Skip to main content
PMC home page
The Journal of Clinical Endocrinology and Metabolism logoLink to The Journal of Clinical Endocrinology and Metabolism
. 2010 Oct 20;96(1):E225–E232. doi: 10.1210/jc.2010-1733

The Metabolic Phenotype of Prader-Willi Syndrome (PWS) in Childhood: Heightened Insulin Sensitivity Relative to Body Mass Index

Andrea M Haqq 1, Michael J Muehlbauer 1, Christopher B Newgard 1, Steven Grambow 1, Michael Freemark 1
PMCID: PMC3038476  PMID: 20962018

Abstract

Context: Insulin sensitivity is higher in patients with Prader-Willi syndrome (PWS) than in body mass index-matched obese controls (OCs). Factors contributing to the heightened insulin sensitivity of PWS remain obscure. We compared the fasting levels of various hormones, cytokines, lipids, and liver function tests in 14 PWS patients and 14 OCs with those in 14 age- and gender-matched lean children (LC). We hypothesized that metabolic profiles of children with PWS are comparable with those of LC, but different from those of OCs.

Results: Leptin levels were comparable in PWS patients and OCs, suggesting comparable degrees of adiposity. Glucose levels were comparable among groups. However, fasting insulin concentrations and homeostasis model assessment insulin resistance index were lower in PWS patients than in OCs (P < 0.05) and similar to LC. Moreover, high-density lipoprotein levels were lower and triglycerides higher in OCs (P < 0.05) but not PWS patients. Total adiponectin, high-molecular-weight (HMW) adiponectin and the HMW to total adiponectin ratio were higher in PWS patients (P < 0.05) than in OCs and similar to LC. High-sensitivity C-reactive protein and IL-6 levels were higher in OCs than in PWS patients or LC (P < 0.05). Nevertheless, PAI-1 levels were elevated in both OC and PWS patients. There were no group differences in glucagon-like peptide-1, macrophage chemoattractant protein-1, TNFα, IL-2, IL-8, IL-10, IL-12p40, IL-18, resistin, total or low-density lipoprotein cholesterol, aspartate aminotransferase, or alanine aminotransferase.

Conclusions: The heightened insulin sensitivity of PWS patients relative to OCs is associated with higher levels of adiponectin and lower levels of high-sensitivity C-reactive protein and IL-6. Future studies will determine whether PWS children are protected from obesity comorbidities such as type 2 diabetes, hyperlipidemia, and nonalcoholic fatty liver disease.

The heightened insulin sensitivity of Prader-Willi syndrome relative to obese controls is associated with higher levels of adiponectin and lower levels of pro-inflammatory cytokines, hsCRP, and IL-6.

Prader-Willi syndrome (PWS) is a relatively common genetic obesity syndrome, with incidence of one in 10,000 to one in 15,000 live births. Infants with PWS have hypotonia, poor suck, decreased arousal, and failure to thrive and often require tube feedings for several weeks to months. This period is followed by progressive obesity by 1–6 yr of age, insatiable appetite, short stature secondary to deficient GH secretion, delayed puberty, possible central adrenal insufficiency, delayed motor and cognitive development, behavioral difficulties, and sleep disturbances (1).

PWS is an imprinting disorder caused by the lack of expression of paternally derived genes on chromosome 15q11–13. The genetics of PWS is complex; several candidate genes are implicated including the SNURF-SNRPN gene (encoding small nucleolar RNAs) (2), necdin (encoding necdin protein) (3), and MAGEL2 (a necdin related protein) (4). It is clear that the genes disrupted in PWS include a gene or set of genes that play vital roles in the regulation of body fat and metabolism.

Interestingly, numerous studies have shown that PWS individuals demonstrate a state of relative hypoinsulinemia despite profound obesity. Fasting insulin concentrations are lower in PWS compared with body mass index (BMI)-matched children (5), and insulin sensitivity is increased (6,7). Compared with BMI-matched controls, children with PWS have decreased insulin responses to mixed meals and oral and iv glucose loading (8). Additionally, the first- and second-phase insulin secretory responses are significantly lower in PWS adults compared with obese controls during an iv glucose tolerance test (6). Finally, increased sensitivity to exogenous insulin has been observed in PWS (9).

The factors contributing to the heightened insulin sensitivity of children with PWS remain obscure. Some studies implicate abnormal partitioning of body fat and lean mass: whole-body magnetic resonance imaging has found that PWS adults have greater fat mass relative to fat-free mass, but significantly less visceral adiposity than BMI-matched controls (10). Additionally, preliminary studies from our group and others have found higher adiponectin concentrations and increased ratios of high-molecular-weight (HMW) to total adiponectin in PWS children compared with BMI-matched controls (7,11,12,13,14).

In this study we compared measures of insulin sensitivity and the levels of various hormones, cytokines, lipids, and liver function tests in PWS children and BMI-matched controls with those in age- and gender-matched lean children. We hypothesized that the insulin sensitivity and metabolic profiles of children with PWS would be comparable with those of lean children (LCs), but would differ from those of BMI-matched obese controls (OCs).

Subjects and Methods

The study protocol was approved by the Institutional Review Board of the Duke University Medical Center. Informed written consent from the parents/guardians and assent from each child was obtained.

