help button home button Endocrine Society JCEM
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Journal of Clinical Endocrinology & Metabolism, doi:10.1210/jc.2006-2855
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
92/6/2046    most recent
Author Manuscript (PDF)
Right arrow Submit a related Letter to the Editor
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Misra, M.
Right arrow Articles by Klibanski, A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Misra, M.
Right arrow Articles by Klibanski, A.
Related Collections
Right arrow Calcium and Bone Metabolism
Right arrow Metabolism
Right arrow Neuroendocrinology and Pituitary
The Journal of Clinical Endocrinology & Metabolism Vol. 92, No. 6 2046-2052
Copyright © 2007 by The Endocrine Society

Relationships between Serum Adipokines, Insulin Levels, and Bone Density in Girls with Anorexia Nervosa

Madhusmita Misra, Karen K. Miller, Jennalee Cord, Rajani Prabhakaran, David B. Herzog, Mark Goldstein, Debra K. Katzman and Anne Klibanski

Neuroendocrine Unit (M.M., K.K.M., J.C., R.P., A.K.), Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts 02114; Harris Center (D.B.H.), Massachusetts General Hospital, Boston, Massachusetts 02114; Division of Adolescent Medicine (M.G.), MassGeneral Hospital for Children and Harvard Medical School, Boston, Massachusetts 02114; and Division of Adolescent Medicine (D.K.K.), Department of Pediatrics, Hospital for Sick Kids, Toronto, Canada M5G 1X8

Address all correspondence and requests for reprints to: Madhusmita Misra, M.D., BUL 457, Neuroendocrine Unit, Massachusetts General Hospital, 55 Fruit Street, Boston, Massachusetts 02114. E-mail: mmisra{at}partners.org.


   Abstract
Top
 Abstract
Introduction
 Subjects and Methods
 Results
 Discussion
 References
 
Background: Adolescents with anorexia nervosa (AN) have low bone mineral density (BMD). Adipokines and insulin play an important role in bone metabolism in healthy individuals. However, their association with bone metabolism in AN is unknown.

Objective: The aim of the study was to determine whether adipokines and insulin are independently associated with measures of BMD in adolescents with AN and controls.

Design/Methods: Levels of adiponectin and insulin, fasting and after oral glucose, were evaluated in 17 AN patients and 19 controls (age, 12–18 yr), in whom hormonal parameters [GH, IGF-I, cortisol, estradiol, leptin, ghrelin, and peptide YY (PYY)] had been previously determined. Body composition, bone mineral content, and BMD at the lumbar spine, hip, femoral neck, and total body were assessed by dual energy x-ray absorptiometry. Two bone formation and bone resorption markers were examined.

Setting: The study was conducted at a General Clinical Research Center.

Results: Adiponectin differed between AN subjects and controls after controlling for fat mass and decreased in both after oral glucose (P = 0.02 and 0.07). On regression modeling, independent associations were observed of: 1) body mass index and adiponectin with lumbar spine bone mineral apparent density Z-scores (r2 = 0.45); 2) lean mass, PYY, and ghrelin with hip Z-scores (r2 = 0.55); 3) adiponectin and lean mass with femoral neck-bone mineral apparent density Z-scores (r2 = 0.34); and 4) lean mass, PYY, GH, and ghrelin with total body-bone mineral content/height Z-scores (r2 = 0.64), for the combined group. Adiponectin was also independently associated with BMD, and insulin was associated with bone turnover markers in the groups considered separately.

Conclusions: Adiponectin contributes significantly to the variability of bone density, and insulin contributes to bone turnover markers in adolescent girls.


   Introduction
Top
 Abstract
 Introduction
Subjects and Methods
 Results
 Discussion
 References
 
ADOLESCENT GIRLS WITH anorexia nervosa (AN) are at risk for low bone mineral density (BMD) (1), which is associated with hypogonadism, low IGF-I (1, 2), high cortisol and peptide YY (PYY) (3, 4) levels, and decreased lean mass (1, 5). However, these hormonal and body composition alterations account for only 20–60% of the variability of BMD Z-scores in AN and healthy adolescents (3, 5, 6), and other factors that contribute to the variability of BMD are unknown.

An important and striking feature of AN is very low fat mass. Although the relationship between lean mass and BMD is well known (1, 7), the relationship between fat-dependent factors and BMD is less clear. Adipokines, such as adiponectin, leptin, and IL-6, affect bone metabolism (8, 9, 10, 11, 12, 13, 14, 15) and are altered in AN. In addition, insulin, whose secretion is directly related to fat mass, exerts anabolic effects on bone (16). Fat mass is associated positively with leptin and insulin but inversely with adiponectin (reviewed in Ref. 17). Girls with AN have low insulin and leptin levels (18). However, we have reported normal adiponectin in adolescents with AN (18), and both high (19, 20) and low (21) adiponectin levels have been reported in adults with AN. Adiponectin receptors are expressed on osteoblasts and osteoclasts (8, 10, 11, 12), and high adiponectin levels are associated with low BMD in healthy adults (22, 23). A role for adiponectin in suppressing osteoprotegerin (OPG) and increasing expression of receptor activator of nuclear factor-{kappa}B (RANK) ligand (RANKL) has been described, suggesting that high adiponectin levels may cause increased osteoclastic activity and low BMD (9). Adiponectin, however, is also reported to increase osteoblastic activity and decrease osteoclastic activity in animals, which should be associated with increased bone formation and decreased resorption (10, 11, 12). Therefore, although insulin and adiponectin play an important role in maintaining normal bone mass, it is unknown whether they are independently associated with bone in AN.

