4R mutant children (MC4) matched for age, sex, and BMI, 3 leptin-deficient children (OB), and 13 PWS children (PWS). Values are mean ± se. P values are given in parentheses for each control group for comparison with PWS.

In non-PWS subjects, ghrelin levels were inversely correlated with age (r = 0.36, P = 0.007), insulin (r = 0.55, P < 0.001), and BMI (r = 0.62, P < 0.001), but not leptin (see Figs. 2, 3³, and 4⁴, respectively). On multiple linear regression analysis, significant independent relationships between fasting serum ghrelin concentrations and BMI and insulin were demonstrated (P = 0.03 and 0.01, respectively) (Table 2). In PWS, however, no significant correlation was seen between fasting ghrelin concentrations and age, BMI, leptin, or insulin.

Figure 2.

Negative relationship between serum ghrelin and age is shown in all control groups (NC, OC, MC4, and OB) (r = 0.36, P = 0.007).

Figure 3.

Negative relationship between serum ghrelin and insulin is shown in all control groups (NC, OC, MC4, and OB) (r = 0.55, P < 0.001).

Table 2.

Multiple linear regression analysis of the relationship between ghrelin and other measures in control groups (normal weight, obese, melanocortin 4 receptor gene mutations, and leptin gene mutations)

Independent variablesa	Standard coefficient	P
BMI	−0.43	0.03
Age	−0.19	0.10
Leptin	0.11	0.57
Insulin	−0.36	0.01

Dependent variable: ghrelin.

^a

r^b for the model = 0.50.

Discussion

Ghrelin is a potent GH secretagogue and, when given in pharmacologic doses, has been shown to induce obesity in animals and stimulate appetite in humans (3, 4, 6). The key finding in the present study is the 3- to 4-fold higher ghrelin levels in the PWS compared with obese matched control subjects. These control groups include children with single genetic mutations associated with extreme obesity including leptin deficiency and mutations in the MC4-R. This corroborates similar findings in PWS adults showing that overnight fasting ghrelin levels were 4.5-fold higher in PWS adults compared with obese controls and 2.5-fold higher than lean controls (12). Also agreeing with data from adults (7), the present data also demonstrate an inverse relationship between BMI and ghrelin levels in non-PWS subjects. It is, therefore, not unexpected that ghrelin levels in PWS, although higher than the lean normal controls, did not reach statistical significance. In addition, we show an inverse relationship between age of the non-PWS subjects (ranging from 4.4 to 15.1 yr) and ghrelin levels. On the other hand, levels of ghrelin in the PWS children did not correlate with age, BMI, or insulin. Therefore, we speculate that ghrelin may be an orexigenic factor contributing to the markedly increased appetite and obesity seen in PWS. This is supported by a recent study that showed ghrelin administered to achieve serum ghrelin concentrations to levels roughly twice that found with fasting, potently stimulated appetite and food intake in adults following an overnight fast (6).

The regulation of ghrelin secretion is poorly understood. Ghrelin secretion is inhibited by ingestion of nutrients (4), which in turn stimulate secretion of a number of gut and pancreatic hormones that might be candidate regulators of ghrelin. It is also possible that ghrelin is inversely regulated by leptin, given ghrelin’s proposed role in body weight regulation. In the present study, the relationship between ghrelin concentrations and age, BMI, leptin, and insulin was examined in the normal control subjects. Whereas age was inversely related to ghrelin levels in normal control subjects by simple regression analysis, only BMI and insulin (not age or leptin) independently predicted ghrelin levels on multiple linear regression analysis. The finding that ghrelin levels in subjects with leptin deficiency were appropriate for their degree of obesity provides evidence that leptin does not have a regulatory role in ghrelin secretion. Therefore, although our data support a role for insulin (or insulin resistance, of which increased insulin levels are a marker) in ghrelin regulation, we cannot be sure that the lower ghrelin levels associated with increasing BMI and insulin levels is independent of other confounding variables such as a higher energy intake.