Subjects

Subjects with PWS were recruited from local pediatric endocrinology clinics and also by advertisements through several national PWS organizations. OCs and lean subjects were recruited from local community pediatricians and the insulin resistance and obesity clinics at Duke University Medical Center. Fourteen children with PWS (median age and BMI Z-score: 11.4 and 2.15, respectively), 14 age-, gender-, and BMI-matched obese control children without PWS (median age and BMI Z-score 12.0 and 2.35 yr, respectively), and 14 age- and gender-matched LC (median age and BMI Z-score 12.3 and −0.6 yr, respectively) were studied. Subjects with chronic secondary illnesses such as diabetes mellitus, liver or kidney disease or active malignancy, or those taking investigational drugs were excluded. Nine PWS subjects were taking a stable dose of GH at the time of study. Referring physicians made the initial decisions whether to treat with GH; the mean GH dose of each subject was 0.02 mg/kg/d. All PWS subjects had free T4 and TSH levels in the normal range (either endogenous or on replacement); one PWS subject was on thyroid replacement to treat central hypothyroidism. Four subjects with PWS had uniparental disomy; all others had a deletion in chromosome 15. Table 1 shows the subjects’ clinical characteristics.

Table 1.

Baseline metabolic parameters of PWS, BMI-matched controls (OCs), and LC

PWS (n = 14) OC (n = 14) LC (n = 14)
Age (yr) 11.4 (7.1, 14.6)a 12.0 (10.6, 14.2)a 12.3 (10.8, 13.8)a
Males/females 5/9a 6/8a 8/6a
BMI Z-score 2.2 (1.5, 2.6)a 2.3 (1.8, 2.7)a −0.6 (−1.1, +0.04)b
Glucose (mg/dl) 87.8 (81.8, 101.5)a 94.9 (89.0, 99.5)a 99.0 (93.0, 101.0)a
Insulin (μIU/ml) 13.49 (10.17, 29.38)a 23.20 (17.17, 34.70)b 10.16 (7.32, 13.23)a
HOMA-IR 3.2 (2.5, 5.7)a 5.1 (4.1, 8.6)b 2.3 (1.8, 3.2)a
Total adiponectin (ng/ml) 8,769 (6,364, 15,046)a 5,798 (3,217, 7,353)b 9,923 (9,016, 13,512)a
HMW adiponectin (ng/ml) 5,371 (2,763, 11,868)a 2,523 (819, 3,147)b 5,645 (4,689, 7,649)a
HMW to total adiponectin ratio 0.71 (0.48, 0.84)a 0.40 (0.27, 0.51)b 0.54 (0.52, 0.61)a,b
Leptin (ng/ml) 31.5 (12.9, 42.7)a 32.2 (15.6, 54.5)a 2.2 (1.9, 4.2)b
Total adiponectin to leptin ratio 257.1 (126.5, 1105.6)a 132.3 (62.1, 366.2)a 4,724.0 (2,147.7, 5,224.6)b
Total GLP-1 (pg/ml) 9.96 (6.87, 15.48)a 11.56 (10.85, 14.05)a 8.825 (6.26, 12.01)a
Open in a new tab

Data reported as medians (interquartile range); Kruskal-Wallis test was used for between-group comparisons. Groups with different letter superscripts are significantly different (P < 0.05) from each other. 

Anthropometric measures

Weight was measured in subjects wearing light clothing using the same calibrated digital scale. Height was measured in triplicate using a Harpenden stadiometer. BMI Z-scores were calculated using EpiInfo version 3.3.2 (Centers for Disease Control and Prevention, Atlanta, GA).

Experimental design

All subjects were admitted to the General Clinical Research Center at Duke University Medical Center. Blood samples were obtained after an observed 10-h overnight fast.

Assays

Blood samples were collected after an overnight fast between 0800 and 1000 into red-top Vacutainer tubes (for serum) or purple Vacutainer tubes (for plasma). Aprotinin (500 kIU/ml of blood; Roche, Indianapolis, IN) was added to each vacutainer. Serum samples were allowed to clot on ice for 30 min and each sample was centrifuged, and the serum or plasma was removed and stored at −70 C until time of the assay. All analytes were measured in duplicate. Fasting plasma insulin and leptin were performed by RIA (Linco, St. Charles, MO). Plasma total and HMW adiponectin were done by ELISA from ALPCO (Salem, NH). Serum total glucagon-like peptide-1 (GLP-1) was measured using electrochemiluminescent assay technology, with commercial kits from Meso Scale Discovery (Gaithersburg, MD). Fasting cytokines [plasma TNF-α, IL-10, IL-6, IL-1b, IL-2, IL-12p40, IL-8, macrophage chemoattractant protein-1(MCP-1), plasminogen activator inhibitor-1 (PAI-1), interferon (IFN)-γ, IL-12p70, serum IL-18, and resistin] were measured via a multiplex chemiluminescent assay using a Searchlight Plus CCD imager (Pierce Biotechnology, Woburn, MA). These cytokines were measured either because they are known mediators [TNF-α, IL-6, C-reactive protein (CRP), MCP-1, PAI-1, and resistin] (15,16) of insulin resistance or their contribution to insulin resistance was unclear (IL-10, IL-1b, IL-2, IL-12p40, IL-8, IFNγ, IL-12p70, and serum IL-18). Fasting serum high-sensitivity CRP, and plasma free (nonesterified) fatty acids (FFAs), total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL), triglycerides, and glucose were performed using a Hitachi 911 autoanalyzer (Tokyo, Japan). Reagents for FFAs were supplied by Wako Chemical USA (Richmond, VA), and for the remaining analytes by Roche Diagnostics. Plasma aspartate aminotransferase (AST) and alanine aminotransferase (ALT) were measured using a Beckman-Coulter UniCel DxC600 with reagents from Beckman (Fullerton, CA). Homeostasis model assessment was calculated using the formula: fasting glucose (milligrams per deciliter) × fasting insulin (microinternational units per milliliter)/405. Coefficients of variation, on average, are less than 10% for all assays reported.

