Download this issue

Vijay K. YADAV, PhD
Gerard KARSENTY, MD, PhD
Department of Genetics and
Development
Patricia DUCY, PhD
Department of Pathology
Columbia University Medical
Center, New York, NY
USA

Serotonin is a bioamine synthesized in the brain and gut that regulates diverse functions from mood to gastrointestinal tract motility. This diversity in serotonin function(s) is achieved through one or several of its 14 distinct receptor(s) expressed on the target cells. The emerging concept that brain- and gut-derived serotonin regulate bone remodeling in opposite manner has revealed novel mechanism(s) by which bone mass is regulated and maintained. Advances in our understanding of serotonin synthesis, receptor activation, and participation in distinct regulatory networks demonstrate a role for serotonin in osteoblast and osteoclast functions. This review focuses on this new “expanded serotonin biology” and discusses how drugs targeting serotonin synthesis or signaling can be harnessed for treating low-bone-mass diseases.

Skeleton in vertebrates serves multiple mechanical, hematopoietic, and endocrine functions.¹ In order to perform its functions properly, skeleton continuously renews itself through a homeostatic process known as bone remodeling—a process carried out by osteoblasts and osteoclasts to maintain a fine balance between bone formation and resorption.¹ Bone remodeling occurs constantly and simultaneously in numerous parts of skeleton and the maintenance of a normal, healthy skeletal mass depends on continuous exchange of information taking place among osteoblasts, osteoclasts, osteocytes, constituents of the bone matrix, and other organs.¹ Therefore, understanding what factors are influencing bone mass in the context of other signals is important. The fact that osteoporosis is a heritable trait provides an opportunity to use modern molecular genetics to obtain mechanistic insights that were previously unobtainable. If we could find genetic variants with known or at least tractable functions that are unequivocally associated with osteoporosis, we might be able to build up a picture of what sorts of biological factors determine why some people are more susceptible to osteoporosis than others.

Lrp5: a multifaceted molecule

The low-density lipoprotein receptor (LDLR)-related protein (Lrp)-5 is part of a subset of the LDLR family of cell surface proteins.^2,3 Since its cloning in 1998, Lrp5 has taken biologists to voyages of discoveries from lipoprotein clearance to glucose homeostasis to bone remodeling. Not surprisingly, it has been shown to bind to multiple ligands and activate a multitude of downstream cascades in distinct cell types to regulate different processes (Figure 1). In hepatocytes, Lrp5 binds apolipoprotein E (ApoE) and plays a role in the hepatic clearance of ApoE-containing chylomicron remnants, a major plasma lipoprotein carrying diet-derived cholesterol.^4,5 In pancreatic islets, Lrp5 regulates insulin secretion and consequently Lrp5-deficient animals are glucose intolerant.⁶ Consistent with its role in glucose homeostasis, the LRP5 gene is mapped within the region (IDDM4) linked to type 1 diabetes on chromosome 11q13.⁷ LRP5 is also the gene responsible for osteoporosis- pseudoglioma (OPPG) syndrome and high-bone-mass (HBM) syndrome in humans due to an isolated change in bone formation.^8-10

The main question surrounding Lrp5 biology, since its identification as the cause of OPPG, has been to define how its absence can cause the developmental onset of blindness and postnatal onset of osteoporosis characterizing this disease.^8-10 Several recent studies have now shed¹¹ new light on the mechanisms associated with these two functions of Lrp5. Indeed, ample studies have conclusively demonstrated that Lrp5 uses the Norrin and Wnt signaling pathways during embryogenesis to regulate vascularization in the eyes.^12-14 That dysregulation of Wnt signaling plays a role in the development of blindness in a Lrp5-dependent manner fuelled interest in this signaling pathway, leading to the identification of critical Wnt-dependent mechanisms involved in controlling early differentiation of osteochondroprogenitor cells during embryogenesis as well as osteoblast and osteoclast functions.^11,15-17 Some of these targets have already made it to preclinical trials, viz, sclerostin.¹⁸ However, and to our dismay, using an unbiased microarray approach, we serendipitously identified that the mechanism through which Lrp5 loss- and gain-of-function mutations regulate bone formation is by regulating serotonin production in the gut.¹⁹ This dual role of Lrp5—one developmental (directly dependent on Wnt signaling in the eye) and the other postnatal (relying on the indirect effect of gutderived serotonin on bone cells)—is consistent with the multifunctionality of Lrp5, which participates in a wide variety of signaling cascade(s).

Figure 1. Lrp5 ligands and targets.

We should emphasize that the fact that the deletion of Lrp5 in osteoblasts progenitors or mature osteoblasts did not result in a discernible effect on bone mass in our studies does not exclude, however, that Lrp5 could play a role, in a Wntdependentmanner, or not, in regulating the response of osteocytes to mechanical loading.^20-21 Further studies analyzing, in parallel, mice deficient in Lrp5 globally as well as conditionally in osteocytes will be pivotal to address this specific point.

Ever-expanding tenets of serotonin biology

Serotonin (5-hydroxytryptamine) was discovered in 1948 as a factor causing vascular contractions, hence the name of the molecule serotonin (L, serum + Gk, tonos, tone).²² Since then serotonin biology has expanded exponentially and it is now recognized as a pivotal regulator in many central and peripheral functions.²³ Serotonin is generated through an enzymatic pathway in which L-tryptophan is converted into L-5-OHtryptophan by an enzyme called tryptophan hydroxylase (Tph); this intermediate product is then converted to serotonin by an aromatic L-aminoacid decarboxylase.^24,25 There are two Tph genes that catalyze the rate-limiting step in serotonin biosynthesis: Tph1 and Tph2. Tph1 is expressed mostly, but not only, in enterochromaffin cells of the gut and is responsible for the production of peripheral serotonin.²³ Tph2 is expressed exclusively in raphe neurons of the brainstem and is responsible for the production of serotonin in the brain.²⁵ Remarkably, serotonin does not cross the blood–brain barrier; therefore it should be viewed from a functional point of view as two distinct molecules depending on their site of synthesis.²⁴ Brainderived serotonin (BDS) acts as a neurotransmitter, while gut-derived serotonin (GDS), till now, has only been appreciated as an autocrine/paracrine signal that regulates mammary gland biogenesis, liver regeneration, and gastrointestinal tract motility.^23,26,27

Figure 2. Gut-bone endocrine axis: potential molecular targets for therapeutic interventions.

Regulation of bone formation through gut-derived serotonin

Our work on the mechanism(s) underlying OPPG and HBM led us to identify Lrp5 as one of the regulators of GDS (Figure 2).¹⁹ Conditional inactivation of Lrp5 and Tph1 in the gut cells identified that GDS functions as a hormone that directly inhibits osteoblast proliferation and bone formation.¹⁹ We reasoned that if GDS was acting as a hormone one or several of its 14 receptors must be expressed on osteoblasts. Indeed, primary osteoblasts expressed three serotonin receptors: Htr1b, 2a, and 2b. Mice with either global deletion of Htr2a or osteoblast-specific deletion of Htr2b did not display skeletal phenotypes; however, that was not the case for mice with global or osteoblast-specific deletion of Htr1b gene.¹⁹ These latter animals displayed a high bone mass phenotype, although of lower magnitude than the one displayed by the mice expressing HBM mutation of Lrp5 (G171V) in the gut, suggesting that there may be, yet to be identified, mediators of the HBM mutations. Nevertheless, the fact that the HBM phenotype of Htr1b-deficient animals was similar in magnitude to one of the mice that had suppressed levels of GDS demonstrated that it is through Htr1b receptor that GDS regulates bone formation.¹⁹ Htr1b is a Gi-protein–coupled receptor and, consistent with its role in neurons, it inhibited cAMP production and phosphokinase A (PKA)-mediated cAMP response element binding protein (CREB) phosphorylation in primary osteoblasts. These results identified that a cAMPPKA- CREB pathway regulates osteoblasts proliferation and bone formation.¹⁹ Yet, this signaling pathway in the osteoblasts is utilized by many other receptors, and future studies would need to dissociate how this selectivity of Htr1b action on osteoblast proliferation is achieved.

Negative association of peripheral serotonin levels with bone mass in humans

The identification of a gut-derived serotonin-bone endocrine axis (Figure 2) begged the question of its biomedical importance in humans. Modder et al²⁸ analyzed serum serotonin levels in a population-based sample of 275 women and related these to bone mineral densities (BMD) at distinct skeletal sites and bone microstructural parameters. They found that serum serotonin levels were inversely associated in these women with body and spine areal bone mineral density (aBMD) as well as with femur neck total and trabecular volumetric bone mineral density (vBMD).²⁸ Moreover, multiple LRP5 mutations associated with decreased BMDs have been analyzed and all published studies thus far show that these mutations are associated with a 2-to-4 increase in serum serotonin levels.^19,29 Conversely, analysis of two HBM patients in the US as well as a recent study of 9 HBM European patients, who harbor the T253I gain-of-function mutation of LRP5, showed that their serotonin concentrations in platelet-poor plasma were significantly lower compared to sex- and age-matched controls.^19,30 Collectively, these studies performed in different continents by different investigators, provide convincing evidence to support a physiological role for circulating serotonin in negatively regulating bone formation in humans related to one it plays in mice.

