Osteoporos Int (2005) 16: 424429 DOI 10.

1007/s00198-004-1711-5

O R I GI N A L A R T IC L E

Decreased osteoprotegerin and increased bone turnover in young female patients with major depressive disorder and a lifetime history of anorexia nervosa
Kai G. Kahl Sebastian Rudolf Leif Dibbelt rn Gehl Beate M. Stoeckelhuber Hans-Bjo Fritz Hohagen Ulrich Schweiger

Received: 9 January 2004 / Accepted: 2 July 2004 / Published online: 5 August 2004 International Osteoporosis Foundation and National Osteoporosis Foundation 2004

Abstract Low bone mineral density (BMD) is a frequent, often persistent complication in patients with major depressive disorder (MDD) and anorexia nervosa (AN) that increases the risk of pathologic fractures. The pathogenetic process underlying osteopenia in MDD and AN is still unclear, although several factors, including a dysbalance of cytokines, are associated with loss of bone mass. Alterations in the serum levels of cytokines have been observed in patients with MDD, AN, and other psychiatric disorders. Therefore, we examined serum levels of cytokines, markers of bone turnover, and BMD in 13 patients with MDD and a lifetime history of AN. Bone turnover markers (osteocalcin and C-terminal degradation products of type I collagen) and tumor necrosis factor a (TNF-a) in patients were signicantly increased compared with those of the control group. Osteoprotegerin (OPG) in patients was signicantly decreased. Eight of 13 patients (62%) displayed osteopenia at the lumbar spine. TNF-a correlated signicantly with C-terminal degradation products of type I collagen, an osteoclastic marker, but signicantly negatively with OPG. Our data suggest that

TNF-a and OPG may play a role in the pathogenetic process underlying osteopenia in these patients. Keywords Anorexia nervosa Major depression Osteopenia Osteoprotegerin Tumor necrosis factor a

Introduction
Following initial case reports of osteoporosis-related fractures in patients with major depressive disorder (MDD) and anorexia nervosa (AN) [1, 2], studies using quantitative computerized tomography and photon absorptiometry showed a decrease of BMD in these patients [3, 4]. Low BMD is associated with the development of osteoporosis and an increased risk for pathologic fractures. The estimated risk increases by a factor of 1.5 to 3 for each reduction of standard deviation (SD) in BMD [5]. However, the pathophysiologic process underlying bone loss in these patients has not been suciently described. Therefore identication of factors associated with the development of low BMD may be crucial for identifying groups at risk for osteoporosis and for preventing further bone mineral reduction. Bone tissue is continuously built by a coordinated process of bone resorption and bone formation. Increased bone resorption either by activation of osteoclastic cells or by inhibition of osteoblastic cells has been described as an important pathophysiologic factor in disorders associated with decreased BMD [6]. Factors thought to contribute to increased bone resorption include an increased activity of the hypothalamus-pituitary-adrenal (HPA) system followed by a subsequent increase in concentrations of serum cortisol, decreased concentrations of estrogen, and a dysbalance of cytokines [7]. Of these, TNF-a and IL-6 have been discussed in connection with bone loss via an increase in the activity of osteoclastic cells [8], while insulin-like growth factor I (IGF-I) and leptin have been associated with

K.G. Kahl (&) S. Rudolf F. Hohagen U. Schweiger Klinik fu r Psychiatrie und Psychotherapie, Medizinische Universita t Schleswig-Holstein, Campus Lu beck, Ratzeburger Allee 160, 23538 Lu beck, Germany E-mail: Kahl.K@psychiatry.uni-luebeck.de Tel.: +49-451-5002980 Fax: +49-451-5004780 L. Dibbelt Institut fu r Klinische Chemie, Medizinische Universita t Schleswig-Holstein, Campus Lu beck, Ratzeburger Allee 160, 23538 Lu beck, Germany B.M. Stoeckelhuber H.-B. Gehl Institut fu t r Radiologie, Medizinische Universita Schleswig-Holstein, Campus Lu beck, Ratzeburger Allee 160, 23538 Lu beck, Germany