Statistical analyses

Because of the nonnormal distribution for many variables, all data are presented as the median and interquartile range. Between-group comparisons were performed using the nonparametric Kruskal-Wallis test (for three group comparisons) and the Wilcoxon rank-sum test (for two group comparisons). Spearman rank correlation coefficients were used to measure monotonic associations between variables. Statistical analyses were performed using SAS 9.1 for Windows (SAS Institute, Cary, NC) and SigmaStat (SPSS, Inc., Chicago, IL). Two-sided P values at the standard 0.05 level were used to determine statistical significance.

Results

We studied 14 subjects with PWS, 14 age-, gender-, and BMI-matched OCs, and 14 age- and gender-matched LC. Subject characteristics and comparisons are shown in Table 1. Groups were comparable for age and sex distribution. Median BMI Z values were comparable among subjects with PWS and OCs but lower in LC.

Fasting insulin, homeostasis model assessment insulin resistance index (HOMA-IR), total and HMW adiponectin, leptin, and total GLP-1

Relative to LC, PWS subjects and OCs had higher fasting leptin levels (P < 0.05). Yet fasting insulin levels and HOMA-IR in PWS subjects were comparable with those in LC and 42 and 37% lower (P < 0.05 for both), respectively, than those in OCs. Likewise, total and HMW adiponectin levels in PWS subjects were comparable with those in LC and 51 and 112% higher (P < 0.05 for both), respectively, than those in OCs. These findings were not explained by differences in age, gender, or pubertal status among the groups (data not shown). There were no significant differences in total GLP-1 among the three groups (Table 1).

Fasting adipocytokines

Relative to LC, high-sensitivity CRP levels were more than 7-fold higher in OCs (P < 0.05); in contrast, high-sensitivity CRP levels in PWS subjects were comparable with LC. IL-6 levels in children with PWS were 41% lower (P < 0.05) than those in OCs. Conversely, circulating PAI-1 levels in PWS subjects and OCs were 2- to 3-fold higher (P < 0.05 for both) than in LC. There were no significant differences in plasma MCP-1, TNFα, IL-2, IL-8, IL-10, IL12p40, IL-18, or resistin among the three groups.

Small intergroup differences in the levels of other ILs are difficult to interpret because the circulating concentrations are highly variable and generally fell less than 20 pg/ml (Table 2).

Table 2.

Fasting adipocytokine measurements in PWS, BMI-matched controls (OCs), and LCs

PWS (n = 14) OC (n = 14) LC (n = 14)
TNF-α (pg/ml) 20.40 (6.50, 38.00)a 46.50 (11.10, 88.60)a 47.05 (17.40, 98.30)a
CRP (mg/liter) 1.40 (0.30, 5.30)a,b 2.20 (0.60, 5.20)a 0.30 (0.30, 0.80)b
PAI-1 (pg/ml) 29,284.80 (15,553.85, 82,937.40)a 38,914.00 (22,762.40, 59,151.20)a 13,603.60 (11,429.10, 23,033.00)b
IL-6 (pg/ml) 11.65 (6.55, 14.80)a 19.90 (17.90, 43.60)b 22.45 (11.40, 33.10)a,b
IL-12p70 (pg/ml) 1.65 (1.05, 4.40)a 7.15 (1.70, 9.90)a,b 7.90 (3.50, 23.20)b
IL-1b (pg/ml) 0.80 (0.25, 1.70)a 1.10 (0.60, 4.70)a,b 4.35 (1.20, 8.70)b
IL-10 (pg/ml) 2.55 (1.25, 4.35)a 5.20 (2.60, 10.90)a 7.20 (2.80, 19.00)a
IL-2 (pg/ml) 8.70 (2.10, 11.70)a 22.80 (5.90, 30.00)a 19.85 (4.30, 38.40)a
IL-12p40 (pg/ml) 8.95 (6.05, 14.70)a 17.35 (3.60, 53.20)a 29.40 (10.60, 49.30)a
IL-8 (pg/ml) 14.55 (7.60, 23.60)a 11.05 (8.60, 37.80)a 16.25 (11.60, 20.90)a
IFNγ (pg/ml) 7.90 (2.55, 12.20)a 14.90 (5.10, 41.80)a 10.05 (8.80, 31.30)a
IL-18 (pg/ml) 105.95 (89.40, 149.80)a 99.30 (87.50, 110.80)a 93.35 (81.20, 106.50)a
MCP-1 (pg/ml) 450.80 (335.35, 633.75)a 415.95 (354.70, 519.20)a 435.80 (409.60, 468.30)a
Resistin (pg/ml) 10,067.20 (8,534.10, 14,787.20)a 12,125.25 (10,379.90, 16,161.10)a 10,374.65 (8,509.50, 14,470.00)a
Open in a new tab

Data reported as medians (interquartile range); Kruskal-Wallis test used. Groups with different letter superscripts are significantly different (P < 0.05) from each other. 

Lipids, FFAs, and liver function tests

Fasting plasma total and LDL cholesterol and total FFAs were comparable among the three groups. However, relative to LC, HDL cholesterol levels were 38% lower (P < 0.001) and plasma triglycerides (TGs) 2-fold higher (P < 0.001) in OCs but not in PWS subjects (Table 3). The ratio of TG to HDL was more than 2.5 fold higher in OCs than in LC (P < 0.05).

Table 3.