Brain-derived serotonin: an expected player in the regulation of bone mass

In our quest to understand the serotonin regulation of bone mass in vertebrates, we then inactivated Tph2, the gene that catalyzes the rate-limiting step in the biosynthesis of BDS. The absence of serotonin in the brain resulted in a severe low-bone-mass phenotype affecting the axial (vertebrae) and appendicular (long bones) skeleton.³¹ This phenotype was secondary to a decrease in bone formation parameters (osteoblast numbers and bone formation rate) and to an increase in bone resorption parameters (osteoclast surface and circulating Dpd levels).³¹ Hence, BDS is a positive and powerful regulator of bone mass accrual acting on both arms of bone remodeling.³¹

While we were doing these studies we noticed, upon opening the abdominal cavities, that Tph2-deficient animals had a dramatic decrease in their adipose mass.³¹ This prompted us to analyze in great detail their energy metabolism phenotype. The decrease in their fat mass was due, in part, to the fact that these mice ate less and spent much more energy compared to their wild-type littermates.³¹ This observation was not entirely surprising since serotonin is known to play important roles in many other physiological processes. However, what caught our attention was the fact that the three most notable phenotypes of adult Tph2-deficient animals ie, decrease in bone mass, decrease in appetite, and increase in energy expenditure are a mirror image of what is observed in mice that lack leptin.^32,33

Leptin is an adipocyte-derived hormone that regulates many functions, viz, appetite, energy expenditure, bone mass, etc.^34-39 Studies in the last 16 years have highlighted a more complete neural and neurochemical circuit diagram for the leptin regulation of these functions.^34-36 These neural circuits involve many distinct neuronal populations in the brain, including neurons of arcuate, Ventromedial, and lateral hypothalamus, and neurons of the nucleus tractus solitarius (NTS) etc.^34-36,40 Three correlative experiments suggested that leptin might signal in the serotonin neurons, among others, to regulate some of its downstream functions. First, the leptin receptor (ObRb) is expressed on serotonin neurons located in the raphe nuclei of brainstem, where BDS is produced, and is functional.^31,41 Second, serotonin neurons project to the key hypothalamic nuclei responsible for the regulation of appetite, energy expenditure, and bone mass.⁴² Third, patients on selective serotonin reuptake inhibitors (SSRIs) have been reported to have changes in their appetite and bone mass.^43,44 To explore that leptin might utilize serotonin as one of its downstream mediators to regulate these three functions, we inactivated the leptin receptor in different nuclei of the hypothalamus or in the serotonergic neurons of the brainstem.³¹ Mice lacking ObRb either in Sf1-expressing neurons of the ventromedial hypothalamus (VMH) nuclei or in Pomc-expressing neurons of the arcuate (ARC) nuclei had normal sympathetic activity, bone remodeling parameters, and bone mass; they also had normal appetite and energy expenditure, and when fed a normal diet, did not develop an obesity phenotype.^40,45 In contrast, mice that lack ObRb in Sert-Cre positive serotonin neurons (ObRbSERT-/-) developed a high bone mass phenotype; they also had an increase in appetite and displayed low-energy expenditure. As a result, ObRbSERT-/- mice, when fed a normal diet, developed an obesity phenotype. These genetic studies demonstrated that leptin signals, in part, in the serotonin neurons of the brainstem to regulate, bone mass, appetite, and energy expenditure (Figure 3). The identification of serotonin as one of its mediators adds to the list of the multitude of messengers (viz, dopamine, melanocortins, etc) utilized by leptin in the brain to affect peripheral functions.^31,34-36

The demonstration that a leptin-dependent central control of bone mass, appetite, and energy expenditure occurs, among other neural relays, through its ability to inhibit serotonin production, raised questions about the location and identity of serotonin receptors on hypothalamic neurons mediating these functions. Double fluorescence in situ hybridization and nuclei- specific gene inactivation experiments revealed that serotonin promotes bone mass accrual through Htr2c receptors expressed on the VMH nuclei, while appetite was promoted through Htr2b and Htr1a receptors expressed on ARC nuclei of the hypothalamus. Further analysis revealed that Htr2c receptor expression on VMH nuclei is en route to the sympathetic center of the brain, while Htr1a and Htr2b achieve their functions on appetite most likely through modulation of melanocortin signaling (Figure 3). These studies emphasized that with respect to the bone mass and energy metabolism effects of leptin signaling in the brain, a systems approach involving anatomically distinct neural elements will provide a more complete explanation of leptin actions in the brain.

Gain of function in serotonin signaling and bone mass

Our loss of function studies with GDS and BDS dissociated the role played by peripheral and central serotonin signaling in the regulation of bone mass.^19,31 As with any other study, these studies raised many more questions than they answered. For instance, what effect would an increase in serotonin signaling have on the bone mass? SSRIs are a class of drugs that do exactly that.⁴⁶ Based on these effects, this class of drugs is prescribed to cure many psychiatric disorders associated with diminished serotonin signaling and their therapeutic actions are diverse, ranging from efficacy in the treatment of depression to obsessive-compulsive disorder, panic disorder, bulimia, and other conditions. The plethora of biological substrates, receptors, and pathways for serotonin are candidates to mediate not only the therapeutic actions of SSRIs, but also their side effects.⁴⁷ In a cohort of 5008 communitydwelling adults, Richards et al⁴⁴ revealed that patients that were taking SSRIs had increased risk of hip fractures. Because SSRIs readily cross the blood–brain barrier, it raised the question as to the site of action of these drugs to produce their deleterious actions on bone.

Figure 3. Neuronal relays underlying leptin regulation of bone mass, appetite, and energy expenditure.

Several approaches have been used in the past to understand this deleterious effect of SSRIs on bone mass. Gustaffson et al,⁴⁸ using naïve rats as a model of serotonin effect on bone mass, analyzed site-specific alterations in the long bone when rats were injected daily with serotonin. These authors reported a decrease in trabecular bone mass and an increase in cortical thickness in long bones. The negative influence of serotonin injections on the trabecular bone mass in their study is consistent with our and earlier mouse genetic studies. We reported that mice harboring a loss of function mutation for Lrp5 gene have increased levels of GDS and a low bone mass at vertebral sites.¹⁹ Battaglino et al⁴⁹ tested the direct effects of SSRIs on bone mass and they consistently observed an increase in trabecular bone mass in these animals.⁴⁹ These latter results, given the effects of SSRIs in humans, were surprising at the time they were reported, but with the advancement of knowledge related to serotonin signaling in the brain and periphery we can today explain these results. Likely the observed effects were due to the fact that, under the conditions tested in their study, SSRIs were having more profound influences on BDS, a positive regulator of bone mass. Warden et al,⁵⁰ taking another approach for a model of chronic use of SSRIs, reported that mice that lack serotonin transporter (Htt-/- mice) have decreased bone mass at both cortical and trabecular sites.

Their study is consistent with Richards et al⁴⁴ and other clinical reports that show that patients taking SSRIs often have a decrease in bone mass. Surprisingly, Htt-/- mice have undetectable levels of serotonin in their blood and a twofold reduction in brain serotonin content (VKY, unpublished observations). The low bone mass observed in Htt-/- mice would suggest that BDS compared to GDS has a dominant role in the overall regulation of bone mass through serotonin. Indeed, analyses of mice lacking both the Tph1 and Tph2 genes display a low bone mass phenotype demonstrating that despite accounting for >5% of total serotonin pool in the body, BDS dominates in the overall regulation of bone mass.³¹ Since SSRIs cross the blood–brain barrier, and osteoporosis is only observed when they are taken in the long term, development of SSRIs with selective central actions would be worth exploring in the future for curing depression while minimizing their side effects on bone.

Therapeutic implications of serotonin regulation of bone mass

The richness and complexity of the serotonin modulation of bone mass discussed in this review provide both a pharmacologic opportunity and a challenge. On the one hand, the involvement of specific serotonin receptors on osteoblasts and hypothalamic neurons provides an opportunity to pharmacologically target these specific receptors for the treatment of osteoporosis. On the other hand, the fact that each of these serotonin receptors participates in multiple physiologic processes presents a challenge, since even a drug targeting a single serotonin receptor is likely to have effects on multiple body systems. For example, although Htr2c agonists may be used to increase bone mass through its effect in the brain, their clinical use would be limited by their effects on other organ systems, such as sympathetic tone or melanocortin signaling.^31,51 Fortunately, the system is less complex and more amenable to therapeutic interventions in the periphery. Since the effect of GDS, a negative regulator of bone formation, is dominant there, one would be able to suppress its levels mildly in order to avoid side effects of drugs targeting the receptors directly. This way one would be able to maintain basal level of signaling in other systems dependent on serotonin while at the same time getting the therapeutic outcome in sensitive systems such as bone, which responds robustly to >50% modulation in peripheral serotonin levels.¹⁹

As GDS is a potent inhibitor of osteoblast proliferation and bone formation, we tested the contention that pharmacologically suppressing GDS would be able to prevent, or cure, gonadectomy- induced bone loss. Serendipitously, we came across an inhibitor that was inhibiting peripheral serotonin production without having any detectable effect on brain serotonin content.⁵² This is, and will be, a prerequisite for any drug that is going to target serotonin synthesis or signaling, as brain serotonin has opposite influence on bone mass accrual and in fact is beneficial to bone. The drug, LP533401, a Tph1 inhibitor, was effective in preventing and even curing osteoporosis in mice and rats at an oral dose of less than 25 mg/kg/day through an isolated increase in bone formation.⁵³ The effect of Tph1 inhibitors on bone mass establishes that inhibition of GDS biosynthesis can rescue ovariectomy-induced osteoporosis in the mouse through an anabolic mechanism. These studies further validate the role of GDS as a regulator of bone formation and provide foundation for the development of other molecules that target the Tph1/Htr1b/osteoblast pathway for the treatment of low bone mass diseases, either alone or in combination with other existing therapies (Figure 2).

Future studies would be necessary to investigate four specific issues: First, the absolute threshold levels at which suppression in peripheral serotonin signaling is anabolic to the bone. Second, to analyze in more detail plasma- vs serumvs platelet-derived serotonin in the regulation of bone mass. Third, to thoroughly characterize any toxicity or side effects the drugs that target this pathway might have on any of the functions of other peripheral organs. Fourth, and most importantly, if these types of drugs can be used to treat low bone mass conditions associated with specific genetic mutations in mouse models of human diseases such as osteoporosis pseudoglioma.