425

restoration of bone mass [4]. Other disease- and medication-related processes such as impaired uid and electrolyte balance, dietary and vitamin D deciency, and decreased exercise and exposure to sunshine could play a role in bone loss [9], but the available studies report no evidence in favor of these factors. Epidemiologic studies revealed that in 60% of cases, AN is associated with comorbid MDD [10]. Both disorders are reportedly associated with a dysregulation of the HPA system and a dysregulation of cytokines during acute states of disease [11, 12, 13, 14, 15, 16, 17]. Furthermore, there are reports of alterations of the HPA system or increased serum levels of TNF-a in subgroups of patients suering from MDD or AN even in the absence of the acute state of the disorder [18, 19, 20, 21]. Our study examined BMD, markers of bone turnover, and cytokines in a naturalistic sample of patients with current MDD who had remitted from AN. We chose this group for two reasons. Since comorbidity with aective disorders is the rule rather than the exception in patients with AN, this group is more representative of AN patients in the general population. Furthermore, the chronicity of psychiatric symptoms in this group makes the detection of biochemical mechanisms of osteopenia more probable. The majority of cases of depression and anorexia nervosa occur in the context of personality disorders [22, 23, 24]. Thus we chose to study only patients with MDD and a lifetime history of AN in the context of borderline personality disorder (BPD) in order to further limit the heterogeneity of the sample.

Table 1 Demographic data,T-values, and bone turnover marker in patients and controls (CTRL). A signicant reduction of BMD was found at all bone regions examined when means of patientsZvalues were compared with those of a reference population Patients (n=13) Age (years) BMI (g/m2) Lumbar spine Z-score Osteopenia Right femur Z-score Osteopenia Left femur Z-score Osteopenia Nondominant forearm Z-score Osteopenia Osteocalcin (ng/ml) Crosslaps (pmol/l) 29.27.3 19.91.5 )0.840.86a 8/13 (62%) )1.010.66b 8/13 (62%) )0.0930.68b 7/13 (54%) )1.270.69b 10/13 (77%) 21.98.2a 3,7411,946c CTRL (n=13) 26.35.7 20.41.3 -

13.33.5 2,327802

a Represents <.01 dierence between patients and data from a reference population b Represents <.001 dierence between patients and data from a reference population c Represents <.05 dierence between the groups

Methods and materials

Thirteen unmedicated female patients consecutively admitted to our hospital who met DSM-IV diagnostic criteria for MDD and lifetime AN were included in this study. Exclusion criteria were current AN, amenorrhoea, schizophrenia, oligophrenia, pregnancy, estrogen-deciency, infectious or (auto-) inammatory disease and an age of 17 years or younger. Remission from AN was dened as a weight continuously exceeding a minimum of 17.5 kg/m2 and reappearance of menstruation intervals between 21 and 38 days. Diagnosis was made using the German version of the Structured Clinical Interview for DSM-IV (SCID I and II). Thirteen healthy women of similar age and body mass index (BMI) served as the control group for the metabolic and immunologic parameters (CTRL). To exclude possible chronic AN in CTRL, control subjects also underwent the Structured Clinical Interview for DSM-IV and a questionnaire concerning eating habits. Sociodemographic data are given in Table 1. The study was approved by the local ethics committee. All women in the patient and the control group gave their written informed consent. Physical activity in both groups was comparable in that no study subject reported intensive exercise or physical labor and no study subject was bedridden continuously for more than 3 days during the