Fasting lipids and FFA and liver function measurements in PWS, BMI-matched controls (OCs), and LCs

PWS (n = 14) OC (n = 14) LC (n = 14)
Cholesterol (mg/dl) 132.00 (118.50, 142.50)a 123.50 (115.00, 154.00)a 124.00 (110.00, 138.00)a
LDL (mg/dl) 78.50 (73.10, 88.85)a 85.95 (66.00, 107.40)a 66.50 (63.10, 87.50)a
HDL (mg/dl) 49.1 (34.7, 52.6)a,b 32.4 (29.3, 44.2)a 52.6 (43.3, 68.1)b
TG (mg/dl) 55.5 (28.0, 73.5)a,b 80.0 (67.0, 104.0)a 40.0 (36.0, 53.0)b
TG to HDL ratio 1.3 (0.5, 2.4)a,b 2.3 (1.7, 3.3)a 0.8 (0.6, 1.0)b
Total FFA (mmol/liter) 0.43 (0.37, 0.63)a 0.35 (0.26, 0.44)a 0.28 (0.19, 0.450)a
ALT (U/liter) 21.80 (16.50, 28.00)a 23.00 (21.00, 27.50)a 19.00 (17.00, 24.00)a
AST (U/liter) 27.80 (24.50, 36.00)a 25.80 (20.00, 32.50)a 32.00 (27.00, 40.00)a
AST to ALT ratio 1.3 (0.9, 1.8)a 1.0 (0.8, 1.4)a 1.5 (1.4, 1.8)a
Open in a new tab

Data reported as medians (interquartile range); Kruskal-Wallis test used. Groups with different letter superscripts are significantly different (P < 0.05) from each other. 

No differences in AST and ALT levels were observed among the three groups. However, in females analyzed separately, ALT was higher in OCs than in PWS subjects (P = 0.006); AST and AST to ALT ratio were not different. In males analyzed separately, no group differences were observed in measurements of ALT, AST, or AST to ALT ratio. However, a trend toward a relatively lower AST to ALT ratio was seen in the OCs compared with PWS subjects or LC (P = 0.06).

Correlations of HOMA-IR to other metabolic parameters

We analyzed the correlations of HOMA-IR with age, BMI, BMI Z-score, glucose, insulin, total adiponectin, HMW adiponectin, HMW to total adiponectin, leptin, total adiponectin to leptin, total GLP-1, TNF-α, CRP, PAI-1, IL-6, IL-12p70, IL-1b, IL-10, IL-2, IL-12p40, IL-8, IFNγ, IL-18, MCP1, resistin, cholesterol, LDL, HDL, TG, TG to HDL ratio, total FFAs, ALT, AST, and AST to ALT ratio in PWS, BMI-matched, and lean control children (Table 4). In PWS children, HOMA-IR correlated positively with age (P < 0.05), BMI (P < 0.01), BMI Z-score (P < 0.01), insulin (P < 0.01), GLP-1 (P < 0.01), CRP (P < 0.05), and PAI-1 (P < 0.05); HOMA-IR correlated negatively in the PWS group with total adiponectin, HMW adiponectin, HMW to total adiponectin ratio, and total adiponectin to leptin ratio (P < 0.01 for all). In OC children, HOMA-IR correlated positively with BMI (P < 0.01), BMI Z-score (P < 0.01), insulin (P < 0.01), leptin (P < 0.01), and CRP (P < 0.05); HOMA-IR correlated negatively in the OC group with total adiponectin to leptin ratio (P < 0.05) and HDL (P < 0.01). In LC, fewer significant correlations were demonstrated; HOMA-IR correlated positively only with age (P < 0.05), insulin (P < 0.01), and total GLP-1 (P < 0.01).

Table 4.

Spearman correlations of HOMA-IR to other metabolic parameters in PWS, BMI-matched controls (OCs), and LC

PWS children HOMA-IR OC children HOMA-IR LC HOMA-IR
Age 0.58a 0.35 0.56a
BMI 0.68b 0.88b 0.51
BMI Z-score 0.65b 0.81b 0.16
Glucose 0.02 0.51 0.06
Insulin 0.97b 0.96b 0.99b
Total adiponectin −0.80b −0.31 0.11
HMW adiponectin (ng/ml) −0.80b −0.36 −0.14
HMW to total adiponectin ratio −0.74b −0.31 −0.20
Leptin (ng/ml) 0.48 0.77b 0.06
Total adiponectin to leptin ratio −0.70b −0.63a −0.2
Total GLP-1 (pg/ml) 0.87b −0.10 0.68b
TNF-α (pg/ml) 0.13 −0.48 0.13
CRP (mg/liter) 0.64a 0.54a −0.16
PAI-1 (pg/ml) 0.70a 0.02 0.20
IL-6 (pg/ml) 0.38 −0.25 0.39
IL-12p70 (pg/ml) 0.15 −0.01 −0.16
IL-1b (pg/ml) 0.27 0.18 0.06
IL-10 (pg/ml) 0.23 −0.44 −0.17
IL-2 (pg/ml) 0.17 −0.19 0.50
IL-12p40 (pg/ml) 0.44 −0.26 0.14
IL-8 (pg/ml) 0.36 −0.28 −0.05
IFNγ (pg/ml) 0.07 −0.04 0.48
IL-18 (pg/ml) 0.30 −0.21 −0.00
MCP-1 (pg/ml) 0.34 −0.21 0.14
Resistin (pg/ml) −0.07 0.31 0.00
Cholesterol (mg/dl) 0.02 −0.29 −0.24
LDL (mg/dl) 0.04 −0.23 −0.04
HDL (mg/dl) −0.30 −0.67b −0.18
TG (mg/dl) 0.27 0.13 −0.04
TG to HDL ratio 0.32 0.32 0.06
Total FFAs (mmol/liter) −0.25 0.45 −0.12
ALT (U/liter) 0.37 0.15 −0.15
AST (U/liter) 0.15 0.27 −0.23
AST to ALT ratio −0.27 0.43 0.13
Open in a new tab
a

P < 0.05. 

b

P < 0.01. 