As research on the role of serotonin and its receptors in bone physiology progresses, the difficulty of these challenges will become clearer. In the process we will likely discover new therapeutic targets for osteoporosis treatments as well as gain a better understanding of the beauty and complexity of bone biology. _

1. Rodan GA, Martin TJ. Therapeutic approaches to bone diseases. Science. 2000;289:1508-1514.
2. Twells RC, Metzker ML, Brown SD, et al. The sequence and gene characterization of a 400-kb candidate region for IDDM4 on chromosome 11q13. Genomics. 2001;72:231-242.
3. Hey PJ, Twells RCJ, Phillips MS, et al. Cloning of a novel member of the lowdensity lipoprotein receptor family. Gene. 1998;216:103-111.
4. Figueroa DJ, Hess JF, Ky B, Brown SD, et al. Expression of the type I diabetesassociated gene LRP5 in macrophages, vitamin A system cells, and the Islets of Langerhans suggests multiple potential roles in diabetes. J Histochem Cytochem. 2000;48:1357-1368.
5. Kim DH, Inagaki Y, Suzuki T, et al. A new low density lipoprotein receptor related protein, LRP5, is expressed in hepatocytes and adrenal cortex, and recognized apolipoprotein E. J Biochem. 1998;124:1072-1076.
6. Fujino T, Asaba H, Kang MJ, et al. Low-density lipoprotein receptor-related protein 5 (LRP5) is essential for normal cholesterol metabolism and glucose-induced insulin secretion. Proc Natl Acad Sci U S A. 2003;100:229-234.
7. Luo DF, Buzzetti R, Rotter JI, et al. Confirmation of three susceptibility genes to insulin-dependent diabetes mellitus: IDDM4, IDDM5 and IDDM8. Hum Mol Genet. 1996;5:693-698.
8. Gong Y, Slee RB, Fukai N, et al. LDL receptor-related protein 5 (LRP5) affects bone accrual and eye development. Cell. 2001;107:513-523.
9. Johnson ML, Gong G, Kimberling W, Recker SM, Kimmel DB, Recker RR. Linkage of a gene causing high bone mass to human chromosome 11(11q12-13). Am J Hum Genet. 1997;60:1326-1332.
10. Kato M, Patel MS, Levasseur R, et al. Cbfa1-independent decrease in osteoblast proliferation, osteopenia, and persistent embryonic eye vascularization in mice deficient in Lrp5, a Wnt coreceptor. J Cell Biol. 2002;157:303-314.
11. Glass DA 2nd, Karsenty G. In vivo analysis of Wnt signaling in bone. Endocrinology. 2007;148:2630-2634.
12. Lobov IB, Rao S, Carroll TJ, et al. WNT7b mediates macrophage-induced programmed cell death in patterning of the vasculature. Nature. 2005;437: 417-421.
13. Junge HJ, Yang S, Burton JB, et al. TSPAN12 regulates retinal vascular development by promoting Norrin- but not Wnt-induced FZD4/beta-catenin signaling. Cell. 2009;139:299-311.
14. Ye X, Wang Y, Cahill H, et al. Norrin, frizzled-4, and Lrp5 signaling in endothelial cells controls a genetic program for retinal vascularization. Cell. 2009;139: 285-298.
15. Holmen SL, Zylstra CR, Mukherjee A, et al. Essential role of beta-catenin in postnatal bone acquisition. J Biol Chem. 2005;280:21162-21168.
16. Day TF, Guo X, Garrett-Beal L, Yang Y.Wnt/beta-catenin signaling in mesenchymal progenitors controls osteoblast and chondrocyte differentiation during vertebrate skeletogenesis. Dev Cell. 2005;8:739-750.
17. Hill TP, Spater D, Taketo MM, Birchmeier W, Hartmann C. Canonical Wnt/betacatenin signaling prevents osteoblasts from differentiating into chondrocytes. Dev Cell. 2005;8:727-738.
18. Li X, Ominsky MS, Warmington KS, et al. Sclerostin antibody treatment increases bone formation, bone mass, and bone strength in a rat model of postmenopausal osteoporosis. J Bone Miner Res. 2009;24:578-588.
19. Yadav VK, Ryu JH, Suda N, et al. Lrp5 controls bone formation by inhibiting serotonin synthesis in the duodenum. Cell. 2008;135:825-837.
20. Sawakami K, Robling AG, Ai M, et al. The Wnt co-receptor LRP5 is essential for skeletal mechanotransduction but not for the anabolic bone response to parathyroid hormone treatment. J Biol Chem. 2006;281:23698-23711.
21. Robinson JA, Chatterjee-Kishore M, Yaworsky PJ, et al. Wnt/beta-catenin signaling is a normal physiological response to mechanical loading in bone. J Biol Chem. 2006;281:31720-31728.
22. Rapport MM, Green AA, Page IH. Serum vasoconstrictor, serotonin; isolation and characterization. J Biol Chem. 1948;176:1243-1251.
23. Gershon MD, Tack J. The serotonin signaling system: from basic understanding to drug development for functional GI disorders. Gastroenterology. 2007; 132:397-414.
24. Mann JJ, McBride PA, Brown RP, et al. Relationship between central and peripheral serotonin indexes in depressed and suicidal psychiatric inpatients. Arch Gen Psychiatry. 1992;49:442-446.
25. Walther DJ, Peter JU, Bashammakh S, et al. Synthesis of serotonin by a second tryptophan hydroxylase isoform. Science. 2003;299:76.
26. Matsuda M, Imaoka T, Vomachka AJ, et al. Serotonin regulates mammary gland development via an autocrine-paracrine loop. Dev Cell. 2004;6:193-203.
27. Lesurtel M, Graf R, Aleil B, et al. Platelet-derived serotonin mediates liver regeneration. Science. 2006;312:104-107.
28. Modder UI, Achenbach SJ, Amin S, Riggs BL, Melton LJ, Khosla S. Relation of serum serotonin levels to bone density and structural parameters in women. J Bone Miner Res. 2009;10.1359/jbmr.090721.
29. Saarinen A, Saukkonen T, Kivela T, et al. Low density lipoprotein receptor– related protein 5 (LRP5) mutations and osteoporosis, impaired glucose metabolism and hypercholesterolaemia. Clin Endocrinol (Oxf). 2009;72:481-488.
30. Frost M, Andersen T, Yadav V, Brixen K, Karsenty G, Kassem M. Patients with high-bone-mass phenotype due to Lrp5-T253I mutation have low plasma levels of serotonin. J Bone Miner Res. 2010;25:673-675.
31. Yadav VK, Oury F, Suda N, et al. A serotonin-dependent mechanism explains the leptin regulation of bone mass, appetite, and energy expenditure. Cell. 2009; 138:976-989.
32. Friedman JM, Halaas JL. Leptin and the regulation of body weight in mammals. Nature. 1998;395:763-770.
33. Spiegelman BM, Flier JS. Obesity and the regulation of energy balance. Cell. 2001;104:531-543.
34. Grill HJ. Leptin and the systems neuroscience of meal size control. Front Neuroendocrinol. 2010;31:61-78.
35. Friedman JM. Leptin and the regulation of body weight. Harvey Lect. 1999; 95:107-136.
36. Farooqi IS, Matarese G, Lord GM, et al. Beneficial effects of leptin on obesity, T cell hyporesponsiveness, and neuroendocrine/metabolic dysfunction of human congenital leptin deficiency. J Clin Invest. 2002;10:1093-1103.
37. Ducy P, Amling M, Takeda S, et al. Leptin inhibits bone formation through a hypothalamic relay: a central control of bone mass. Cell. 2000;100:197-207.
38. Elefteriou F, Ahn JD, Takeda S, et al. Leptin regulation of bone resorption by the sympathetic nervous system and CART. Nature. 2005;434:514-520.
39. Takeda S, Elefteriou F, Levasseur R, et al. Leptin regulates bone formation via the sympathetic nervous system. Cell. 2002;111:305-317.
40. Balthasar N, Coppari R, McMinn J, et al. Leptin receptor signaling in POMC neurons is required for normal body weight homeostasis. Neuron. 2004;42: 983-991.
41. Scott MM, Lachey JL, Sternson SM, et al. Leptin targets in the mouse brain. J Comp Neurol. 2009;514:518-532.
42. Moore RY, Halaris AE, Jones BE. Serotonin neurons of the midbrain raphe: ascending projections. J Comp Neurol. 1978;180:417-438.
43. Michelson D, Amsterdam JD, Quitkin FM, et al. Changes in weight during a 1-year trial of fluoxetine. Am J Psychiatry. 1999;156:1170-1176.
44. Richards JB, Papaioannou A, Adachi JD, et al. Effect of selective serotonin reuptake inhibitors on the risk of fracture. Arch Intern Med. 2007;167:188-194.
45. Dhillon H, Zigman JM, Ye C, et al. Leptin directly activates SF1 neurons in the VMH, and this action by leptin is required for normal body-weight homeostasis. Neuron. 2006;49:191-203.
46. Mann JJ. The medical management of depression. N Engl J Med. 2005;353: 1819-1834.
47. Heath MJ, Hen R. Serotonin receptors. Genetic insights into serotonin function. Curr Biol. 1995;5:997-999.
48. Gustafsson BI, Westbroek I, Waarsing JH, et al. Long-term serotonin administration leads to higher bone mineral density, affects bone architecture, and leads to higher femoral bone stiffness in rats. J Cell Biochem. 2006;97:1283-1291.
49. Battaglino R, Vokes M, Schulze-Spate U, et al. Fluoxetine treatment increases trabecular bone formation in mice. J Cell Biochem. 2007;100:1387-1394.
50. Warden SJ, Robling AG, Sanders MS, Bliziotes MM, Turner CH. Inhibition of the serotonin (5-hydroxytryptamine) transporter reduces bone accrual during growth. Endocrinology. 2005;146:685-693.
51. Heisler LK, Cowley MA, Kishi T, et al. Central serotonin and melanocortin pathways regulating energy homeostasis. Ann N Y Acad Sci. 2003;994:169- 174.
52. Liu Q, Yang Q, Sun W, et al. Discovery and characterization of novel tryptophan hydroxylase inhibitors that selectively inhibit serotonin synthesis in the gastrointestinal tract. J Pharmacol Exp Ther. 2008;325:47-55.
53. Yadav VK, Balaji S, Suresh PS, et al. Pharmacological inhibition of gut-derived serotonin synthesis is a potential bone anabolic treatment for osteoporosis. Nat Med. 2010;16:308-312.