year before examination. BMI of all patients (range 18.823.9 kg/m2) and control subjects (range 18.5 22.1 kg/m2) was above the lower 10th percentile at the time of examination according to the German National Nutrition Survey [25]. In all control subjects, menstruation was reported to be regular with intervals ranging from 24 to 35 days during the last year. Bone mineral density (BMD) was measured in all patients by means of dual-energy X-ray absorptiometry (DXA) at the lumbar spine, right femur, left femur, and forearm of the nondominant hand using a Lunar Prodigy Densitometer equipped with software containing reference data for an adult population (version 2.15.092; Lunar, Wisconsin, USA). Patients individual BMD values were expressed as Z-scores and compared with these reference data. Osteopenia was dened according to the WHO guidelines as a T-score )1 [26]. BMD was not measured in CTRL because of German government regulations restricting the use of radiation in healthy women of reproductive age. Serum from each participant was collected during the early menstrual phase (between day 3 and day 6 postmenstruation) at 07:30 a.m. and stored at )40C until analysis. Laboratory markers of bone turnover (serum osteocalcin and serum C-terminal degradation products of type I collagen, referred to as crosslaps), intact parathyroid hormone (PTH), and 1,25-hydroxyvitamin D were determined using commercial IRMA (Nichols Institute Diagnostics, Germany) and ELISA kits (Osteometer Biotech, Denmark), respectively. Serum cortisol was determined by RIA (DPC, Bad Naunheim, Germany). TNF-a receptors I and II, interleukin 2 (IL-2), interferon c (IFN-c), insulin-like growth factor I (IGFI), leptin, and osteoprotegerin (OPG) were determined

426

using available ELISA kits according to the manufacturers instructions (all from R&D Systems, Germany). High-sensitivity ELISA kits were used (HS Quantikinine; R&D Systems, Germany) to determine tumor necrosis factor a (TNF-a) and interleukin 6 (IL-6). Data were analyzed using SPSS (version 10.0). Power calculation showed that a sample size of 13 in the patient group is sucient to detect a dierence in the T-score of )0.8 (SD 0.8) with a power of 80%. Groups were compared by means of t-test. Further analysis was performed with MANCOVA. Spearman coecients of correlation were calculated. A p value below 0.05 was considered to be signicant. All values are given as mean + SD where appropriate.

Table 2 Serum hormones, levels of fasting cortisol, and cytokines in patients compared with controls (CTRL) Patients (n=13) Cortisol (nmol/l) Estrogen (pg/ml) 1,25-Hydroxyvitamin D (ng/ml) Parathormone (pg/ml) TNF-a (pg/ml) TNF-RI (pg/ml) TNF-RII (pg/ml) IL-6 (pg/ml) IGF-I (ng/ml) OPG (pmol/l) Leptin (ng/ml)
a

CTRL (n=13) 530108 8067.8 79.638.4 26.012.9 2.00.5 1,260157 2,287104 0.80.3 18632 3.50.6 12.711.4

596177 6943.3 82.424.6 25.28.5 2.91.0a 1,227173 2,169290 1.21.1 18457 3.00.7a 15.315.9

Represents <.05 dierence between the groups

Results
Sociodemographic data are given in Table 1. We found no dierences between the groups with regard to age (df=24; p=.24) and BMI (df=24; p=.24). Patients had developed AN by a mean age of 14.3 years (1.5 years) and had been in remission from AN for 3.3 year (3.1 years) before examination. No correlation was found between the onset of AN or the time remission and BMD in any bone region examined (data not shown). Osteopenia (T-value < )1 SD) was observed in 8/13 (62%) patients at the lumbar spine (Table 1), 8/13 (62%) at the right femur, 7/13 (54%) at the left femur, and in 10/13 (77%) at the forearm of the nondominant hand (Table 1). The t-test revealed signicant dierences concerning the bone regions examined when Z-values of patients were compared with the reference data for an adult population matched for age, sex, and BMI (lumbar spine: df=12; p=.04; right femur: df=12; p<.001; left femur df=12; p<.001; forearm: df=12; p<.001) (Table 1). MANCOVA with the dependent variables TNF-a, OPG, osteocalcin, and crosslaps revealed a signicant eect of group (F=4.84; df=4.19; p=.007) but no eect of BMI (F=.36; df=4.19; p=.83) or age (F=1.00; df=4.19; p=.65). Post hoc analysis revealed signicantly higher concentrations of TNF-a (F=6.1; df=1.22; p=.021), osteocalcin (F=11.6; df=1.22; p=.003), and crosslaps (F=6.7; df=1.22; p=.016) in patients when compared with the CTRL. OPG was lower than in CTRL (F=4.4; df=1.22 ; p=.048) (Tables 1 and 2). Serum levels of TNF-a correlated positively with crosslaps (r=.42; p=.032) but negatively with OPG (r=).40; p=.042) (Fig. 1). Osteocalcin concentrations correlated positively with crosslaps (r=.57; p=.002). No relevant correlations were found between age, BMI, age at onset of AN, duration of AN, time since remission from AN, and BMD at any bone region examined or serum cytokine levels (data not shown). No dierences between the groups were found for fT3, fT4, thyroid stimulating hormone, calcium, magnesium, phosphate (data not shown), or the other hormones and cytokines examined (Table 2).