Discussion

This is the first study to compare systematically the metabolic profiles of hormones, cytokines, lipids, and liver function tests in PWS subjects and OCs relative to lean age- and gender-matched children. Leptin levels were comparable in PWS and OC, suggesting a comparable degree of adiposity. However, insulin concentrations and HOMA-IR were lower in PWS compared with OCs and similar to those in LC, suggesting that PWS children are more insulin sensitive than OCs and similar to LC. Furthermore, total adiponectin, HMW adiponectin, and the HMW to total adiponectin ratio were all significantly higher in the PWS subjects compared with BMI-matched controls. Surprisingly, the degree of elevation in adiponectin rivaled that seen in the lean children, consistent again with the increased insulin sensitivity demonstrated in both PWS and lean children compared with OCs. Similarly, the findings of elevated CRP, IL-6, and TGs and the relatively low levels of HDL in OC but not in PWS subjects also suggest a higher degree of insulin sensitivity in PWS relative to body weight. On the other hand, PAI-1 levels were elevated in both OC and PWS subjects relative to LC. No differences were observed in total GLP-1, MCP-1, TNFα, IL-2, IL-8, IL-10, IL-12p40, IL-18, resistin, total and LDL cholesterol, AST, or ALT among the three groups. In PWS children, HOMA-IR correlated positively with age (P < 0.05), BMI (P < 0.01), BMI Z-score (P < 0.01), insulin (P < 0.01), GLP-1 (P < 0.01), CRP (P < 0.05), and PAI-1 (P < 0.05); HOMA-IR correlated negatively in the PWS group with total adiponectin, HMW adiponectin, HMW to total adiponectin ratio, and total adiponectin to leptin ratio (P < 0.01 for all).

Our results are similar to those of other researchers who found a lower degree of insulin resistance and higher insulin sensitivity in PWS children than in equally obese children who do not have PWS (17,18,19). Most recently Sohn et al. (20) have reported increased insulin sensitivity and expression of the adiponectin receptor, decreased expression of IL-6 in peripheral blood mononuclear cells, and decreased TGs in PWS children compared with obese children; TNFα was not significantly different between groups.

There are several potential mechanisms that might account for the higher insulin sensitivity in PWS compared with OCs. First, the PWS children have a unique body composition that is demonstrable, even in underweight infants: increased percent body fat with preferential deposition in sc rather than visceral adipose tissue (10,11,21,22). Because visceral fat plays a key role in determination of insulin sensitivity, the selective deposition of sc fat would confer a greater degree of insulin sensitivity in PWS children. Eiholzer et al. (22) has also raised the possibility that even young underweight children with PWS do not sense fat stores by the hypothalamic so-called adipostat correctly because they may be leptin resistant. However, this theory does not seem plausible because leptin resistance is thought to predispose to peripheral insulin resistance.

Second, increases in total adiponectin, a preferential increase in the HMW form of adiponectin, and increased HMW to total adiponectin ratio might contribute to the increased insulin sensitivity in PWS. Plasma adiponectin levels correlate inversely with obesity and positively with insulin sensitivity in both healthy subjects and subjects with diabetes. Furthermore, plasma adiponectin concentrations are reduced in obese adolescents with high visceral adipose tissue stores compared with those with lower visceral adipose tissue (23). Negative correlations between concentrations of adiponectin and glucose, insulin, and TG levels, BMI, percentage body fat, and HOMA-IR have been demonstrated in non-PWS children (24).

Finally, in normal puberty there is a physiological increase in insulin resistance, related to increased GH secretion and IGF-I production. Therefore, it is possible that the GH deficiency associated with PWS contributes partially to heightened insulin sensitivity. However, in our cohort, nine of 14 PWS children were treated with a stable dose of GH, and IGF-I values were comparable between the PWS and OC groups; this is therefore not likely the mechanism of increased insulin sensitivity found in PWS.

Interestingly, some reports also demonstrate decreased insulin secretion from pancreatic β-cells in children with PWS (17). Some authors theorize that this is due to decreased vagal parasympathetic efferent tone to the pancreas, which exerts control on normal insulin secretion (6). Supportive of this theory is the finding in PWS of decreased pancreatic polypeptide secretion, which reflects alterations in autonomic nervous system function (25).

Higher levels of total adiponectin and a preferential increase in the HMW form in PWS children and adults have been reported by our group and others (7,11,12,26,27). However, few of the studies in children included a lean comparison group. In contrast to our findings, studies in PWS adults report that total adiponectin is higher than in OCs but lower than in lean subjects (26,27). However, another study in PWS children reported an increase in total adiponectin in PWS children comparable with sex- and age-matched controls (12). These results suggest that adiponectin levels might decline with age in PWS. In support of this theory is that total and HMW adiponectin levels were approximately twice as high in young infants with PWS (median age 22 months) as in older children, consistent with the fact that insulin sensitivity declines with age (13). Additional studies are needed to address the possible association of adiponectin with age.

Our study found relatively higher concentrations of CRP and IL-6 in OC children than in PWS or lean subjects. Cytokines released by the adipose tissue, such as adiponectin, resistin, IL-6, CRP, and TNF-α, link adiposity with insulin resistance in children (28). TNF-α and IL-6 act to inhibit insulin signaling in the adipose tissue.