Keywords: serotonin; gut; bone; osteoblast; osteoclast

Download this issue

Anne TETI, PhD

Nadia RUCCI, MD
Department of Experimental
Medicine, University of L’Aquila
L’Aquila, ITALY

Bone is a tissue of central importance, maintaining several relationships with other organs. Among these, the immune system, with which it shares molecular pathways, transcription factors, and several cytokines responsible for bone and immune cell regulation. A paradigm of this crosstalk comes from the studies of Hiroshi Takayanagi on the mechanisms underlying the development of rheumatoid arthritis, demonstrating the central role of a subset of T lymphocytes in the induction of exaggerated osteoclast activity, thus leading to erosion in the affected joints. RANKL/RANK (receptor activator of nuclear factor–kappaB [ligand]) is an important pathway shared by bone and the immune system. This pathway is essential for both osteoclastogenesis and lymphocyte differentiation, so that diseases due to inactivating mutations of RANKL or RANK, such as osteopetrosis, result in immunological defects in addition to altered bone phenotype. This review focuses on the description of the principal molecules/pathways shared with the immune system, which under both physiological and pathological conditions, regulate bone remodeling by acting on osteoclast formation and activity. We propose that the evidence available today strongly points to the osteoclast as a cell with immunological properties, in addition to its role in bone resorption.

The perception of bone as a static organ has changed dramatically over the past several years. The literature has clearly shown that bone is a tissue of central importance and that, in addition to its role in locomotion and in the regulation of calcium and phosphate homeostasis, bone actively maintains multiple relationships with other organs.

Recent observations have evidenced crucial crosstalk between bone and the immune system, thus leading to the launch of a new interdisciplinary field, osteoimmunology.¹ Indeed, several cytokines, molecular pathways, and transcription factors are shared by the immune and skeletal systems. Moreover, immune cells, like bone cells, arise from hematopoietic stem cells (HSCs) found in the bone marrow, which is physically as well as functionally associated with bone tissue. Interestingly, cell differentiation from HSCs has been shown to be subject to a fine regulation by the osteoblasts, which form the HSC niche.² Kollet et al have consistently found that, once subjected to specific stressful stimuli, activated osteoclasts degrade endosteal components, thus promoting the mobilization of hematopoietic progenitors.³

Studies on autoimmune diseases, such as rheumatoid arthritis, performed by Hiroshi Takayanagi, have provided a pivotal contribution in development of the field of osteoimmunology, with identification of a subset of a T cell population that produces high quantities of interleukin (IL)-17, a pro-osteoclastogenic cytokine that increases osteoclast differentiation by direct and indirect mechanisms, thus leading to bone destruction.¹ Conversely, animal models lacking molecules pivotal for the regulation of the immune system frequently show an abnormal osteoclast phenotype.¹

Based on this evidence, we believe that a more extensive investigation of the mechanisms underlying the bone-immune interplay could allow the identification of new strategies for the management of immune system and bone disorders. In this review, we summarize the recent findings that have contributed to consolidation of the field of osteoimmunology, with particular focus on the close relationship between the osteoclasts and the immune cells.

The bone remodeling process

It is well known that bone tissue is in dynamic flux, continually renewed lifelong by a physiological process termed bone remodeling.^4,5 This process is mandatory for the replacement of immature bone with mechanocompetent bone, as well as for repair of fractures and for proper calcium balance. Indeed, it has been estimated that at least 10% of bone is renewed per year.

Bone remodeling follows the activation-resorption-formation (ARF) sequence (Figure 1). The first step, called the activation phase, starts with stimulation of the lining cells, quiescent osteoblasts, which, in response to appropriate stimuli, increase their own surface expression of receptor activator of the nuclear factor-kappaB (NF-κB) ligand (RANKL), which in turn interacts with its receptor RANK (receptor activator of NF-κB), expressed by preosteoclasts. RANKL/RANK interaction triggers preosteoclast fusion and differentiation to multinucleated osteoclasts. Once differentiated, osteoclasts polarize, adhere to the bone surface, and dissolve bone (resorption phase), then they undergo apoptosis, which is a physiological process, required to prevent excessive bone resorption.

After this resorptive process, there is an intermediate phase preceding bone formation, called a reversal phase, during which some macrophage-like uncharacterized mononuclear cells are observed at the site of remodeling, whose function consists of removal of debris produced during matrix degradation.

The final step, bone formation, is triggered by several growth factors stored in the bone matrix and released after its degradation, including bone morphogenetic proteins (BMPs), fibroblast growth factors (FGFs), and transforming growth factor–β (TGFβ), which are likely to be responsible for recruitment of osteoblasts in the resorbed area. Once recruited, osteoblasts produce new bone matrix, initially not mineralized (osteoid), and then they promote its mineralization, thus completing the bone remodeling process.

Under physiological conditions, the coupling of bone formation with previous resorption occurs faithfully. In contrast, an imbalance between the resorption and formation reflects improper bone remodeling, which in turn affects the bone mass, eventually leading to a pathological condition.

Osteoblast regulation of osteoclastogenesis

Although the principal function of the osteoblasts is to synthesize bone matrix proteins and to promote the process of mineralization, a crucial role of osteoblasts in osteoclast biology has been clearly demonstrated by the release of key molecules that regulate osteoclastogenesis and bone resorption. Osteoclasts are multinucleated cells that arise from the monocyte/ macrophage cell line.⁶ Starting from multipotent HSCs, transcription factor PU.1, along with the macrophage-colony stimulating factor (M-CSF), allows the commitment toward a common progenitor for macrophages and osteoclasts (Figure 2). In particular, PU.1 positively regulates the M-CSF receptor, c-Fms, while M-CSF stimulates proliferation of osteoclast precursors and upregulates RANK expression. With the expression of c-Fms and RANK receptors, the precursors become fully committed to osteoclast lineage.⁷ The main source of RANKL in bone is the osteoblast, which expresses RANKL on its membrane surface, thus inducing osteoclast differentiation by interacting with the RANK receptor expressed by the osteoclast precursors. Therefore, triggering of the RANKL/RANK pathway requires a cell-cell contact (Figure 3, page 344). However, lower quantities of soluble RANKL are also released after enzymatic cleavage of the surface molecule by metalloproteinase (MMP)-14. Another key molecule produced by osteoblasts that interfere with the RANKL/RANK pathway is osteoprotegerin (OPG), a decoy receptor for RANKL⁸ with an osteoprotective role. Indeed, OPG is a secreted protein sharing the same structure of the extracellular domain of RANK so that it binds RANKL, preventing its interaction with RANK and subsequent inhibition of osteoclastogenesis.

Figure 1. The bone remodeling process.

Figure 2. Schematic representation of osteoclastogenesis.

RANKL/RANK signaling

RANKL is a type II membrane protein belonging to the TNF superfamily, while its receptor RANK is a type I membrane protein. Osteotropic hormones and factors such as 1,25-dihydroxyvitamin D₃ [1,25(OH)₂D₃], parathyroid hormone (PTH), prostaglandin E₂ (PGE₂), and IL-11 upregulate the expression of RANKL in osteoblast/stromal cell plasma membrane. As previously mentioned, RANKL interacts with its receptor RANK, located on the preosteoclast surface, which in turn activates signaling by recruiting adaptor molecules belonging to the TNF-receptor–associated factors (TRAF) family (Figure 3). Indeed, the RANK cytoplasmic tail contains three binding sites for TRAF6⁹ and this interaction is mandatory for osteoclast differentiation, since TRAF6 knockout mice develop osteopetrosis.¹⁰ Binding of TRAF⁶ to RANK induces trimerization of TRAF6, leading to activation of nuclear factor–kappaB (NFkappaB) and of mitogen-activated protein kinases (MAPKs).¹¹

Figure 3. RANKL/RANK pathway activation.

NF-κB includes a family of dimeric transcription factors, which reside in the cytoplasmunder nonstimulated conditions. However, they enter the nucleus upon cell stimulation by various factors, including RANKL. NF-κB is central to the osteoclastogenic process since the double knockout of the p52 and p50 subunits leads to blockade of osteoclast formation.¹²

Another transcription factor crucial for osteoclast differentiation is activator protein 1 (AP-1) complex, which consists of c-Fos, c-Jun, and ATF proteins. In particular, c-Fos is specifically induced by RANK and is critical for osteoclastogenesis, since c-Fos knockout mice develop osteopetrosis due to the lack of osteoclasts.¹³

NF-κB upregulates the expression of another key molecule for osteoclast differentiation, nuclear factor of activated T cells, cytoplasmic 1 (NFATc1) transcription factor.^14,15 This initial induction requires the interaction of NF-κB with NFATc2, which is recruited to the NFATc1 promoter independently of RANKL stimulation.¹⁶ The essential role of NFATc1 in osteoclastogenesis was demonstrated by evidence that NFATc1- deficient embryonic stem cells did not differentiate into osteoclasts, while the ectopic expression of NFATc1 induced osteoclast differentiation also in the absence of RANKL.¹⁷

Chromatin immunoprecipitation experiments revealed that NFATc1 is recruited to the NFATc1 promoter region 24 hours after RANKL stimulation, and this occupancy persists during the terminal differentiation of osteoclasts, thus indicating a mechanism of autoamplification.¹⁸

In cooperation with AP-1, PU.1, NF-κB, and microphthalmia-associated transcription factor (MITF), NFATc1 regulates the transcription of several target genes involved in osteoclast differentiation and function (Figure 3). These include cathepsin K, calcitonin receptor, tartrate-resistant acid phosphatase (TRAcP),^17,19 β3 integrin, osteoclastassociated receptor (OSCAR),⁷ and dendritic cell–specific transmembrane protein (DC-STAMP), the latter involved in osteoclast fusion.

RANKL/RANK signaling is shared by bone and the immune system

When we talk about the role of RANKL/ RANK in the immune system, we need to point out that RANKL, also known as TNF-related activationinduced cytokine (TRANCE) according to the nomenclature of the immune system, and its receptor RANK, were first identified as molecules expressed by T cells and dendritic cells, respectively, and their physical interaction increased the ability of dendritic cells to stimulate naive T cell proliferation as well as dendritic cell survival.

Therefore, the RANKL/RANK pathway was “born” in an immunologic context. At the same time, bone researchers identified the so called osteoclast differentiation factor (ODF), expressed by the osteoblasts, which increased osteoclast formation,²⁰ and OPG, an osteoblast-derived secreted member of the TNF receptor family, which, in contrast, inhibited osteoclast development and bone resorption acting as a decoy receptor. The molecule able to interact with OPG, named OPG-ligand (OPGL),²¹ was then identified. Finally, bone researchers and immunologists joined in the conclusion that RANKL/TRANCE, ODF, and OPGL are the same molecule, and that RANKL-expressing T cells can also activate osteoclasts, thus mimicking the effect of osteoblasts. Based on this evidence it is not surprising to find bone loss in patients with disorders characterized by abnormal activation of the immune system, such as rheumatoid arthritis or other chronic inflammatory diseases.