We hypothesized that high adiponectin, low leptin, and low insulin levels in adolescent AN girls would be independently associated with low BMD and bone turnover markers. Therefore, we determined associations of BMD and bone turnover markers with adipokines (adiponectin, leptin, IL-6) and insulin in AN subjects and controls, in whom we have previously examined other variables associated with bone metabolism such as lean mass, GH-IGF-I, cortisol, estradiol, and PYY (2, 3, 4, 5).


   Subjects and Methods
Top
 Abstract
 Introduction
 Subjects and Methods
Results
 Discussion
 References
 
Subject selection and experimental protocol

We enrolled 17 adolescent girls meeting DSM-IV criteria for AN and 19 healthy girls. Girls with AN did not differ from controls for chronological (16.5 ± 1.5 vs. 15.5 ± 1.8 yr) or bone age (16.0 ± 1.4 vs. 15.8 ± 2.0 yr). Duration since diagnosis ranged from 1–36 months. AN girls were referred by eating disorder providers in the New England area, and healthy controls recruited through mailings to pediatricians. Our Institutional Review Board approved the study, and informed consent and assent were obtained.

Subjects were admitted to our General Clinical Research Center for frequent sampling for GH, cortisol, ghrelin, and leptin every 30 min (2000–0800 h). Area under the curve was calculated using Cluster (http://mljohnson.pharm.virginia.edu/downloads.html). Fasting blood was drawn for adiponectin, insulin, IGF-I, estradiol, IL-6, osteocalcin (OC), carboxyterminal propeptide of type 1 procollagen (PICP), and PYY. A second morning urine sample was obtained for N-telopeptide (NTX) and deoxypyridinoline (DPD). Subjects were administered 100 g oral glucose, and insulin and adiponectin were measured at 0, 30, and 60 min. Body composition and BMD at the lumbar spine (LS), hip, and total body (TB) were determined by dual energy x-ray absorptiometry (Hologic 4500; Hologic Corp., Waltham, MA). Bone mineral apparent density (BMAD) was calculated for the LS and femoral neck (FN) using established formulae (24) as a surrogate measure for volumetric BMD. Z-scores for BMAD and TB bone mineral content/height (TB-BMC/ht) were calculated using the applet of Bachrach, Hastie, and Narasimhan (http://www-stat-class.stanford.edu/pediatric-bones/). AN girls were reexamined after a 10% gain in body mass index (BMI) (n = 9).

IGF-I, estradiol, IL-6, PYY, adiponectin, insulin, and bone turnover markers (2, 3, 5, 6, 25, 26); overnight data for GH, cortisol, leptin, and ghrelin (2, 4, 18, 27); and measures of LS-BMD, LS-BMAD, and hip BMD (2, 4, 5, 6, 25) have been previously reported. We have not previously reported adiponectin levels after oral glucose and after adjusting for fat mass and BMI, or measures of FN-BMAD, TB-BMC/ht, and associations with body composition and hormonal parameters, or the relationship of adiponectin and insulin with BMD and bone markers.

Biochemical assessment

We used RIA to measure adiponectin [Linco Diagnostics, Inc., St. Charles, MO; lowest detectable concentration, 0.001 mg/liter; coefficient of variation (CV) 6.4–8.4%], insulin (Linco Research Inc., St. Charles, MO; sensitivity, 2 µIU/ml; CV, 2.2–4.4%), leptin (Linco Diagnostics, Inc.; sensitivity, 0.5 ng/ml; CV, 3.4–8.3%), ghrelin (Phoenix Pharmaceuticals, Belmont, CA, sensitivity, 2 pg/ml; CV, 9.9–10.5%), cortisol (Diagnostic Products Corp., Los Angeles, CA; sensitivity, 0.21 µg/dl; CV, 2.5–4.1%), estradiol (Diagnostic Systems Laboratories, Inc., Webster, TX; limit of detection, 2.2 pg/ml; CV, 6.5–8.9%), and PICP (Diasorin, Inc., Stillwater, MN; limit of detection, 25 ng/ml; CV, 1.3–3.8%). PYY was measured using a RIA (Phoenix Pharmaceuticals Inc., Belmont, CA), with 100% cross-reactivity with 1–36 and 3–36 human PYY. Immunoradiometric assay (Nichols Institute Diagnostics, San Juan Capistrano, CA) was used to measure GH (detection limit, 0.05 ng/ml; CV, 2.4–9.4%), IGF-I (Nichols Institute Diagnostics; detection limit, 30 ng/ml; CV, 3.1–4.6%), and OC (Nichols Institute Diagnostics; minimum detection limit, 0.5 ng/ml; CV, 3.5–5.2%). IL-6 was measured using a sandwich enzyme immunoassay (R&D Systems Inc., Minneapolis, MN; minimum detectable limit, 0.039 pg/ml; CV, 6.9–7.4%). OPG was measured by Amgen, Inc. (Thousand Oaks, CA), using their in-house human endogenous ELISA (minimum detection limit, 2.34 pg/ml; CV, 0.69–2.7%) (measures monomeric and dimeric OPG, and OPG bound to RANKL). We used ELISA to measure NTX (Ostex International, Inc., Seattle, WA; limit of detection, 20 nmol bone collagen equivalent; CV, 5–19%) and DPD (Quidel, Inc., San Diego, CA; limit of detection, 1.1 nmol/liter; CV, 4.3–8.4%). We did not measure SHBG, which could provide estimates of free levels of gonadal steroids. Samples were stored at –80 C until analysis and run in duplicate in a single batch.