In contrast to CRP and IL-6, PAI-I levels were increased in both PWS children and OCs relative to LC. PAI-1 is an inhibitor of the fibrinolytic system. Therefore, increased concentrations of circulating PAI-1 lead to a state of hypofibrinolysis, resulting in impaired removal of vascular thrombi. PAI-1activity has been associated with coronary heart disease, type 2 diabetes, and obesity (29). It is now known that human adipose tissue secretes and expresses PAI-1; specifically, it is the stromal compartment consisting of macrophages infiltrating adipose tissue that express PAI-1 (30). Some authors believe that PAI-1 levels depend not strictly on absolute fat mass but rather reflect fat redistribution and therefore ectopic fat storage (i.e. liver fat) (31). Indeed, several studies indicate that plasma PAI-1 levels are closely correlated with the visceral fat area determined by computed tomography scan in humans (32,33). However, a study of adipose tissue biopsies of 22 obese individuals (BMI 33.1–60.2 kg/m2) showed that sc adipose tissue secreted greater amounts of PAI-1 and had higher PAI-1 mRNA levels than visceral adipose tissue (34). PAI-1 expression correlated strongly with the cell size of sc adipose tissue. The storage of fat in sc stores in PWS may explain the increased concentrations of PAI-1 in both PWS children and OCs relative to lean subjects. A recent study of 12 young adults with PWS, 12 obese subjects matched for percentage body fat and central abdominal fat mass, and 10 healthy normal-weight subjects found that the PWS subjects demonstrated a similar degree of insulin resistance but increased low-grade inflammation (higher IL-6 levels and increased neutrophil activation markers, CD66b and CD11b) compared with obese controls (35). Reasons for the discrepancies between these results and our findings include: 1) the authors selected obese and PWS groups for degree of abdominal adiposity; 2) many PWS subjects had obstructive sleep apnea, which is known to increase immune cell activation; 3) none of their PWS subjects had ever been treated with GH; and 4) all subjects were adults in their late 20s. Additionally, another study by Höybye (36) showed that IL-6 and CRP levels were elevated in PWS adults compared with values given as normal range for the assays; TNF-α levels were normal in agreement with our study. Limitations to the Höybye study include lack of an appropriate control group. Finally, a study in seven adults with PWS compared with seven obese non-PWS and seven lean subjects showed higher plasma concentrations of CRP, complement component C3, IL-18, and IL-6 in PWS adults (37). In contrast, we found no increases in IL-18 in children with PWS. It is possible that the inflammatory profile of adults with PWS differs from that of children included in this study.

A subset of our patients (nine of 14) was on stable doses of GH; this represents a possible limitation of the study. However, other investigations showed that 12 months of GH therapy did not reduce IL-6, TNF-α, or CRP levels in adults with PWS; therefore, GH is unlikely to have significantly impacted the levels of inflammatory cytokines in our study (36). Another study shows that 12 months of GH therapy did not significantly increase adiponectin in adults with PWS; therefore, GH is unlikely to have significantly impacted the levels of adiponectin in our study (37). Pubertal status is not available on all subjects; this is another limitation to our study. However, hypogonadism has been shown to be associated with increased systemic inflammation and features of metabolic syndrome (38). Therefore, hypogonadism is unlikely to explain the relatively protected metabolic phenotype of the PWS subjects in our study.

Our findings suggest possible mechanisms for the relative paucity of metabolic complications in obese PWS children and adults compared with those of equal adiposity. For example, Goldstone et al. (10) has found that PWS adult women had lower amounts of visceral fat and were more insulin sensitive than women with exogenous obesity (findings confirmed now in larger cohorts of PWS subjects), and a small number of studies have demonstrated lower incidence of metabolic comorbidities associated with PWS (10,17). Other small studies have reported deaths in various cohorts of adults with PWS; uniparental disomy is reported by some as an independent risk factor of death. Reported causes of death have included cardiorespiratory failure, gastric rupture, sudden collapse, profound hypoglycemia, and stroke (39). The oldest PWS individual reported in the literature is a 71-yr-old obese and hypertensive female (without diabetes) who died of acute cardiorespiratory failure after a viral upper respiratory tract infection (40). Further long-term studies are needed to determine whether the increased insulin sensitivity relative to total body fat demonstrated in PWS children will translate into long-term clinical benefit.

Acknowledgments

We acknowledge Ms. Juanita Cuffee (study coordinator) for help with recruitment of study subjects. We thank all the families and children who participated in these studies. We acknowledge additional support from the Alberta Diabetes Institute and the Women & Children’s Health Research Institute at University of Alberta.

Footnotes

This work was supported by Grant 1K23-RR-021979 (to A.M.H.). M.J.M. and C.B.N. was supported by the Sarah W. Stedman Nutrition and Metabolism Center. This study was conducted through the Duke University Medical Center, General Clinical Research Center Grant MO1-RR-30 (National Center for Research Resources, Clinical Research Centers Program, National Institutes of Health) and additional funding from the Sarah W. Stedman Nutrition and Metabolism Center.

Disclosure Summary: The authors have nothing to disclose.

First Published Online October 20, 2010

Abbreviations: ALT, Alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CRP, C-reactive protein; FFA, free (nonesterified) fatty acid; GLP-1, glucagon-like peptide-1; HDL, high-density lipoprotein; HMW, high molecular weight; HOMA-IR, homeostasis model assessment insulin resistance index; IFN, interferon; LC, lean children; LDL, low-density lipoprotein; MCP-1, macrophage chemoattractant protein-1; OC, obese control; PAI-1, plasminogen activator inhibitor-1; PWS, Prader-Willi syndrome; TG, triglyceride.