Immunological role of the RANKL/RANK pathway

As far as the role of RANKL in the immune system is concerned, it has been demonstrated that, in addition to bone phenotype, due to the lack of osteoclasts, RANKL-deficient mice show a defect in the development of secondary lymphoid tissue.²² However, these mice do not present a severe immunodeficiency, likely due to the fact that lack of RANKL in T cells is compensated by CD40.²³ RANKL also seems to be important for mammary development²⁴ and has been found to be involved in inflammatory bowel diseases by stimulating dendritic cells.²⁵

Figure 4. Ig-like receptors and regulation of osteoclastogenesis.

Recent evidence identified a role for RANKL as a chemokine that can attract RANK-expressing tumor cells and osteoclasts,^26,27 thus pointing to a role of this factor in tumor- induced bony metastases.

Finally, a recent study (2009) identified an unexpected role of RANKL/RANK in the central nervous system, showing that this pathway was involved in thermoregulation and central fever response in inflammation.²⁸

RANKL/RANK–linked diseases

The versatility of the RANKL/RANK axis mirrors the complexity of the diseases in which this pathway is lacking. Among them, osteopetrosis is a rare genetic disorder characterized by sclerosis of the skeleton due to reduced or complete lack of osteoclast function and, as a consequence, impairment of bone resorption.²⁹ This disease is clinically very heterogeneous, ranging from severe to asymptomatic. Autosomal recessive osteopetrosis (ARO) is the most severe form of osteopetrosis, usually diagnosed within the first year of life and in patients with a resultant life expectancy of 3 to 4 years. Similar clusters of patients with ARO harbor mutations in the genes encoding for RANKL and RANK.^30,31 In contrast with all the other forms characterized by a normal or increased number of osteoclasts that are unable to resorb bone, obviously this is an osteoclast- poor ARO form.

Importantly, beside bone phenotype, there are also immunological defects, such as hypogammaglobulinemia due to impairment in immunoglobulin-secreting B cells. This is in line with evidence showing the importance of RANKL/RANK in the immune system, which should be taken into account in the management of this form of ARO. Indeed, it has been demonstrated that two ARO patients harboring RANK mutations exhibited impaired fever during pneumonia.²⁸

Ig-like receptors and osteoclast regulation

Beside the well-known RANKL/RANK pathway, osteoblasts can regulate osteoclast differentiation by interacting with immunoglobulin (Ig)-like receptors, such as OSCAR, whose ligand has not yet been clearly identified. These receptors are associated with immunoreceptor tyrosine-based activation motif (ITAM)-harboring adaptor molecules DAP12 and Fcreceptor common gamma subunit (FcRγ). The role of the latter molecules in osteoclast regulation has been underlined by evidence that mice deficient in both DAP12 and FcRγ have an osteopetrotic phenotype.³² Phosphorylation of the ITAM sequence in DAP12 or FcRγ, resulting after RANK activation, allows the recruitment of splenocyte tyrosine kinase (SYK) and resultant activation of phospholipase C gamma (PLCγ), which in turn triggers calcium signaling. Calcium signaling promotes osteoclastogenesis by activating calcium/calmodulin-dependent protein kinase type IV (CAMKIV), which concurs to c-Fos and calcineurin activation, both cooperating to potentiate NFATc1 autoamplification (Figure 4).¹ Among the molecules that have a dual role in the regulation of immune cells and osteoclasts, a recent study³³ identified the transcription factor B lymphocyte-induced maturation protein–1 (Blimp1). This is a transcriptional repressor involved in the differentiation of B lymphocytes toward plasma cells by direct repression of the transcription factors Pax5, Bcl, and Myc.³⁴ Nishikawa and colleagues demonstrated that Blimp1 stimulates osteoclastogenesis by repressing the transcription factors IFN regulatory factor-8 (IRF-8) and v-Maf musculo-aponeurotic fibrosarcoma oncogene family, protein B (MafB), both negatively affecting osteoclastogenesis.^35,36

Inflammatory cytokines and osteoclastogenesis

Among the cells of the immune system, T lymphocytes play a crucial role in the regulation of osteoclastogenesis, which is indeed the result of a balance between positive and negative factors produced by T cells. As far as the RANKL/ RANK pathway is concerned, it has been demonstrated that activated T cells express RANKL on their surface, thus activating osteoclastogenesis by cell–cell contact.³⁷ Activated T cells also produce IL-10, IL-12, and IL-18, which, in contrast, negatively affect osteoclastogenesis.³⁸ As described below, the CD4+ T helper (T_H)-cell subset T_H1 and T_H2 produce interferon gamma (IFN-γ), which suppresses RANKL signaling by degrading TRAF6, and IL-4, another cytokine with an anti-osteoclastogenic role.

Other cells of the immune system, such as the macrophages, contribute to osteoclast differentiation and function by producing IL-1, IL-6, and TNF-α.^20,39 Moreover, a recent study⁴⁰ showed that lipopolysaccharides (LPS) upregulated the expression of membrane RANKL in human blood neutrophils. LPS-activated neutrophils were then able to stimulate osteoclastogenesis and bone resorption in co-cultures with osteoclast precursors.

Finally, a recent report from Rifas and Weitzmann⁴¹ described the identification of a new T cell cytokine, called secreted osteoclastogenic factor of activated T cells (SOFAT), which induces both osteoblastic IL-6 production and osteoclast formation in the absence of osteoblasts or RANKL, and was insensitive to the effects of the RANKL inhibitor OPG.

Immune diseases and osteoclast activation

_ Rheumatoid arthritis
One of the milestones that was pivotal in defining the new field of osteoimmunology came from research by Takayanagi et al on rheumatoid arthritis.¹ This is an autoimmune disease characterized by inflammation of synovial joints, with CD4+ T-lymphocyte infiltration and synovial cell proliferation, leading to severe bone destruction mediated by osteoclasts.⁴² The clinical feature of bone loss is not restricted to the affected joints, since systemic osteoporosis can also occur.⁴³ Although the increased inflammatory cytokine levels present in affected joints may contribute to enhanced osteoclastogenesis, the mechanism of systemic osteoporosis associated with arthritis remains unclear.

Recent reports have highlighted the crucial role of osteoclasts in the development of this disease. Indeed, it has been demonstrated that osteoclast-deficient mice were protected from bone erosion in arthritis models, despite the presence of inflammation.^44,45 Moreover, high RANKL expression has been detected specifically in the synovium of rheumatoid arthritis patients.⁴⁶ In rheumatoid arthritis affected joints, different cell types can be found, including macrophages, fibroblasts, dendritic cells, plasma cells, and CD4+ T helper cells. Takayanagi¹ demonstrated that in the latter cell type there is a subpopulation, defined as osteoclastogenesis TH cells (THO cells), which, at variance with CD4+ TH 1 cells, does not produce the anticlastogenic cytokines IFN-γ and IL-4, but secretes high quantities of IL-17. This cytokine, in turn, triggers RANKL expression by synovial fibroblasts. IL-17 also stimulates local inflammation, thus inducing macrophages to secrete proinflammatory cytokines such as TNF-α, IL-1, and IL-6. These cytokines in turn activate osteoclastogenesis, directly as well as by stimulating RANKL expression by synovial fibroblasts. Finally, it has been shown that T_HO cells themselves express RANKL, thus activating osteoclastogenesis by direct induction of precursor differentiation.

_ Psoriatic arthritis
This is a disease characterized by musculoskeletal inflammation, and several studies have reported the crucial role of TNF in its pathogenesis. Elevated levels of TNF have been found in the sera, synovial fluid, and synovial membranes of psoriatic patients.⁴⁷ A marked reduction in inflammation and progressive joint damage was consistently observed in patients treated with anti-TNF drugs, which is not only due to their ability to reduce inflammation, but also to reduce osteoclast activation, since it is well known that TNF promotes osteoclast formation. On the other hand, recent reports showed that TNF can affect bone formation by inducing Dickkopf-related protein 1 (DKK-1) to impair bone-forming osteoblast development via inhibition of Wnt signaling.

Is the osteoclast an immune cell?

Based on the aforementioned evidence, it has been hypothesized that osteoclasts are cells that belong to the immune system.⁴⁸ This raises the question as to why there is a need for an immune cell to resorb bone.⁴⁹ Chambers⁵⁰ had previously proposed that the bone matrix is recognized by osteoclasts as a peculiar “foreign body.” In fact, as described in the above (see the bone remodeling process), during the resting condition the bone matrix is covered by a layer of osteoblasts, or lining cells (Figure 1), which segregates the bone matrix from the interstitial fluid, thus probably preventing recognition by the immune system. An external stimulus, such as an inflammatory response, or exposure to PTH/parathyroid hormone– related protein (PTHrP), could trigger osteoblast retraction, so that the mineralized bone matrix can be exposed and recognized as a “foreign body” by immune cells, which have all the requirements to induce osteoclast formation and bone resorption.

Under physiological conditions, this process, once activated, must be switched off and, in fact, there are several paracrine and autocrine mechanisms that negatively regulate osteoclast activity. Consequently, osteoblasts are recalled in the previously resorbed site where they refill the lacunae with newformed bone matrix and again segregate the bone surface from the interstice so that “foreign bone” is no longer exposed to the immune system. If the negative regulation of osteoclast activity fails, this process proceeds longer than necessary, thus resulting in excess bone resorption, with pathological consequences.