Discussion
One major nding of our study is that over 50% of young women with MDD and a lifetime history of AN display osteopenia at all bone regions examined. These women had remitted from AN 3.3 years earlier on average. Although the mean BMD reduction was relatively mild, these young patients are prone to further development of osteoporosis and may constitute a risk group for pathologic fractures.

Fig. 1Correlation curves. TNFa was positively correlated with markers of bone resorption (crosslaps) and negatively correlated with OPG

427

Studies examining patients with MDD have been carried out in patients at a mean age of more than 40 years [3]. More recent studies examining bone mass in patients recovered from AN outside the context of MDD and BPD pointed to a near normal volumetric BMD [27]. Another study examining 19 women 21 years after recovery from AN found that their lumbar BMD did not dier signicantly from that of healthy controls [28]. The lower bone density in our patients may potentially be explained by the comorbidity with MDD, by the relatively short period since their remission from AN, incomplete recovery (e.g., persistent abnormalities of food composition), or by delayed age of bone peak mass. Another important nding is that MDD/AN women had signicantly higher levels of serum crosslaps, which is an osteoclastic marker, and of osteocalcin, which is a marker for bone formation. These ndings point to a high rate of bone-turnover osteopenia in our patients. Other studies on alterations of markers of bone turnover yielded inconsistent results, pointing either to a low or a high rate of bone turnover in MDD or AN [4, 9, 29]. Our results are in accordance with those of a recent n and coworkers, who found high rates of study by Herra bone remodeling in patients with rst-episode MDD but no correlation of bone turnover parameters with either serum-cortisol or IL-6 [9]. We found high concentrations of serum TNF-a and low concentrations of OPG in patients compared with controls. TNF-a correlated negatively with OPG but positively with crosslaps, which are a marker for bone loss. These ndings extend the existing literature and indicate that these cytokines may play a role in the bone loss observed in our patients. TNF-a has been described as a proresorptive cytokine that induces the genesis of osteoclasts by increasing the concentrations of ligandreceptor activator of nuclear factor-jB (RANKL), an osteoclast-activating factor [30]. OPG represents an endogenous receptor antagonist that neutralizes the biologic eects of the RANKL and prevents bone loss. Consistent with these ndings, OPG reportedly inhibits the proliferation, dierentiation, survival, and fusion of osteoclastic precursor cells and promotes osteoclast apoptosis in vitro [31]. In animal models, OPG knockout mice displayed severe osteoporosis due to excessive bone resorption [32, 33]. OPG, RANKL, and the receptor activator of nuclear factor-jB (RANK) are thought to represent a novel cytokine system that is capable of regulating the proliferation, activation, differentiation, fusion, and apoptosis of osteoclastic cells. Dysregulation of this cytokine system is thought to underlie bone loss due to dierent somatic disorders [8]. Our ndings of increased TNF-a and decreased OPG suggest that both cytokines may be involved in the bone loss observed in our patients, and that TNF-a may, at least in part, be linked to the activation of osteoclastic cells. Interestingly, we found low OPG associated with higher concentrations of bone turnover markers in our