References

  1. Bray GA, Dahms WT, Swerdloff RS, Fiser RH, Atkinson RL, Carrel RE 1983 The Prader-Willi syndrome: a study of 40 patients and a review of the literature. Medicine (Baltimore) 62:59–80 [PubMed] [Google Scholar]
  2. Runte M, Hüttenhofer A, Gross S, Kiefmann M, Horsthemke B, Buiting K 2001 The IC-SNURF-SNRPN transcript serves as a host for multiple small nucleolar RNA species and as an antisense RNA for UBE3A. Hum Mol Genet 10:2687–2700 [DOI] [PubMed] [Google Scholar]
  3. Muscatelli F, Abrous DN, Massacrier A, Boccaccio I, Le Moal M, Cau P, Cremer H 2000 Disruption of the mouse Necdin gene results in hypothalamic and behavioral alterations reminiscent of the human Prader-Willi syndrome. Hum Mol Genet 9:3101–3110 [DOI] [PubMed] [Google Scholar]
  4. Mercer RE, Kwolek EM, Bischof JM, van Eede M, Henkelman RM, Wevrick R 2009 Regionally reduced brain volume, altered serotonin neurochemistry, and abnormal behavior in mice null for the circadian rhythm output gene Magel2. Am J Med Genet B Neuropsychiatr Genet 150B:1085–1099 [DOI] [PubMed] [Google Scholar]
  5. Lindgren AC, Hagenäs L, Ritzén EM 1999 Growth hormone treatment of children with Prader-Willi syndrome: effects on glucose and insulin homeostasis. Swedish National Growth Hormone Advisory Group. Horm Res 51:157–161 [DOI] [PubMed] [Google Scholar]
  6. Schuster DP, Osei K, Zipf WB 1996 Characterization of alterations in glucose and insulin metabolism in Prader-Willi subjects. Metabolism 45:1514–1520 [DOI] [PubMed] [Google Scholar]
  7. Haqq AM, Muehlbauer M, Svetkey LP, Newgard CB, Purnell JQ, Grambow SC, Freemark MS 2007 Altered distribution of adiponectin isoforms in children with Prader-Willi syndrome (PWS): association with insulin sensitivity and circulating satiety peptide hormones. Clin Endocrinol (Oxf) 67:944–951 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Zipf WB, O'Dorisio TM, Cataland S, Dixon K 1983 Pancreatic polypeptide responses to protein meal challenges in obese but otherwise normal children and obese children with Prader-Willi syndrome. J Clin Endocrinol Metab 57:1074–1080 [DOI] [PubMed] [Google Scholar]
  9. Sareen C, Ruvalcaba RH, Kelley VC 1975 Some aspects of carbohydrate metabolism in Prader-Willi syndrome. J Ment Defic Res 19:113–119 [DOI] [PubMed] [Google Scholar]
  10. Goldstone AP, Thomas EL, Brynes AE, Bell JD, Frost G, Saeed N, Hajnal JV, Howard JK, Holland A, Bloom SR 2001 Visceral adipose tissue and metabolic complications of obesity are reduced in Prader-Willi syndrome female adults: evidence for novel influences on body fat distribution. J Clin Endocrinol Metab 86:4330–4338 [DOI] [PubMed] [Google Scholar]
  11. Kennedy L, Bittel DC, Kibiryeva N, Kalra SP, Torto R, Butler MG 2006 Circulating adiponectin levels, body composition and obesity-related variables in Prader-Willi syndrome: comparison with obese subjects. Int J Obes 30:382–387 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Festen DA, van Toorenenbergen A, Duivenvoorden HJ, Hokken-Koelega AC 2007 Adiponectin levels in prepubertal children with Prader-Willi syndrome before and during growth hormone therapy. J Clin Endocrinol Metab 92:1549–1554 [DOI] [PubMed] [Google Scholar]
  13. Haqq AM, Grambow SC, Muehlbauer M, Newgard CB, Svetkey LP, Carrel AL, Yanovski JA, Purnell JQ, Freemark M 2008 Ghrelin concentrations in Prader-Willi syndrome (PWS) infants and children: changes during development. Clin Endocrinol (Oxf) 69:911–920 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Joo EY, Hong SB, Sohn YB, Kwak MJ, Kim SJ, Choi YO, Kim SW, Paik KH, Jin DK 2010 Plasma adiponectin level and sleep structures in children with Prader-Willi syndrome. J Sleep Res 19:248–254 [DOI] [PubMed] [Google Scholar]
  15. Hotamisligil GS 2006 Inflammation and metabolic disorders. Nature 444:860–867 [DOI] [PubMed] [Google Scholar]
  16. Gupta A, Ten S, Anhalt H 2005 Serum levels of soluble tumor necrosis factor-α receptor 2 are linked to insulin resistance and glucose intolerance in children. J Pediatr Endocrinol Metab 18:75–82 [DOI] [PubMed] [Google Scholar]
  17. Krochik AG, Ozuna B, Torrado M, Chertkoff L, Mazza C 2006 Characterization of alterations in carbohydrate metabolism in children with Prader-Willi syndrome. J Pediatr Endocrinol Metab 19:911–918 [DOI] [PubMed] [Google Scholar]
  18. Eiholzer U, Stutz K, Weinmann C, Torresani T, Molinari L, Prader A 1998 Low insulin, IGF-I and IGFBP-3 levels in children with Prader-Labhart-Willi syndrome. Eur J Pediatr 157:890–893 [DOI] [PubMed] [Google Scholar]
  19. Talebizadeh Z, Butler MG 2005 Insulin resistance and obesity-related factors in Prader-Willi syndrome: comparison with obese subjects. Clin Genet 67:230–239 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Sohn YB, Kwak MJ, Kim SJ, Park SW, Kim CH, Kim MY, Kwon EK, Paik KH, Jin DK 2010 Correlation of adiponectin receptor expression with cytokines and insulin sensitivity in growth hormone (GH)-treated children with Prader-Willi syndrome and in non-GH-treated obese children. J Clin Endocrinol Metab 95:1371–1377 [DOI] [PubMed] [Google Scholar]