Conclusions

It is now clear that bone is a tissue of central importance, therefore, when we study the molecular mechanisms underlying bone remodeling and bone pathological events, we cannot ignore its multiple interactions with other tissues. The discipline of osteoimmunology has shown that osteoclasts and immune cells share a common origin. These two types of cells arise from the HSCs in the bone marrow, another organ closely related to bone. Immunology has also clarified the involvement of bone cells in the development of diseases initially classified in an immunological context, and has identified the central role of some cytokines, known to be produced by immune cells, in the regulation of bone cells. Furthermore, recent advances suggest the potential involvement of osteoclasts and osteoblasts in the regulation of HSCs directed to an immunological commitment.We believe that these findings should encourage immunologists and bone researchers to continue investigating this field, all the more so as better understanding of the relationships between bone and immune cells could help identify new strategies for the management of patients suffering from bone diseases. _

1. Takayanagi H. Osteoimmunology: shared mechanisms and crosstalk between the immune and bone systems. Nat Rev Immunol. 2007;7:292-304.
2. Calvi LM, Adams GB, Weibrecht KW, et al. Osteoblastic cells regulate the hematopoietic stem cell niche. Nature. 2003;425:841-846.
3. Kollet O, Dar A, Shivtiel S, et al. Osteoclasts degrade endosteal components and promote mobilization of hematopoietic progenitor cells. Nat Med. 2006; 12:657-664.
4. Zaidi M. Skeletal remodelling in health and disease. Nat Med. 2007;13:791-801.
5. Datta HK, Ng WF, Walker JA, et al. The cell biology of bone metabolism. J Clin Pathol. 2008;61:577-587.
6. Udagawa N, Takahashi N, Akatsu T, et al. Origin of osteoclasts: mature monocytes and macrophages are capable of differentiating into osteoclasts under a suitable microenvironment prepared by bone marrow-derived stromal cells. Proc Natl Acad Sci U S A. 1990;87:7260-7264.
7. Teitelbaum SL. Osteoclasts: what do they do and how do they do it? Am J Pathol. 2007;170:427-435.
8. Simonet WS, Lacey DL, Dunstan CR, et al. Osteoprotegerin: a novel secreted protein involved in the regulation of bone density. Cell. 1997;89:309-319.
9. Gohda J, Akiyama T, Koga T, et al. RANK-mediated amplification of TRAF6 signaling leads to NFATc1 induction during osteoclastogenesis. EMBO J. 2005; 24:790-799.
10. Naito A, Azuma S, Tanaka S, et al. Severe osteopetrosis, defective interleukin- 1 signalling and lymph node organogenesis in TRAF6-deficient mice. Genes Cell. 1999;4:353-362.
11. Kobayashi N, Kadono Y, Naito A, et al. Segregation of TRAF6-mediated signalling pathways clarifies its role in osteoclastogenesis. EMBO J. 2001;20: 1271-1280.
12. Franzoso G, Carlson L, Xing L, et al. Requirement for NF-kappaB in osteoclast and B-cell development. Genes Dev. 1997;11:3482-3496.
13. Wang ZQ, Ovitt C, Grigoriadis AE, et al. Bone and haematopoietic defects in mice lacking c-fos. Nature. 1992;360:741-745.
14. Takatsuna H, Asagiri M, Kubota T, et al. Inhibition of RANKL-induced osteoclastogenesis by (-)-DHMEQ, a novel NF-κB inhibitor, through downregulation of NFATc1. J Bone Miner Res. 2005;20:653-662.
15. Li F, Matsuo K, Xing L, et al. Over-expression of activated NFATc1 plus RANKL rescues the osteoclastogenesis defect of NF-κB p50/p52 double knockout splenocytes. J Bone Miner Res. 2004;19:S2.
16. Asagiri M, Sato K, Usami T, et al. Autoamplification of NFATc1 expression determines its essential role in bone homeostasis. J Exp Med. 2005;202:1261-1269.
17. Takayanagi H, Kim S, Koga T, et al. Induction and activation of the transcription factor NFATc1 (NFAT2) integrate RANKL signalling for terminal differentiation of osteoclasts. Dev Cell. 2002;3:889-901.
18. Takayanagi H. The Role of NFAT in Osteoclast Formation. Ann N Y Acad Sci. 2007;1116:227-237.
19. Matsuo K, Galson DL, Zhao C, et al. Nuclear factor of activated T-cells (NFAT) rescues osteoclastogenesis in precursors lacking c-Fos. J Biol Chem. 2004; 279:26475-26480.
20. Suda T, Takahashi N, Udagawa N, et al. Modulation of osteoclast differentiation and function by the new members of the tumor necrosis factor receptor and ligand families. Endocr Rev. 1999;20:345-357.
21. Lacey DL, Timms E, Tan HL, et al. Osteoprotegerin ligand is a cytokine that regulates osteoclast differentiation and activation. Cell. 1998;93:165-176.
22. Kong YY, Feige U, Sarosi I, et al. Activated T cells regulate bone loss and joint destruction in adjuvant arthritis through osteoprotegerin ligand. Nature. 1999; 402:304-309.
23. Bachmann MF, Wong BR, Josien R, et al. TRANCE, a tumor necrosis factor family member critical for CD40 ligand-independent T helper cell activation. J Exp Med. 1999;189:1025-1031.
24. Fernandez-Valdivia R, Mukherjee A, Ying Y, et al. The RANKL signaling axis is sufficient to elicit ductal side-branching and alveologenesis in the mammary gland of the virgin mouse. Dev Biol. 2009;328:127-139.
25. Moschen AR, Kaser A, Enrich B et al. The RANKL/OPG system is activated in inflammatory bowel disease and relates to the state of bone loss. Gut. 2005;54: 479-487.
26. Jones DH. Nakashima T, Sanchez OH, et al. Regulation of cancer cell migration and bone metastasis by RANKL. Nature. 2006;440:692-696.
27. Rucci N, Millimaggi D, Mari M, et al. Receptor activator of NF-êB ligand enhances breast cancer–induced osteolytic lesions through upregulation of extracellular matrix metalloproteinase inducer/CD147. Cancer Res. 2010 Jul 14. [Epub ahead of print] PMID:20631064.
28. Hanada R, Leibbrandt A, Hanada T, et al. Central control of fever and female body temperature by RANKL/RANK. Nature. 2009;462:505-509.
29. Whyte MP. Osteopetrosis. In: Royce PM, Steinman B, eds. Connective Tissue and its Heritable Disorders: Medical, Genetic, and Molecular Aspects. New York, NY: Wiley-Liss; 2002:753-770.
30. Sobacchi C, Frattini A, Guerrini MM, et al. Osteoclast-poor human osteopetrosis due to mutations in the gene encoding RANKL. Nat Genet. 2007;39:960- 962.
31. Guerrini MM, Sobacchi C, Cassani B, et al. Human osteoclast-poor osteopetrosis with hypogammaglobulinemia due to TNFRSF11A (RANK) mutations. Am J Hum Genet. 2008;83:64-76.
32. Mocsai A, Humphrey MB, Van Ziffle JA, et al. The immunomodulatory adapter proteins DAP12 and Fc receptor α chain (FcRα) regulate development of functional osteoclasts through the Syk tyrosine kinase. Proc Natl Acad Sci U S A. 2004;101:6158-6163.
33. Nishikawa K, Nakashima T, Hayashi M, et al. Blimp1-mediated repression of negative regulators is required for osteoclast differentiation. Proc Natl Acad Sci U S A. 2010;107:3117-3122.
34. Martins G and Calame K. Regulation and functions of Blimp-1 in T and B lymphocytes. Annu Rev Immunol. 2008;26:133-169.
35. Zhao B, Takami M, Yamada A, et al. Interferon regulatory factor-8 regulates bone metabolism by suppressing osteoclastogenesis. Nat Med. 2009;15:1066-1071.
36. Kim K, Kim JH, Lee J, et al. MafB negatively regulates RANKL-mediated osteoclast differentiation. Blood. 2007;109:3253–3259.
37. Kong YY, Yoshida H, Sarosi I, et al. OPGL is a key regulator of osteoclastogenesis, lymphocyte development and lymph-node organogenesis. Nature. 1999; 397:315-323.
38. Takayanagi H, Ogasawara K, Hida S, et al. T-cell-mediated regulation of osteoclastogenesis by signalling cross-talk between RANKL and IFN-gamma. Nature. 2000;408:600-605.
39. Boyle WJ, Simonet WS, Lacey DL. Osteoclast differentiation and activation. Nature. 2003;423:337-342.
40. Chakravarti A, Raquil MA, Tessier P, et al. Surface RANKL of Toll-like receptor 4–stimulated human neutrophils activates osteoclastic bone resorption. Blood. 2009;114:1633-1644.
41. Rifas L.Weitzmann MN. A novel T cell cytokine, secreted osteoclastogenic factor of activated T cells, induces osteoclast formation in a RANKL-independent manner. Arthritis Rheum. 2009;60:3324-3335.
42. Sato K, Takayanagi H. Osteoclasts, rheumatoid arthritis, and osteoimmunology. Curr Opin Rheumatol. 2006;18:419-426.
43. Schett G, Hayer S, Zwerina J, et al. Mechanisms of disease: the link between RANKL and arthritic bone disease. Nat Clin Pract Rheumatol. 2005;1:47-54.
44. Pettit AR, Ji H, von Stechow D, et al. TRANCE/RANKL knockout mice are protected from bone erosion in a serum transfer model of arthritis. Am J Pathol. 2001;159:1689-1699.
45. Redlich K, S. Hayer, R. Ricci, et al. Osteoclasts are essential for TNF-alphamediated joint destruction. J Clin Invest. 2002;110:1419-1427.
46. Takayanagi H, Iizuka H, Juji T, et al. Involvement of receptor activator of nuclear factor kappaB ligand/osteoclast differentiation factor in osteoclastogenesis from synoviocytes in rheumatoid arthritis. Arthritis Rheum. 2000;43: 259-269.
47. Kofi A, Mensah MS, Schwarz EM, et al. Altered bone remodelling in psoriatic arthritis. Curr Rheumatol Rep. 2008;10:311-317.
48. Baron R. Arming the osteoclast. Nat Med. 2004;10:458-460.
49. Del Fattore A, Teti A, Rucci N. Osteoclast receptors and signaling. Arch Biochem Biophys. 2008;47:147-160.
50. Chambers TJ, Darby JA, Fuller K. Mammalian collagenase predisposes bone surfaces to osteoclastic resorption. Cell Tissue Res. 1985;241:671-675.