Homeostasis model assessment-insulin resistance (HOMA-IR) was calculated [fasting glucose (mmol/liter) x fasting insulin (µU/ml)/22.5]. Cortisol can be converted to SI units (nmol/liter) by dividing by 0.0363.

Statistical methods

Data are presented as mean ± SD and were analyzed using the JMP program (version 4, SAS Institute Inc., Cary, NC). Student t test was used to calculate differences between means. When data were not normally distributed, logarithmic conversions were performed to approximate a normal distribution. Univariate and mixed model stepwise regression analyses (P = 0.15 for entry, and P = 0.10 to leave the model) were used to determine associations with BMD and bone turnover markers. Univariate analyses were used as a preliminary screening test, and because our conclusions were based not on these analyses but on the final regression models, we did not adjust for multiple comparisons. Data are presented for the group as a whole because of the great degree of overlap of adiponectin levels in AN and controls. However, we also describe results of regression modeling within individual groups. Paired t tests were used to compare data from baseline to weight recovery.


   Results
Top
 Abstract
 Introduction
 Subjects and Methods
 Results
Discussion
 References
 
Baseline characteristics

Baseline characteristics have been previously reported (1, 2, 3, 4, 5, 6, 18, 27) and are summarized in Table 1Go. Compared with controls, AN had lower BMI, fat mass, IGF-I, leptin, and estradiol, and higher GH, ghrelin, cortisol, and PYY. Bone density and PICP were lower in AN.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Baseline characteristics in adolescent girls with AN and healthy adolescents

 
Adiponectin and measures of insulin resistance

Adiponectin levels did not differ in AN subjects vs. controls, although adiponectin 60 min after oral glucose trended higher in AN. The ratio of adiponectin to fat mass and BMI was also higher in AN (Table 2Go). After oral glucose, adiponectin decreased in both groups. On paired analysis, adiponectin decreased from 13.5 to 11.2 mg/liter in AN (P = 0.02) and from 11.9 to 8.7 mg/liter in controls (P = 0.07). The absolute decrease in adiponectin did not differ in AN vs. controls (–2.3 ± 3.6 vs. –3.2 ± 7.3 mg/liter). Fasting insulin and HOMA-IR were lower in AN subjects (Table 2Go). No associations were observed between adiponectin and insulin. Adiponectin was weakly associated with lean mass and leptin (r = 0.30; P = 0.07 for both), but not with fat mass or BMI.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Adipokines and measures of insulin resistance in adolescent girls with AN and healthy adolescents 0, 30, and 60 min after ingestion of a 100-g oral glucose load

 
Relationship of adipokines and other hormones with bone density and bone turnover measures

Correlation analyses of adipokines and measures of insulin resistance with BMD for the group as a whole are shown in Table 3Go, and with bone turnover markers in Table 4Go.


View this table:
[in this window]
[in a new window]

  		TABLE 3. Associations of bone density measures with adipokines and measures of insulin resistance

 

View this table:
[in this window]
[in a new window]

  		TABLE 4. Associations of bone turnover markers with adipokines and measures of insulin resistance

 
Adipokines, insulin, and bone metabolism. Adiponectin correlated inversely with LS-BMAD and its Z-scores (Fig. 1Go). Leptin was positively associated with BMD at each site, whereas no associations were noted between IL-6 and BMD. Adipokines were not associated with bone turnover markers. OPG correlated inversely with leptin, but not with adiponectin or insulin. Insulin was directly associated with Z-scores for LS-BMAD, total hip BMD, and TB-BMC/ht, PICP, and NTX/creatinine.


Figure 1
View larger version (8K):
[in this window]
[in a new window]

  		FIG. 1. Relationship between adiponectin levels and LS BMAD and BMAD Z-scores. An inverse association was observed on univariate analysis between adiponectin levels and LS BMAD and BMAD Z-scores for the group as a whole (r = –0.33, P = 0.05; and r = –0.37, P = 0.03, respectively).

 
Other hormones and body composition measures and bone metabolism. Positive associations were observed of: 1) BMI with LS-BMAD, hip BMD, TB-BMC/ht, and their Z-scores (r ≥ 0.36; P ≤ 0.03); 2) lean mass with hip BMD, FN-BMAD, TB-BMC/ht, and their Z-scores (r ≥ 0.35; P ≤ 0.04); 3) IGF-I with Z-scores for LS-BMAD, hip BMD, FN-BMAD, and TB-BMC/ht (r ≥ 0.32; P ≤ 0.05); and 4) estradiol with TB-BMC/ht (r = 0.32; P = 0.02). Ghrelin showed weak positive associations with LS-BMAD, and GH and estradiol with hip BMD (P ≤ 0.10). Inverse associations were observed between: 1) PYY and Z-scores for LS-BMAD, hip BMD, and TB-BMC/ht (r ≤ –0.36; P ≤ 0.05); 2) cortisol and Z-scores for LS-BMAD and TB-BMC/ht (r ≤ –0.38; P ≤ 0.02); and 3) OPG with LS-BMAD Z-scores (r = –0.35; P = 0.04).