patients, whereas in most physiologic and pathologic situations, concentrations of OPG and bone turnover markers are both elevated or reduced. However, in glucocorticoid-induced osteoporosis a decrease of OPG in association with an increase of bone resorption has been described [34, 35]. In line with others who found a dysregulation of the hypothalamus-pituitary-adrenal system in MDD (reviewed in [12]), serum cortisol concentrations were elevated in our depressed patients when compared with healthy women, although not statistically signicant. The lack of signicance may be explained by dierent methods in cortisol measurements. Determination of cortisol concentrations by urine cortisol or serum cortisol proles more precisely represents alterations of the HPA system than single fasting cortisol measurements. However, it seems reasonable to hypothesize that the observed ndings of decreased OPG in association with increased bone turnover may at least in part be explained by increased concentrations of glucocorticoids in our patients. Depressed women with a lifetime history of AN may be a special group at high risk for the development of osteoporosis early in life. Unlike for MDD-associated osteoporosis, several studies exist on the treatment of osteoporosis in AN. Estrogen replacement therapy failed to normalize bone mass in AN [36]. More recent experimental treatment strategies include hypercaloric diets and recombinant IGF-I [37, 38]. Interestingly, in women suering from postmenopausal osteoporosis, treatment with OPG resulted in a suppression of biochemical markers of bone metabolism [39]. This latter treatment may be an option for future therapy strategies. Similar concentrations of IL-6, IGF-I, leptin, and TNF-a RI/II were found in the two study groups. Furthermore, the ranges of fasting cortisol, PTH, and 1,25hydroxyvitamin D are comparable in both study groups. Mean estrogen concentrations were 69 43.3 pg/ml in the patient and 80 67.8 pg/ml in the control group. In our laboratory, the lower cuto of the reference range is 30 pg/ml. Two women in the patient and three women in the control group had values below the lower reference point. Taken together these ndings do not support the hypothesis that estrogen and vitamin D deciency or alterations in the above-mentioned peptides may play a major role in bone loss observed in our patients. However, this assumption is limited by the low power to detect states of altered gonadal or adrenal function on the basis of single morning blood samples. Unfortunately, we were not able to determine BMD in the healthy women who served as the control group for the serum parameters, which poses another limitation. However, care was taken to exclude the presence of psychiatric and medical disorders in the CTRL group. It is therefore unlikely that probands in the CTRL group will have displayed loss of BMD. Furthermore, the BMD values of patients have been compared with the data from a reference population matched for age, sex, and BMI according to the WHO guidelines. Another

428

limitation is that we were not able to reliably determine the lowest BMI during phases of AN in our patients and that we had no information about sex steroids during adolescence. Therefore, the possible association between lowest BMI or low peak bone mass and present BMD cannot be examined. A general limitation may apply to the fact that information about eating habits and menses was based on self-reports using questionnaires. The retrospective and subjective character of these data has to be taken into account. Patients suered from a high level of psychiatric comorbidity with lifetime AN and current MDD and BPD. This limits the specicity in that immunologic and metabolic alterations can be attributed to a single disease entity. Yet, since the majority of cases of major depression in young women that come to clinical treatment occurs in the context of comorbidity, our data may better represent clinical populations. In summary, we found low bone density, high concentrations of TNF-a, low concentrations of OPG, and high rates of bone turnover markers (osteocalcin and crosslaps) in patients with current MDD and lifetime AN, indicating high rates of bone-turnover osteopenia. We hypothesize a pathogenetic role of TNF-a and OPG in the process underlying the bone loss in this group of patients. Our ndings point to potential new mechanisms that may support reduction of BMD in this patient sample.
Acknowledgements The study was supported by a grant of the Medical University of Schleswig-Holstein (FUL2301).