The mechanism by which ghrelin stimulates appetite has not been fully determined. Evidence now points to ghrelin’s effects in the hypothalamus being mediated through the agouti-related protein (AGRP)/neuropeptide Y (NPY) pathway. Studies demonstrate the presence of ghrelin receptors in multiple hypothalamic nuclei that may be involved in energy homeostasis (14–16). By immunohistochemistry, ghrelin is located in several regions of the hypothalamus (17). Ghrelin when administered centrally, leads to induction of c-fos, a measure of neuronal activation, in the medial arcuate nucleus where NPY/AGRP cells are located (18). Other experiments show that ghrelin is able to induce expression of fos protein in both NPY neurons and AGRP neurons; antibodies and antagonists of NPY and AGRP abolished ghrelin-induced feeding (19). Finally, within the arcuate nucleus itself, there is significant coexpression of GH secretagogue-receptor and NPY mRNA (20). It is plausible that both stomach and hypothalamic-derived ghrelin modulate these hypothalamic centers. For example, stomach-derived ghrelin may interact with the arcuate nucleus, which is less protected by a blood-brain barrier and has some direct connection with the bloodstream. Alternatively, hypothalamic-derived ghrelin may modulate hypothalamic centers located within the blood-brain barrier, such as the paraventricular nucleus.

The increased concentrations of ghrelin found in children with PWS may lead to stimulation of appetite through the hypothalamic NPY/AGRP signaling pathways. Because the human ghrelin gene on chromosome 3p26–25 is not located within the known deleted gene sequence that causes PWS, it is not yet clear how or why this elevation in circulating ghrelin occurs in this syndrome. Because ghrelin is thought to regulate GH secretion, and the majority of PWS children are GH deficient, one might speculate that the increase in ghrelin concentration found in PWS children reflects a lack of feedback inhibition from GH. However, this theory is unlikely given the finding that ghrelin concentrations are not elevated in subjects with GH deficiency and GH therapy in these same individuals does not alter their levels of ghrelin (21). It would be important to know whether most of the increase in ghrelin concentration in PWS is derived from the stomach or from other sources, such as the hypothalamus. Furthermore, because the assay used to measure ghrelin in this study captures all ghrelin-like immunoreactive proteins, it is plausible that the increase in ghrelin concentration is attributable not only to active acylated ghrelin protein, but also to an increase in circulating fragments of ghrelin-derived from degradation or alternative splicing or mutations of the ghrelin gene product, which presumably would be inactive. For instance, it is possible that PWS children may have altered processing of preproghrelin by prohormone convertase 2 into the mature active peptide as supported by previous work that postulates a processing defect of vasopressin in some PWS subjects (22). Therefore, it is possible that the increased ghrelin concentrations demonstrated in PWS children reflect mostly inactive hormone, which may explain why higher levels of GH are not observed in PWS children. A second possibility is that the defect in GH secretion in PWS children is not affected by GH secretagogues such as ghrelin (23, 24). A final speculative explanation is that the elevated ghrelin levels in PWS may lead to lower GH levels through a paradoxical override inhibition, similar to that described before with continuous GH-releasing hormone stimulation of GH (12, 25).

In conclusion, we have found markedly elevated concentrations of ghrelin in children with PWS compared with obese controls and children with monogenic obesity defects in MC4-R or leptin. The absence of increased ghrelin concentrations in children with monogenic obesity defects in MC4-R or leptin suggests that ghrelin may be a unique candidate mediator of the obesity in PWS through mechanisms that include ghrelin’s known ability to increase food intake and decreasing energy expenditure (as demonstrated in rodents). Future studies will need to address the functional activity of the circulating ghrelin in subjects with PWS and whether ghrelin antagonists effectively reduce food intake in PWS and thereby prevent the expression of obesity and its complications.

Acknowledgements

We thank Julia Keogh and Ian Halsall for clinical and biochemical measurements and the patients and families who have generously donated their time to allow this research to be possible.

This study was conducted through the Oregon Health and Science University General Clinical Research Center (M01-RR-00334) and was supported by a grant from Pharmacia, Inc. (A.M.H., S.H.L., and R.G.R.) and by the NIH Grant K-23-DK-02689 (to J.Q.P.).

Abbreviations:

• AGRP,

  Agouti-related protein;

• BMI,

  body mass index;

• CV,

  coefficients of variation;

• MC4,

  melanocortin-4;

• MC4-R,

  MC4 receptor;

• NC,

  normal weight children;

• NPY,

  neuropeptide Y;

• OB,

  children with leptin deficiency;

• OC,

  obese children;

• PWS,

  Prader-Willi Syndrome.