BMI and GH were not associated with bone turnover markers. Ghrelin was weakly associated with NTX/creatinine (r = –0.33; P = 0.05), and cortisol with PICP and NTX/creatinine (r ≤ –0.36; P ≤ 0.03). PYY was associated with OC, NTX/creatinine, and DPD/creatinine (r ≤ –0.41; P ≤ 0.02), and estradiol with PICP (r = –0.39; P = 0.02). IGF-I correlated with PICP and OC (r ≥ 0.35; P ≤ 0.04). Because we used regression modeling to determine covariates that were independently associated with BMD and bone turnover markers, details of univariate analyses within the individual groups are not reported.

Regression modeling to determine independent associations with bone density measures

Regression modeling to eliminate confounding variables was performed for the group as a whole and also for the individual groups, with body composition (BMI and lean mass), adiponectin, insulin, and hormones known to affect bone metabolism (leptin, IGF-I, GH, ghrelin, estradiol, cortisol, and PYY) entered into the model.

AN and healthy adolescent girls (Table 5Go). BMI contributed significantly to the variability of LS-BMAD (r2 = 0.36), and BMI and adiponectin to that of LS-BMAD Z-scores (r2 = 0.45). Lean mass, PYY, and ghrelin contributed 36% and 55%, respectively, to the variability of hip BMD and its Z-scores. Lean mass and IGF-I were independently associated with FN-BMAD (r2 = 0.31), and adiponectin and lean mass with FN-BMAD Z-scores (r2 = 0.34). Lean mass, adiponectin, ghrelin, and cortisol were associated with TB-BMC/ht (r2 = 0.80); and lean mass, PYY, and GH were associated with TB-BMC/ht Z-scores (r2 = 0.65). Adiponectin, lean mass, and ghrelin were independently associated with TB-BMD (r2 = 0.60) and adiponectin and lean mass with TB-BMD Z-scores (r2 = 0.47). Positive associations with BMD in the regression model were observed for lean mass, BMI, GH, ghrelin, and estradiol, and inverse associations were observed for PYY and adiponectin.


View this table:
[in this window]
[in a new window]

  		TABLE 5. Regression modeling (stepwise regression) for bone density measures

 
Regression modeling was also performed for bone turnover markers. PICP was associated independently with insulin, GH, and ghrelin (r2 = 0.39); OC with GH and PYY (r2 = 0.24); NTX/creatinine with insulin, GH, estradiol, PYY, and BMI (r2 = 0.59); and DPD/creatinine with ghrelin, lean mass, and PYY (r2 = 0.44).

AN girls. Lean mass and estradiol contributed significantly to the variability of LS-BMAD and its Z-score (r2 = 0.60 and 0.48). Lean mass, PYY, leptin, ghrelin, and estradiol showed independent associations with hip BMD (r2 = 0.84), and lean mass with hip BMD Z-score (r2 = 0.32), FN-BMAD (r2 = 0.22), and its Z-score (r2 = 0.18). Lean mass, ghrelin, estradiol, BMI, PYY, and GH contributed independently to the variability of TB-BMC/Ht (r2 = 0.99), and adiponectin and lean mass to TB-BMC/Ht Z-scores (r2 = 0.63). Lean mass, leptin, and adiponectin were independently associated with TB-BMD (r2 = 0.74), and lean mass and adiponectin with TB-BMD Z-scores (r2 = 0.58). Lean mass, GH, ghrelin, and estradiol were positively associated, whereas adiponectin, PYY, and leptin were inversely associated with BMD measures in this model. For bone turnover markers, on modeling, cortisol and PYY contributed to the variability of PICP (r2 = 0.41); cortisol, GH, and ghrelin to OC (r2 = 0.68); and cortisol to NTX/creatinine (r2 = 0.13).

Healthy girls. Adiponectin independently contributed to the variability of LS-BMAD, and its Z-score on regression modeling (r2 = 0.30 and 0.35). Hip BMD was independently associated with lean mass and ghrelin (r2 = 0.62), and its Z-score with lean mass, GH, estradiol, and PYY (r2 = 0.81). Adiponectin contributed independently to the variability of FN-BMAD (r2 = 0.26); and adiponectin, lean mass, and GH to that of FN-BMAD Z-score (r2 = 0.55). Lean mass and adiponectin were independently associated with TB-BMC/Ht (r2 = 0.76), and lean mass, GH, adiponectin, and IGF-I with TB-BMC/Ht Z-scores (r2 = 0.90). Ghrelin and lean mass were associated with TB-BMD (r2 = 0.52), and GH and lean mass with TB-BMD Z-scores (r2 = 0.54). Lean mass, GH, IGF-I, ghrelin, and estradiol were positively associated; and PYY, adiponectin, and leptin were inversely associated with BMD measures in this regression model. PICP was independently associated with GH, leptin, adiponectin, lean mass, estradiol, and cortisol (r2 = 0.85); OC with GH and BMI (r2 = 0.42); NTX/creatinine with GH, estradiol, and insulin (r2 = 0.66); and DPD/creatinine with estradiol, leptin, adiponectin, and IGF-I (r2 = 0.78). Adiponectin was positively associated with PICP and inversely with DPD/creatinine.

Effects of weight gain

Weight gain (n = 9 AN) was not associated with changes in adiponectin, adiponectin/BMI, or insulin measures. However, even after this 10% increase in BMI, five of the nine girls were less than 90% ideal body weight for age and height. Adiponectin/fat mass decreased from 0.69 to 0.45 (P = 0.02) with weight gain, but was not associated with changes in BMD.