References
1. Brotman AW, Stern TA (1985) Osteoporosis and pathological fractures in anorexia nervosa. Am J Psychiatry 142:495496 2. Van Vort WB, Rubenstein M, Rose RP (1990) Osteoporosis with pathological hip fractures in major depression. J Geriatr Psychiatry Neurol 3:1012 3. Cizza G, Ravn P, Chrousos GP, Gold PW (2001) Depression: a major, unrecognized risk factor for osteoporosis? Trends Endocrinol Metab 12:198203 oz MT, Argente J (2002) Anorexia nervosa in female 4. Mun adolescents: endocrine and bone mineral density disturbances. Eur J Endocrinol 147:275286 5. Whooley MA, Kip KE, Cauley JA, Ensrud KE, Nevitt MC, Browner WS (1999) Depression, falls and risk of fracture in older women. Study of Osteoporotic Fractures Research Group. Arch Int Med 159:484490 6. Hofbauer LC, Neubauer A, Heufelder AE (2001) Receptor activator of nuclear factor-kappaB ligand and osteoprotegerin: potential implications for the pathogenesis and treatment of malignant bone diseases. Cancer 92:460470 7. Halbreich U, Palter S (1996) Accelerated osteoporosis in psychiatric patients: possible pathophysiologic processes. Schizophr Bull 22:447454 8. Hofbauer LC, Khosla S, Lacey DL, Dunstan CR, Boyle WJ, Riggs BL (2000) The roles of osteoprotegerin and osteoprotegerin ligand in the paracrine regulation of bone resorption. J Bone Miner Res 15:212 n A, Amado JA, Garc a-Unzueta MT, Va zquez-Bar9. Herra lez-Mac as J (2000) Increased bone quero JL, Perera L, Gonza remodelling in rst-episode major depressive disorder. Psychosom Med 62:779782

10. Halmi KA, Eckert E, Marchi P, Sampugnaro V, Apple R, Cohen J (1991) Comorbidity of psychiatric diagnoses in anorexia nervosa. Arch Gen Psychiatry 48:712718 11. Corcos M, Guilbaud O, Paterniti S, Moussa M, Chambry J, Chaouat G, Consoli SM, Jeammet P (2003) Involvement of cytokines in eating disorders: a critical review of the human literature. Psychoneuroendocrinology 28:229249 12. Heuser I (1998). The hypothalamic-pituitary-adrenal axis in depression. Pharmacopsychiatry 31:1013 13. Licinio J, Wong ML, Gold PW (1996) The hypothalamuspituitary-adrenal axis in anorexia nervosa. Psychiatry Res 62:7583 14. Dantzer R, Wollman E, Vitkovic L, Yirmiya R (1999) Cytokines and depression: fortuitous or causative association? Molecular Psychiatry 4:328332 15. Maes M (1999) Major depression and activation of the inammatory response syndrome. Adv Exp Med Biol 461:25 46 16. Kiecolt-Glaser, JK, Glaser R (2002) Depression and immune function: central pathways to morbidity and mortality. J Psychosom Res 53:873876 17. Van West D, Maes M (1999) Activation of the inammatory response system: a new look at the etiopathogenesis of major depression. Neuroendocrinol Lett 20:1117 18. Holsboer F, Lauer CJ, Schreiber W, Krieg JC (1995) Altered hypothalamic-pituitary-adrenocortical regulation in healthy subjects at high familial risk for aective disorders. Neuroendocrinology 62:340347 19. Lanqillon S, Krieg J-C, Bening-Abu-Shach U, Vedder H (2000) Cytokine production and treatment response in major depressive disorder. Neuropsychopharmacology 22:370379 20. Nakai Y, Hamagaki S, Takagi R, Taniguchi A, Kurimoto F (1999) Plasma concentrations of tumor necrosis factor-alpha (TNF-alpha) and soluble TNF-alpha receptors in patients with anorexia nervosa. J Clin Endocrin Metabol 84:12261228 21. Kahl KG, Kruse N, Rieckmann P, Schmidt MH (2004) Cytokine mRNA expression patterns in the disease course of female adolescents with anorexia nervosa. Psychoneuroendocrinology 29:1320 22. Corruble E, Ginestet D, Guel JD (1996) Comorbidity of personality disorders and unipolar major depression: a review. J Aect Disord 37:157170 23. Herzog DB, Keller MB, Lavori PW, Kenny GM, Sacks NR (1992) The prevalence of personality disorders in 210 women with eating disorders. J Clin Psychiatry 53:147152 24. Rossi A, Marinangeli MG, Butti G, Scinto A, Di Cicco L, Kalyvoka A, Petruzzi C (2001) Personality disorders in bipolar and depressive disorders. J Aect Disord 65:38 25. Hebebrand J, Himmelmann GW, Heseker H, Schaefer H, Remschmidt H (1996). Use of percentiles for the body mass index in anorexia nervosa: diagnostic, epidemiological, and therapeutic considerations. Int J Eat Disord 19:359369 26. WHO study group (1994) Assessment of fracture risk and its application to screening for postmenopausal osteoporosis. World Health Organ Tech Rep Ser 843:1129 27. Karlsson MK, Weigall SJ, Duan Y, Seeman E (2000) Bone size and volumetric density in women with anorexia nervosa receiving estrogen replacement therapy and in women recovered from anorexia nervosa. J Clin Endocrinol Metab 85:3177 3182 28. Hartmann D, Crisp A, Rooney B, Rackow C, Atkinson R, Patel S (2000) Bone density of women who have recovered from anorexia nervosa. Int J Eat Disord 27:2935 29. Michelson D, Stratakis C, Hill L, Reynolds J, Galliven E, Chrousos G, Gold P (1996) Bone mineral density in women with depression. N Engl J Med 335:11761181 30. Lam J, Takeshita S, Barker JE, Kanagawa O, Ross FP, Teitelbaum SL (2000) TNF-alpha induces osteoclastogenesis by direct stimulation of macrophages exposed to permissive levels of RANK ligand. J Clin Invest 106:14811488 31. Lacey DL, Timms E, Tan HL, Kelley MJ, Dunstan CR, Burgess T, Elliott R, Colombero A, Elliott G, Scully S, Hsu H,