Keywords: osteoimmunology; bone tissue; immune system; hematopoietic stem cell; osteoclast; cytokine; rheumatoid arthritis; bone remodeling

Download this issue

Bernard CORTET, MD, PhD
Isabelle LEGROUX-GÉROT, MD
Département Universitaire
de Rhumatologie
Université de Lille 2
FRANCE

by B. Cortet and I. Legroux-Gérot, France

Psychiatric diseases may, via direct and indirect mechanisms, induce bone fragility. This is particularly the case with depression and anorexia nervosa. Studies show a moderate decrease in bone mineral density (of the order of 6%) in the spine and hip of depressed patients vs controls. Similarly, a significant increase in fracture risk is observed, with an up to 2-fold increase in hip fracture risk. The mechanisms of bone fragility in depressed subjects are complex, multifactorial, and have yet to be fully elucidated. One of themajor directmechanisms involves endogenous hyperadrenocorticism— which is less pronounced than in Cushing’s syndrome, and may be due in part to a rise in proinflammatory cytokines (notably interleukin 6), which is reported in depressed patients. Also, antidepressant treatment—in particular serotonin reuptake inhibitors—may have a negative impact on bone. Indirect factors, whose role is disputed, include weight loss and cigarette and alcohol abuse, often reported in depressed subjects. Anorexia nervosa (AN) has become a major problem in recent years. AN gives rise to multiple complications and is frequently associated with bone loss, with osteoporosis occurring in 38% to 50%of cases. Estrogen deficiency has long been known to play amajor role, but cannot alone explain bone loss. Recent publications have highlighted the essential role of undernourishment and factors influenced by nutritional status, in particular the growth hormone–insulin-like growth factor I (GH-IGF-I) axis. Themanagement of anorexia nervosa–related bone loss is debated.While restoring menstruation and body weight is mandatory, it does not always ensure correction of bone loss. Studies have failed to show any effectiveness of estrogen treatment.

Certain psychiatric disorders may have a deleterious impact on bone. This occurs via direct and indirect mechanisms, not all of which have been elucidated. Essentially two psychiatric disorders have undergone extensive research to evaluate this relationship: depression and anorexia nervosa (AN). These two diseases will therefore be addressed in this paper.

Depression is a frequent disease, affecting about 16% of the North-American population.¹ One of the first publications to link depression and impact on bone was published by Schweiger et al in 1994.² These authors determined bone mineral density (BMD) by quantitative computerized tomography (QCT) in 70 depressed subjects and 88 controls. The female/male ratio was the same in both groups. The authors found an approximately 15% reduction in BMD in the depressed subjects, after adjustment for age. Subsequently, several articles on the same topic were published in which BMD was measured by dual-energy x-ray absorptiometry (DXA), the current consensus method; but the findings disagreed: certain authors reported a link between depression and low BMD,^3-14 while others found no such link.^15-21 This discrepancy is undoubtedly related to the heterogeneity of the disease itself. In addition, it is to be noted that most of the subjects enrolled in the studies were, quite logically, on antidepressant treatment, often serotonin reuptake inhibitors (SSRIs), known to have an impact on bone.

Figure 1. Mean differences in bone mineral density (BMD) between depressed and nondepressed groups and corresponding 95% confidence intervals (CI) for the spine (A) and hip (B) in 12 studies.

Epidemiology of bone impact in depression

_ Densitometric data
A meta-analysis, avoiding the aforementioned pitfalls, was recently published,²² which included 8 cross-sectional and 6 case-control studies. Cohort studies were also evaluated when densitometric data were available. In all the studies, BMD was determined by DXA. In the cross-sectional studies, confounding factors such as age, gender, menopausal status, weight, and body mass index (BMI) were taken into account in the analysis of the results. The studies are summarized in Figure 1.^{5,7,11-14,16,17,19-24} Of the 14 studies, BMD data for all sites (lumbar spine and hip) were only reported in 12 studies, which were thus finally selected for the meta-analysis. The decrease in BMD was only significant in 6 cases. Mean between-group BMD difference was only slight: 53 mg/cm² (95% confidence interval [CI], 18-87) for the lumbar spine. The difference was very similar for the hip: 52 mg/cm2 (95% CI, 22-83). In the depressed subjects, the percentage decrease in BMD was 5.9% for the lumbar spine and 6% for the hip.

When results were expressed as T-scores and Z-scores, the trend was similar, as expected. The decrease, while real, was onlymodest. Thus,mean T-scores in depressed subjects were –0.73 (95%CI, –1.32/–0.146) for the lumbar spine and –0.627 (95% CI, –1.02/–0.233) for the hip. The Z-scores were again, as expected, fairly similar. Various sensitivity analyses were conducted, but did not change the results. It should, however, be noted that when depression was diagnosed using the Diagnostic and Statistical Manual of Mental Disorders (DSM) criteria, the differences were a little more marked. The results for men, while also significant, were of lesser magnitude.

Depression and fracture risk

Bone densitometry is only a surrogate marker and the most important issue is whether depressed subjects are at greater fracture risk. As is the case for densitometric assessment, this requires taking into account numerous factors well known to influence fracture risk. It should be pointed out that studies aimed at determining fracture risk are few. Moreover, some of them are open to methodological criticism, particularly those of Kessler et al,¹ due to the fact that they are retrospective studies.

Four of the 5 prospective studies on fracture risk available to date concluded that depression was associated with an increase in fracture risk. In the remaining study, by Greendale et al,²⁵ which evidenced no such increase, the authors nonetheless showed that those patients with the highest urinary cortisol level were at increased risk of fracture. The main findings from these studies are summarized in Table I.^15,20,25-27 Thus, for example, in 467 depressed subjects from a cohort of 7414, Whooley et al20 reported an increase in relative nonvertebral fracture risk, of 1.3 (95% CI, 1.1-1.6), after multiple adjustments. Similarly, vertebral fracture risk was 2.1 (95% CI, 1.4-4.2), after multiple adjustments. The authors concluded to a relationship between depression severity and the magnitude of the increase in fracture risk (Table I). Generally speaking, the risk of fracture, particularly peripheral fracture, is conditioned by 3 factors: BMD, the magnitude of the impact, and the angle of the impact. This likely applies to depressed patients as well. Lastly, it was shown that there is an increased risk of falls in depression.^20,28

Table I. Depression and fracture risk.

Pathophysiology of bone impact in depression

_ Hypothalamic-pituitary axis
The various hypotheses are summarized in Figure 2 (page 352). There is substantial evidence to confirm the presence of hyperadrenocorticism in the context of depression. The latter results from chronic exposure to stress, which triggers the release of corticotropin-releasing hormone (CRH) by the paraventricular nucleus of the hypothalamus. The process involves frontal cortex, hippocampus, amygdala, and hypothalamus pathways. There is no autonomous hypersecretion of cortisol in depression, so that cortisol levels are markedly lower than those observed in Cushing’s syndrome. This hypothesis is corroborated by clinical data,^13-23 which show an increase in plasma cortisol. An increase in urinary cortisol has also been observed, though some authors failed to evidence any such increase.²⁹

Figure 2. Pathophysiology of the bone impact of depression.

_ Autonomic nervous system
Animal data suggest that there is a relationship between hyperactivity of the efferent autonomic nervous system and risk of bone demineralization. However, the implications of this finding in depression are still the subject of considerable debate.

_ Leptin
The relationship between leptin metabolism and bone metabolism is complex. Findings relating to circulating leptin levels in depressed subjects are contradictory. Thus, understanding of the pathophysiological role of leptin as a factor liable to explain the bone impact in depression requires further investigation.

_ Impairment of the immune system
Impairment of the immune system in depression has been fairly well established, with an increase in proinflammatory cytokines such as interleukins (IL) 1 and 6 and tumor necrosis factor.³⁰ Cytokines are stimulants of the hypothalamic-pituitary- adrenal axis, which may account for the hyperadrenocorticism observed in depression. This has been confirmed by a recent study by Eskandari et al,²⁹ which also reported a reduction in anti-inflammatory cytokines (IL-10 and IL-13) in depressed subjects. This reduction was only significant for IL-13. However, the study population was small, limiting the scope of the study.

Impact of depression on bone and confounding factors

Confounding factors include the classic risk factors for osteoporosis, which can be present in depressed subjects just as in the general population, and a risk factor specific to depression: antidepressant treatment.

_ Osteoporosis risk factors in depressed subjects
Certain risk factors for bone fragility are more frequently encountered in depressed subjects, eg, tobacco and alcohol abuse. Similarly, one of the symptoms of depression, namely, weight loss, is associated with an increase in fracture risk. These confounding factors are sometimes taken into account in the studies previously cited, but not always, which makes it difficult to interpret findings.

_ Antidepressants and bone metabolism
Antidepressants, in particular SSRIs, undoubtedly are themost important confounding factor. The presence of serotonin receptors on the osteoblasts and osteocytes lends support to the involvement of these agents. Some studies are adjusted to take into account antidepressant intake, but this is not always the case. The adverse effect on bone of SSRIs is supported by in vitro and animals studies. Severe osteoporosis has been evidenced in serotonin-deficient mice. A clinical study by Cauley et al31 showed that only SSRIs (and not tricyclic antidepressants) have an adverse impact on bone. More recently, Diem et al³² were able to show that the effect of SSRIs persisted even after adjustment for the symptoms of depression. They also reported accelerated bone loss in subjects on SSRIs vs nonusers and vs tricyclic antidepressant users.

Obviously, the most important issue is to determine whether antidepressants are associated with an increase in fracture risk. Quite logically, and in line with previous studies, a recent large-scale study³³ in 6763 subjects on tricyclic antidepressants or SSRIs vs 26 341 controls matched for age, gender, and geographic origin, reported an increase in hip and femur fracture risk in patients receiving SSRIs: relative risk (RR), 2.35 (95% CI, 1.94-2.84). An increase in fracture risk, albeit of lesser amplitude, was also found in patients on tricyclic antidepressants: RR, 1.76 (95% CI, 1.45-2.15). The strength of the causal relationship was confirmed by the fact that fracture risk rapidly decreased following antidepressant treatment discontinuation. In patients receiving SSRIs, there was a parallel between the latter’s potency in inhibiting serotonin reuptake and the magnitude of the fracture risk increase. The potential increase in the risk of other fractures due to bone fragility was not clearly established. strength of the causal relationship was confirmed by the fact that fracture risk rapidly decreased following antidepressant treatment discontinuation. In patients receiving SSRIs, there was a parallel between the latter’s potency in inhibiting serotonin reuptake and the magnitude of the fracture risk increase. The potential increase in the risk of other fractures due to bone fragility was not clearly established.