   Discussion
Top
 Abstract
 Introduction
 Subjects and Methods
 Results
 Discussion
References
 
We demonstrate an inverse association between adiponectin and BMD measures, despite the lack of a difference in adiponectin levels between AN subjects and controls. Insulin measures were independently associated with bone turnover markers, but not with BMD for the group as a whole.

Adiponectin was independently associated with BMAD Z-scores at the LS and FN, and also with TB-BMD and TB-BMC/ht for the group as a whole, and remained a significant and independent contributor to the variability of LS, FN, and TB measures for controls and for TB measures in AN subjects on regression modeling. The inverse association between adiponectin and BMD observed in this study in healthy and AN girls is consistent with similar reports in healthy men and women (22, 23) and with the observation that adiponectin activates RANKL and inhibits OPG secretion, which would cause increased osteoclastic activity and decreased bone density (9). RANKL binds to RANK to activate osteoclastic activity and inhibit osteoclast apoptosis (reviewed in Ref. 28). OPG acts as a soluble decoy receptor for RANK, and by preventing RANKL from binding to RANK, OPG prevents osteoclast activation and decreases apoptosis. OPG and RANKL expression are regulated by many factors including estrogen, and hypoestrogenism causes a decrease in OPG and an increase in RANKL (29). Because AN is associated with hypogonadism, one would expect low OPG in this condition (29). However, we have previously reported high OPG levels in AN, which we hypothesized is a compensatory mechanism given the low bone density state in AN (25). In addition to a lack of association between OPG and estrogen, we observed no associations between adiponectin and OPG or between adiponectin and bone turnover markers. It is possible that in severe undernutrition, OPG is driven by unknown factors, and therefore estrogen and adiponectin do not show associations with OPG in AN. In addition, circulating OPG may not be predictive of local OPG at the level of bone, and lack of correlations between OPG and adiponectin may not reflect local adiponectin effects in bone. Thus, although we demonstrate that adiponectin is independently associated with BMD in healthy and AN girls, it is not clear whether this is OPG mediated.

In our multivariate models, in addition to adipokines and insulin, we included hormonal and body composition predictors of bone metabolism previously studied to construct a model that explains as much of the variability in BMD and bone turnover markers as possible. Therefore, in addition to adipokines, we included body composition, ghrelin, GH, IGF-I, cortisol, estradiol, and PYY in stepwise regression analyses. Covariates that were independently associated with BMD in our model included body composition, GH, ghrelin, estradiol, and IGF-I (positive associations); and adiponectin, PYY, leptin, and cortisol (negative associations). Lean mass is a known important predictor of BMD (1, 5, 7), IGF-I is bone anabolic (1, 5, 30), and ghrelin increases osteoblastic activity (31). Cortisol excess has multiple effects on bone resulting in low BMD, and PYY receptor knockout mice have increased bone formation, suggesting that PYY excess may decrease bone formation (32). Hypogonadism is associated with increased bone resorption and therefore low BMD. Inverse associations between GH and BMD have been attributed to GH resistance in AN (2). Leptin-deficient mice have increased bone mass (33), and studies in adults indicate inverse associations between leptin and bone (34, 35). Our data are consistent with these studies. Our regression models developed for BMD measures include variables that would be expected to affect BMD in AN, a condition in which BMI, lean mass, IGF-I, and estradiol are decreased, and PYY and cortisol increased.

Of note, insulin correlated positively with BMD on univariate analysis, but did not emerge as an independent correlate on regression modeling. Conversely, variables that did not appear to be associated or were only weakly associated with BMD on univariate analysis, such as ghrelin, GH, cortisol, IGF-I, and PYY, contributed significantly to the variability of these measures on regression modeling. Also, the direction of association of leptin with bone differed in the regression model (negative) compared with simple correlation (positive association). This suggests complex interactions between hormones and body composition in determining bone metabolism. Of importance, the regression models explained a high proportion of the variability of BMD measures. It is necessary to remember, however, that these regression models do not prove causality.

Although insulin was not independently associated with BMD, it was associated with PICP and NTX/creatinine. This is consistent with the bone anabolic effect of insulin and its known positive effect on osteoblastic activity (16). Insulin and bone turnover markers change relatively acutely in response to short-term weight changes, whereas BMD reflects more long-term changes. This may explain why insulin did not correlate with BMD. Other independent contributors to the variability of bone turnover markers were GH, ghrelin, estradiol, and PYY. Positive associations of GH and ghrelin with these markers are consistent with their bone anabolic effects (31, 36). We have demonstrated that PYY is strongly associated with bone turnover markers (3), and PYY remained a significant contributor to the variability in bone turnover markers after including adiponectin in the model.

In AN, similar to the combined group, covariates independently associated with bone turnover markers were PYY, cortisol, ghrelin, and GH. In controls, in addition to ghrelin, GH, and cortisol, insulin and estradiol were also independently associated with bone turnover markers, with a positive association noted with insulin, consistent with its anabolic effects, and an inverse association of bone resorption markers with estrogen, consistent with its antiresorptive effects. Adiponectin was a minor contributor to the variability of these markers in controls, but not AN, and was associated positively with PICP, a bone formation marker, and inversely with DPD/creatinine, a bone resorption marker in the model. This is in contrast to what one may expect given that adiponectin inhibits OPG, which should increase osteoclastic activity (9). However, Oshima et al. (11) have reported that adiponectin stimulates osteoblasts and inhibits osteoclasts, consistent with our findings. The association of adiponectin with bone turnover markers needs to be explored further in larger studies.