429 Sullivan J, Hawkins N, Davy E, Capparelli C, Eli A, Qian YX, Kaufman S, Sarosi I, Shalhoub V, Senaldi G, Guo J, Delaney J, Boyle WJ (1998) Osteoprotegerin is a cytokine that regulates osteoclast dierentiation and activation. Cell 93:165176 32. Bucay N, Sarosi I, Dunstan CR, Morony S, Tarpley J, Capparelli C, Scully S, Tan HL, Xu W, Lacey DL, Boyle WJ, Simmonet WS (1998) Osteoprotegerin-decient mice develop early onset osteoporosis and arterial calcication. Genes Dev 12:12601268 33. Mizuno A, Amizuka N, Irie K, Murakami A, Fujise N, Kanno T, Sato Y, Nakagawa N, Yasuda H, Mochizuki S, Gomibuchi T, Yano K, Shima N, Washida N, Tsuda E, Morinaga T, Higashio K, Ozawa H (1998) Severe osteoporosis in mice lacking osteoclastogenesis inhibitory factor/ osteoprotegerin. Biochem Biophys Res Commun 247:610615 34. Hofbauer LC, Gori F, Riggs BL, Lacey DL, Dunstan CR, Spelsberg TC, Khosla S (1999) Stimulation of osteoprotegerin ligand and inhibition of osteoprotegerin production by glucocorticoids in human osteoblastic lineage cells: potential paracrine mechanisms of glucocorticoid-induced osteoporosis. Endocrinology 140:43824389 35. Chiodini I, Carnevale V, Torlontano M, Fusilli S, Gugglielmi P, Pileri M, Modoni S, di Giorgio A, Liuzzi A, Minisola S, Cammisa M, Trischitta V, Scillitani A (1998) Alterations of bone turnover and bone mass at dierent skeletal sites due to pure glucocorticoid excess: study in eumenorrheic patients with Cushings syndrome. J Clin Endocrinol Metab 18631867 36. Grinspoon S, Herzog D, Klibanski A (1997) Mechanisms and treatment options for bone loss in anorexia nervosa. Psychopharmacol Bull 33:399404 37. Heer M, Mika C, Grzella I, Drummer C, Herpertz-Dahlmann B (2002) Changes in bone turnover in patients with anorexia nervosa during eleven weeks of inpatient dietary treatment. Clin Chem 48:754760 38. Grinspoon S, Thomas L, Miller K, Herzog D, Klibanski A (2002) Eects of recombinant IGF-I and oral contraceptive administration on bone density in anorexia nervosa. J Clin Endocrinol Metab 87:28832891 39. Becker PJ, Holloway D, Nakanishi A, Arrighi M, Leese PT, Dunstan CR (2001) The eect of a single dose of osteoprotegerin in postmenopausal women. J Bone Miner Res 16:348 360