  21. Eiholzer U, L'allemand D, Schlumpf M, Rousson V, Gasser T, Fusch C 2004 Growth hormone and body composition in children younger than 2 years with Prader-Willi syndrome. J Pediatr 144:753–758 [DOI] [PubMed] [Google Scholar]
  22. Eiholzer U, Blum WF, Molinari L 1999 Body fat determined by skinfold measurements is elevated despite underweight in infants with Prader-Labhart-Willi syndrome. J Pediatr 134:222–225 [DOI] [PubMed] [Google Scholar]
  23. Bacha F, Saad R, Gungor N, Arslanian SA 2004 Adiponectin in youth: relationship to visceral adiposity, insulin sensitivity, and β-cell function. Diabetes Care 27:547–552 [DOI] [PubMed] [Google Scholar]
  24. Jeffery AN, Murphy MJ, Metcalf BS, Hosking J, Voss LD, English P, Sattar N, Wilkin TJ 2008 Adiponectin in childhood. Int J Pediatr Obes 3:130–140 [DOI] [PubMed] [Google Scholar]
  25. Berntson GG, Zipf WB, O'Dorisio TM, Hoffman JA, Chance RE 1993 Pancreatic polypeptide infusions reduce food intake in Prader-Willi syndrome. Peptides 14:497–503 [DOI] [PubMed] [Google Scholar]
  26. Höybye C, Bruun JM, Richelsen B, Flyvbjerg A, Frystyk J 2004 Serum adiponectin levels in adults with Prader-Willi syndrome are independent of anthropometrical parameters and do not change with GH treatment. Eur J Endocrinol 151:457–461 [DOI] [PubMed] [Google Scholar]
  27. Caixàs A, Giménez-Palop O, Giménez-Pérez G, Potau N, Berlanga E, González-Glemente JM, Arroyo J, Laferrère B, Mauricio D 2006 Postprandial adiponectin levels are unlikely to contribute to the pathogenesis of obesity in Prader-Willi syndrome. Horm Res 65:39–45 [DOI] [PubMed] [Google Scholar]
  28. Kelly AS, Steinberger J, Kaiser DR, Olson TP, Bank AJ, Dengel DR 2006 Oxidative stress and adverse adipokine profile characterize the metabolic syndrome in children. J Cardiometab Syndr 1:248–252 [DOI] [PubMed] [Google Scholar]
  29. Hamsten A 1995 Hemostatic function and coronary artery disease. N Engl J Med 332:677–678 [DOI] [PubMed] [Google Scholar]
  30. Bastelica D, Mavri A, Verdierl M, Berthet B, Juhan-Vague I, Alessi MC 2002 Relationships between fibrinolytic and inflammatory parameters in human adipose tissue: strong contribution of TNFα receptors to PAI-1 levels. Thromb Haemost 88:481–487 [PubMed] [Google Scholar]
  31. Alessi MC, Bastelica D, Mavri A, Morange P, Berthet B, Grino M, Juhan-Vague I 2003 Plasma PAI-1 levels are more strongly related to liver steatosis than to adipose tissue accumulation. Arterioscler Thromb Vasc Biol 23:1262–1268 [DOI] [PubMed] [Google Scholar]
  32. Shimomura I, Funahashi T, Takahashi M, Maeda K, Kotani K, Nakamura T, Yamashita S, Miura M, Fukuda Y, Takemura K, Tokunaga K, Matsuzawa Y 1996 Enhanced expression of PAI-1 in visceral fat: possible contributor to vascular disease in obesity. Nat Med 2:800–803 [DOI] [PubMed] [Google Scholar]
  33. Janand-Delenne B, Chagnaud C, Raccah D, Alessi MC, Juhan-Vague I, Vague P 1998 Visceral fat as a main determinant of plasminogen activator inhibitor 1 level in women. Int J Obes Relat Metab Disord 22:312–317 [DOI] [PubMed] [Google Scholar]
  34. Eriksson P, Van Harmelen V, Hoffstedt J, Lundquist P, Vidal H, Stemme V, Hamsten A, Arner P, Reynisdottir S 2000 Regional variation in plasminogen activator inhibitor-1 expression in adipose tissue from obese individuals. Thromb Haemost 83:545–548 [PubMed] [Google Scholar]
  35. Viardot A, Sze L, Purtell L, Sainsbury A, Loughnan G, Smith E, Herzog H, Steinbeck K, Campbell LV 2010 Prader-Willi syndrome is associated with activation of the innate immune system independently of central adiposity and insulin resistance. J Clin Endocrinol Metab 95:3392–3399 [DOI] [PubMed] [Google Scholar]
  36. Höybye C 2006 Inflammatory markers in adults with Prader-Willi syndrome before and during 12 months growth hormone treatment. Horm Res 66:27–32 [DOI] [PubMed] [Google Scholar]
  37. Caixàs A, Giménez-Palop O, Broch M, Vilardell C, Megía A, Simón I, Giménez-Pérez G, Mauricio D, Vendrell J, Richart C, González-Clemente JM 2008 Adult subjects with Prader-Willi syndrome show more low-grade systemic inflammation than matched obese subjects. J Endocrinol Invest 31:169–175 [DOI] [PubMed] [Google Scholar]
  38. Laaksonen DE, Niskanen L, Punnonen K, Nyyssönen K, Tuomainen TP, Salonen R, Rauramaa R, Salonen JT 2003 Sex hormones, inflammation and the metabolic syndrome: a population-based study. Eur J Endocrinol 149:601–608 [DOI] [PubMed] [Google Scholar]
  39. Einfeld SL, Kavanagh SJ, Smith A, Evans EJ, Tonge BJ, Taffe J 2006 Mortality in Prader-Willi syndrome. Am J Ment Retard 111:193–198 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Carpenter PK 1994 Prader-Willi syndrome in old age. J Intellect Disabil Res 38(Pt 5):529–531 [DOI] [PubMed] [Google Scholar]

Articles from The Journal of Clinical Endocrinology and Metabolism are provided here courtesy of The Endocrine Society

RESOURCES