We were surprised to find no associations of adiponectin with fat mass, insulin, and HOMA-IR. However, other studies have been unable to detect associations between adiponectin and total fat mass (22) or reported only weak associations with insulin (34). Adiponectin is a product of adipocytes, and marked reductions in fat mass in AN may blunt the expected increase in adiponectin, causing a loss of association between adiponectin and insulin. Although absolute adiponectin did not differ, the ratio of adiponectin to fat mass and BMI was higher in AN subjects than controls. Adiponectin decreased after oral glucose in AN, and less so in controls. However, the extent of reduction did not differ, indicating normal regulation of adiponectin after a carbohydrate load in both groups.

We and others have reported that AN is associated with increased marrow fat (37, 38). Although we observed no associations between adiponectin and fat mass or trunk fat, it is not known whether adiponectin is associated with marrow fat in AN. Marrow fat secretes adiponectin (39), and adiponectin affects fat and bone cell differentiation in the marrow (40). Recombinant adiponectin inhibits differentiation of stromal preadipocytes in marrow cultures (40) and may increase (10) or decrease (12) osteoblastic activity. However, marrow cultures from adiponectin null mice demonstrate no differences in adipose tissue, although osteogenesis is lower than in wild-type mice (12). There may be differences in effects of exogenous vs. endogenous adiponectin on selective differentiation of marrow progenitors into preadipocytes or preosteoblasts. Additional studies are necessary to understand the relationship between adiponectin, marrow fat, and bone.

We demonstrate that adiponectin is associated inversely and independently with BMD, whereas insulin is positively associated with bone turnover markers in adolescents. Our regression models explain a high proportion of the variability in BMD. However, additional studies and larger numbers of subjects are necessary to better understand the complex interplay of body composition and hormonal factors in determining the state of bone metabolism. Furthermore, this needs to be examined across a spectrum of nutritional conditions to develop a clearer understanding of the effect of nutritional status on bone.


   Footnotes
 
This work was supported in part by National Institutes of Health Grants M01-RR-01066, DK 062249, and K23 RR018851.

Disclosure Statement: The authors have nothing to disclose.

First Published Online March 13, 2007

Abbreviations: AN, Anorexia nervosa; BMAD, bone mineral apparent density; BMD, bone mineral density; BMI, body mass index; CV, coefficient of variation; DPD, deoxypyridinoline; FN, femoral neck; HOMA-IR, homeostasis model assessment-insulin resistance; LS, lumbar spine; NTX, for N-telopeptide; OC, osteocalcin; OPG, osteoprotegerin; PICP, carboxyterminal propeptide of type 1 procollagen; PYY, peptide YY; RANK, receptor activator of nuclear factor-{kappa}B; RANKL, RANK ligand; TB, total body; TB-BMC/ht, TB bone mineral content/height.

Received December 22, 2006.

Accepted March 5, 2007.


   References
Top
 Abstract
 Introduction
 Subjects and Methods
 Results
 Discussion
 References
 

  1. Soyka L, Misra M, Frenchman A, Miller K, Grinspoon S, Schoenfeld D, Klibanski A 2002 Abnormal bone mineral accrual in adolescent girls with anorexia nervosa. J Clin Endocrinol Metab 87:4177–4185[Abstract/Free Full Text]
  2. Misra M, Miller K, Bjornson J, Hackman A, Aggarwal A, Chung J, Ott M, Herzog D, Johnson M, Klibanski A 2003 Alterations in growth hormone secretory dynamics in adolescent girls with anorexia nervosa and effects on bone metabolism. J Clin Endocrinol Metab 88:5615–5623[Abstract/Free Full Text]
  3. Misra M, Miller KK, Tsai P, Gallagher K, Lin A, Lee N, Herzog DB, Klibanski A 2006 Elevated peptide YY levels in adolescent girls with anorexia nervosa. J Clin Endocrinol Metab 91:1027–1033[Abstract/Free Full Text]
  4. Misra M, Miller K, Almazan C, Ramaswamy K, Lapcharoensap W, Worley M, Neubauer G, Herzog D, Klibanski A 2004 Alterations in cortisol secretory dynamics in adolescent girls with anorexia nervosa and effects on bone metabolism. J Clin Endocrinol Metab 89:4972–4980[Abstract/Free Full Text]
  5. Misra M, Aggarwal A, Miller KK, Almazan C, Worley M, Soyka LA, Herzog DB, Klibanski A 2004 Effects of anorexia nervosa on clinical, hematologic, biochemical, and bone density parameters in community-dwelling adolescent girls. Pediatrics 114:1574–1583[Abstract/Free Full Text]
  6. Misra M, Miller K, Stewart V, Hunter E, Kuo K, Herzog D, Klibanski A 2005 Ghrelin and bone metabolism in adolescent girls with anorexia nervosa and healthy adolescents. J Clin Endocrinol Metab 90:5082–5087[Abstract/Free Full Text]
  7. Grinspoon S, Thomas E, Pitts S, Gross E, Mickley D, Miller K, Herzog D, Klibanski A 2000 Prevalence and predictive factors for regional osteopenia in women with anorexia nervosa. Ann Intern Med 133:790–794[Abstract/Free Full Text]
  8. Berner HS, Lyngstadaas SP, Spahr A, Monjo M, Thommesen L, Drevon CA, Syversen U, Reseland JE 2004 Adiponectin and its receptors are expressed in bone-forming cells. Bone 35:842–849[Medline]
  9. Luo XH, Guo LJ, Xie H, Yuan LQ, Wu XP, Zhou HD, Liao EY 2006 Adiponectin stimulates RANKL and inhibits OPG expression in human osteoblasts through the MAPK signaling pathway. J Bone Miner Res 21:1648–1656[CrossRef][Medline]
  10. Luo XH, Guo LJ, Yuan LQ, Xie H, Zhou HD, Wu XP, Liao EY 2005 Adiponectin stimulates human osteoblasts proliferation and differentiation via the MAPK signaling pathway. Exp Cell Res 309:99–109[CrossRef][Medline]
  11. Oshima K, Nampei A, Matsuda M, Iwaki M, Fukuhara A, Hashimoto J, Yoshikawa H, Shimomura I 2005 Adiponectin increases bone mass by suppressing osteoclast and activating osteoblast. Biochem Biophys Res Commun 331:520–526[CrossRef][Medline]
  12. Shinoda Y, Yamaguchi M, Ogata N, Akune T, Kubota N, Yamauchi T, Terauchi Y, Kadowaki T, Takeuchi Y, Fukumoto S, Ikeda T, Hoshi K, Chung UI, Nakamura K, Kawaguchi H 2006 Regulation of bone formation by adiponectin through autocrine/paracrine and endocrine pathways. J Cell Biochem 99:196–208[CrossRef][Medline]
  13. Irwin R, Lin HV, Motyl KJ, McCabe LR 2006 Normal bone density obtained in the absence of insulin receptor expression in bone. Endocrinology 147:5760–5767[Abstract/Free Full Text]
  14. Burguera B, Hofbauer LC, Thomas T, Gori F, Evans GL, Khosla S, Riggs BL, Turner RT 2001 Leptin reduces ovariectomy-induced bone loss in rats. Endocrinology 142:3546–3553[Abstract/Free Full Text]
  15. Reseland JE, Syversen U, Bakke I, Qvigstad G, Eide LG, Hjertner O, Gordeladze JO, Drevon CA 2001 Leptin is expressed in and secreted from primary cultures of human osteoblasts and promotes bone mineralization. J Bone Miner Res 16:1426–1433[CrossRef][Medline]
  16. Thrailkill KM, Lumpkin Jr CK, Bunn RC, Kemp SF, Fowlkes JL 2005 Is insulin an anabolic agent in bone? Dissecting the diabetic bone for clues. Am J Physiol Endocrinol Metab 289:E735—E745
  17. Kadowaki T, Yamauchi T, Kubota N, Hara K, Ueki K, Tobe K 2006 Adiponectin and adiponectin receptors in insulin resistance, diabetes, and the metabolic syndrome. J Clin Invest 116:1784–1792[CrossRef][Medline]
  18. Misra M, Miller KK, Kuo K, Griffin K, Stewart V, Hunter E, Herzog DB, Klibanski A 2005 Secretory dynamics of leptin in adolescent girls with anorexia nervosa and healthy adolescents. Am J Physiol Endocrinol Metab 289:E373—E381
  19. Housova J, Anderlova K, Krizova J, Haluzikova D, Kremen J, Kumstyrova T, Papezova H, Haluzik M 2005 Serum adiponectin and resistin concentrations in patients with restrictive and binge/purge form of anorexia nervosa and bulimia nervosa. J Clin Endocrinol Metab 90:1366–1370[Abstract/Free Full Text]
  20. Pannacciulli N, Vettor R, Milan G, Granzotto M, Catucci A, Federspil G, De Giacomo P, Giorgino R, De Pergola G 2003 Anorexia nervosa is characterized by increased adiponectin plasma levels and reduced nonoxidative glucose metabolism. J Clin Endocrinol Metab 88:1748–1752[Abstract/Free Full Text]
  21. Tagami T, Satoh N, Usui T, Yamada K, Shimatsu A, Kuzuya H 2004 Adiponectin in anorexia nervosa and bulimia nervosa. J Clin Endocrinol Metab 89:1833–1837[Abstract/Free Full Text]
  22. Lenchik L, Register TC, Hsu FC, Lohman K, Nicklas BJ, Freedman BI, Langefeld CD, Carr JJ, Bowden DW 2003 Adiponectin as a novel determinant of bone mineral density and visceral fat. Bone 33:646–651[Medline]
  23. Jurimae J, Rembel K, Jurimae T, Rehand M 2005 Adiponectin is associated with bone mineral density in perimenopausal women. Horm Metab Res 37:297–302[CrossRef][Medline]
  24. Katzman DK, Bachrach LK, Carter DR, Marcus R 1991 Clinical and anthropometric correlates of bone mineral acquisition in healthy adolescent girls. J Clin Endocrinol Metab 73:1332–1339[Abstract]
  25. Misra M, Soyka L, Miller K, Herzog D, Grinspoon S, De Chen D, Neubauer G, Klibanski A 2003 Serum osteoprotegerin in adolescent girls with anorexia nervosa. J Clin Endocrinol Metab 88:3816–3822[Abstract/Free Full Text]
  26. Misra M, Miller KK, Almazan C, Ramaswamy K, Aggarwal A, Herzog DB, Neubauer G, Breu J, Klibanski A 2004 Hormonal and body composition predictors of soluble leptin receptor, leptin, and free leptin index in adolescent girls with anorexia nervosa and controls and relation to insulin sensitivity. J Clin Endocrinol Metab 89:3486–3495[Abstract/Free Full Text]
  27. Misra M, Miller KK, Kuo K, Griffin K, Stewart V, Hunter E, Herzog DB, Klibanski A 2005 Secretory dynamics of ghrelin in adolescent girls with anorexia nervosa and healthy adolescents. Am J Physiol Endocrinol Metab 289:E347—E356
  28. Khosla S 2001 Minireview: the OPG/RANKL/RANK system. Endocrinology 142:5050–5055[Abstract/Free Full Text]
  29. Khosla S, Atkinson E, Dunstan C, O’Fallon W 2002 Effect of estrogen versus testosterone on circulating osteoprotegerin and other cytokine levels in normal elderly men. J Clin Endocrinol Metab 87:1550–1554[Abstract/Free Full Text]
  30. Grinspoon S, Baum H, Peterson S, Klibanski A 1995 Effects of rhIGF-I administration on bone turnover during short-term fasting. J Clin Invest 96:900–906[Medline]
  31. Kim S, Her S, Park S, Kim D, Park K, Lee H, Han B, Kim M, Shin C, Kim S 2005 Ghrelin stimulates proliferation and differentiation and inhibits apoptosis in osteoblastic MC3T3–E1 cells. Bone 37:359–369[Medline]
  32. Baldock PA, Sainsbury A, Couzens M, Enriquez RF, Thomas GP, Gardiner EM, Herzog H 2002 Hypothalamic Y2 receptors regulate bone formation. J Clin Invest 109:915–921[CrossRef][Medline]
  33. Takeda S, Elefteriou F, Levasseur R, Liu X, Zhao L, Parker KL, Armstrong D, Ducy P, Karsenty G 2002 Leptin regulates bone formation via the sympathetic nervous system. Cell 111:305–317[CrossRef][Medline]
  34. Oh KW, Lee WY, Rhee EJ, Baek KH, Yoon KH, Kang MI, Yun EJ, Park CY, Ihm SH, Choi MG, Yoo HJ, Park SW 2005 The relationship between serum resistin, leptin, adiponectin, ghrelin levels and bone mineral density in middle-aged men. Clin Endocrinol (Oxf) 63:131–138[CrossRef][Medline]
  35. Lorentzon M, Landin K, Mellstrom D, Ohlsson C 2006 Leptin is a negative independent predictor of areal BMD and cortical bone size in young adult Swedish men. J Bone Miner Res 21:1871–1878[CrossRef][Medline]
  36. Ohlsson C, Bengtsson B, Isaksson O, Andreassen T, Slootweg M 1998 Growth hormone and bone. Endocr Rev 19:55–79[Abstract/Free Full Text]
  37. Mayo-Smith W, Rosenthal DI, Goodsitt MM, Klibanski A 1989 Intravertebral fat measurement with quantitative CT in patients with Cushing disease and anorexia nervosa. Radiology 170:835–838[Abstract/Free Full Text]
  38. Abella E, Feliu E, Granada I, Milla F, Oriol A, Ribera JM, Sanchez-Planell L, Berga LI, Reverter JC, Rozman C 2002 Bone marrow changes in anorexia nervosa are correlated with the amount of weight loss and not with other clinical findings. Am J Clin Pathol 118:582–588[CrossRef][Medline]
  39. Dicker A, Le Blanc K, Astrom G, van Harmelen V, Gotherstrom C, Blomqvist L, Arner P, Ryden M 2005 Functional studies of mesenchymal stem cells derived from adult human adipose tissue. Exp Cell Res 308:283–290[CrossRef][Medline]
  40. Yokota T, Meka CS, Medina KL, Igarashi H, Comp PC, Takahashi M, Nishida M, Oritani K, Miyagawa J, Funahashi T, Tomiyama Y, Matsuzawa Y, Kincade PW 2002 Paracrine regulation of fat cell formation in bone marrow cultures via adiponectin and prostaglandins. J Clin Invest 109:1303–1310[CrossRef][Medline]



This article has been cited by other articles:


Home page
 J. Clin. Endocrinol. Metab.Home page
M. Misra, D. K. Katzman, J. Cord, S. J. Manning, N. Mendes, D. B. Herzog, K. K. Miller, and A. Klibanski
Bone Metabolism in Adolescent Boys with Anorexia Nervosa
J. Clin. Endocrinol. Metab., August 1, 2008; 93(8): 3029 - 3036.
[Abstract] [Full Text] [PDF]

Home page
 J. Clin. Endocrinol. Metab.Home page
M. Misra, R. Prabhakaran, K. K. Miller, M. A. Goldstein, D. Mickley, L. Clauss, P. Lockhart, J. Cord, D. B. Herzog, D. K. Katzman, et al.
Prognostic Indicators of Changes in Bone Density Measures in Adolescent Girls with Anorexia Nervosa-II
J. Clin. Endocrinol. Metab., April 1, 2008; 93(4): 1292 - 1297.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
92/6/2046    most recent
Author Manuscript (PDF)
Right arrow Submit a related Letter to the Editor
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Misra, M.
Right arrow Articles by Klibanski, A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Misra, M.
Right arrow Articles by Klibanski, A.
Related Collections
Right arrow Calcium and Bone Metabolism
Right arrow Metabolism
Right arrow Neuroendocrinology and Pituitary

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Endocrinology Endocrine Reviews J. Clin. End. & Metab.
Molecular Endocrinology Recent Prog. Horm. Res. All Endocrine Journals