Animal models

Yirmiya et al³⁴ developed an experimental model of depression in the mouse, enabling micro-CT scan analysis of the distal metaphysis of the femur and the vertebrae. In both sites, the authors showed, under physiological and pathological conditions, a decrease in trabecular bone volume, compared with controls.

Anorexia nervosa (NA) has become a major public health concern in industrial countries in recent years. Its prevalence is 0.5%, vs 2%for bulimia. AN is a syndrome combining an exaggerated fear of excessive weight, a disorder of body image, significant weight loss, refusal to maintain a minimum normal weight, and amenorrhea.

The course of the disease is accompanied by a variety of disorders and complications. Bone health is much affected, with a decrease ofmore than 1 standard deviation (SD) in spine and femur neck bone mass in 92% of female patients, which exceeds 2.5 SD in 38% of cases.³⁵ The mechanisms of bone loss in AN patients are multiple: hormonal, endocrine, and nutritional. The disease is more severe when it develops during adolescence, a critical period for acquisition of peak bone mass. Bone mass increases gradually through childhood and accelerates during adolescence to reach a peak during Tanner stages 4 and 5. The greater part of bone mass peak determination seems to be genetic (60% to 80%); the remaining 20% to 40% of determination is influenced by nutritional and hormonal factors.³⁶ For a given age, bone loss is more marked in anorexic women than women with normal BMI and amenorrhea of hypothalamic origin. Forty percent of anorexic women are osteoporotic vs 16% in the second group.³⁷ BMI in women in whom AN develops before age 18 years is significantly lower than those in whom it develops later, reflecting the impact of the disease on bone formation.³⁸

Assessment of the bone impact of anorexia nervosa

_ Bone mineral density
BMD is determined in the spine and femur neck by means of bone densitometry measurements using low-dose radiation. TheWorld Health Organization defines osteoporosis as a BMD that is at least 2.5 SD lower than the mean for young women (T-score < –2.5 SD). However, this definition only applies to postmenopausal women, a fact that must be taken into account when dealing with adolescents who have not always achieved peak bone mass. Lower BMD is consistently reported in anorexic female patients, and osteoporosis is present in about 30% of them.^35,39-41

_ Bone remodeling markers
Bone markers, used to assess bone remodeling, are complementary to BMD determination, but are not diagnostic tools. The most frequently used bone-formation markers are osteocalcin and bone alkaline phosphatase (BAP); bone-resorption markers include deoxypyridinoline (DPD), C-terminal (crosslaps or CTX), and N-terminal (NTX) extension peptides and telopeptides (carboxyl terminal telopeptide of collagen I [ICTP]).^39,40 Thesemarkers aremainly used in postmenopausal women and their interpretation is more difficult in young women and adolescents. The literature shows wide divergence in findings; study populations are frequently small and it is necessary to distinguish between the studies conducted on female adolescents and those conducted on adult anorexic patients. Like postmenopausal women, AN women show an increase in bone resorption, but studies have also shown that there is a marked decrease in bone formation.⁴²

This shows that the bone loss in AN patients is also related to other mechanisms, such as estrogen deficiency, and that nutritional or nutrition-dependent factors are also involved. This is confirmed in the literature by the fact that bone loss in AN patients is more marked than that in women of the same age suffering from hypogonadism.

Few studies have addressed fracture risk in AN populations. Lucas et al⁴³ reported a retrospective study in 208 AN patients over 13 years with 58 fractures. Compared with the expected number of fractures, the risk in AN patients was 3-fold greater. Fractures occurred more frequently in inpatients than in outpatients, and bone insufficiency–related cracks were also more frequent in inpatients. A study in female patients with a mean AN duration of 5.8 years reported a 7-fold greater fracture risk than in healthy women of the same age.⁴⁴ Fractures occurred more frequently at the usual sites (vertebrae, followed by the radius and the distal extremity of the femur).⁴⁴

_ Hormonal factors
Studies of the time course of BMD in female AN patient populations show that when AN is diagnosed before age 18, BMD is significantly lower than when diagnosed at a later age, reflecting the impact of the disease on acquisition of peak bone mass.^{35,39,40,44,45}

Amenorrhea is a diagnostic criterion for AN. Estrogen deficiency is known to play a major role in bone mass loss in the AN population. The mechanisms underlying estrogen deficiency in AN have yet to be fully elucidated. They are probably multifactorial, and include hypothalamic dysfunction, weight loss, and dysregulation of neurotransmitters such as GnRH. The literature shows a correlation between BMD and the duration and age of onset of amenorrhea.^35,39,44-46

Estrogen deficiency alone cannot explain bone loss in anorexic female patients. Bone mass gain precedes resumption of menstrual cycles in recovering anorexic patients, while estrogen therapy does not prevent bone loss in adolescents.⁴² Other factors are involved in bone loss in AN. Bone remodeling is differently affected in AN female patients compared with postmenopausal osteoporotic women. Bone resorption and formation are increased, with a balance in favor of resorption, in postmenopausal women, whereas in AN, although bone resorption is slightly greater, the predominant disorder is decreased bone formation,^37,39 though some authors report that it is normal.³⁵ In any event, it is never increased. Reduced bone formation in AN explains the relative failure of antiresorptive treatments and, particularly, estrogens.⁴² In all, this suggests an essential role for undernourishment and factors influenced by nutritional status in the bone loss of AN.

_ Nutritional and endocrine factors
The role of nutritional and endocrine factors is supported by the literature, which shows a strong correlation between female patient BMD and nutritional indices such as BMI, lean mass, fat mass, insulin-like growth factor–I (IGF-I), and leptin.^37,39 In a previous study, the author and his colleagues showed a correlation between hip BMD and IGF-I in 113 female patients with AN.⁴⁶ Hotta et al⁴⁷ showed that the osteoporotic risk is higher when BMI is less than 15 kg/m². Other authors^39,47 have also reported a correlation between bone formation markers (osteocalcin and BAP) and nutritional markers such as BMI, fat mass percentage, IGF-I, and a negative correlation between estradiol and bone resorption markers.

At puberty, the levels of GH-IGF axis hormones increase to stimulate the proliferation and differentiation of osteoblastic precursors. IGF-I is a bone tropism hormone that stimulates bone formation and growth by acting on osteoblasts and stimulating collagen synthesis. Studies have shown an impairment of the GH-IGF-I axis in AN patients.^48,49 Female AN patients display resistance to GH, with high GH levels, but low IGF-I levels. Stoving et al⁴⁸ monitored 24-hour GH secretion in 8 anorexic female patients and showed an increase in the number, duration, and intensity of GH peaks. The authors also showed an increase in basal secretion (20-fold vs 4 fold for pulsatile secretion). The increase in the intensity of GH peaks is ascribed to weight loss, while the number of peaks is related to hypoestrogenism. There was no difference in GH half-life in anorexic patients compared with healthy controls. Sacchi et al49 published similar results. Several authors have reported a decrease in IGF-I levels, but also in binding proteins (IGFBP), in particular IGFBP3 and 2, in anorexic female patients,^50,51 sometimes with an increase in IGFBP1. The decrease in circulating binding-protein levels may in part explain the resistance to GH, preventing the transfer of IGF-I toward the target organs. In addition, IGFBP3 is reported to be a good predictive factor for bone loss in anorexic patients, independently of BMI and IGF-I.

Figure 3. Relationship between circulating leptin level (abscissa:
tertiles 1, 2 and 3) and bone mineral density expressed as lumbar
spine Z-score (as ordinates).

An important role is played by the hormone leptin, an antiorexigenic adipokine secreted by adipose tissue. Leptin’s physiological effects on bone are debated, particularly as they differ depending on whether its peripheral or central action is considered. Measuring BMD and several hormonal factors in a recent cohort study of 103 young women with AN,⁵² the authors found a mean Z-score of –1.17 for the spine, –1.33 for the hip, and –1.11 for the femur neck, and a modest, but significant, positive correlation between leptin levels and spinal BMD (r = 0.30). The correlations were significant, but of lesser amplitude, for the femur neck and whole hip (r = 0.23 and r = 0.21, respectively). Multiple regression analysis showed that 27%of spinal BMD variability was explained by differences in duration of amenorrhea and leptin levels. Figure 3, which plots BMD values as a function of leptin level divided into 3 tertiles, shows a marked and significant difference between the patients in the lowest tertile (mean Z-score: –1.25) and higher tertile (mean Z-score: +0.75).⁵²

Hyperadrenocorticism and calcium and vitamin D deficiency are reported in AN, in some cases compounded by excessive exercise. Thus, high cortisol levels can be found,^38,46 although the circadian rhythm is spared. Similarly, the dexamethasone suppression test frequently evidences an increase in urinary free cortisol. Hyperadrenocorticism may be related to impairment of hypothalamic function or CRH hypersecretion. Grinspoon et al35 reported hyperadrenocorticism in only 22% of anorexic patients with severe bone loss. We found similar results in our study.⁴⁶ Audi et al⁵³ did not find any signif- icant difference in urinary free cortisol levels between AN patients and controls. This suggests that while hyperadrenocorticism is a potential cause of bone loss, it is not the only mechanism involved. The role of vitamin and calcium deficiency in bone loss remains uncertain. In the study by Audi et al⁵³ vitamin D deficiency (25-OH-D3) was observed in 24.6% of AN patients. Urinary calcium was somewhat higher in the group of AN patients in the active phase, and lower in those having regained weight, but still with amenorrhea, and those who had recovered. Soyka et al⁴⁰ reported dietary calcium deficiency (<1300 mg/day) in 42% of the AN patients in their study population, but also in 50% of the controls. Similarly, vitamin D deficiency was present in 42% of AN patients and 44% of controls.

Course of bone loss after weight recovery

A few studies have addressed the time course of BMD in recovered anorexic patients.^40,44,54 Despite the improvement in bone mass with body weight normalization, certain studies report persistent osteopenia in a high proportion of postanorexic patients. Hartman et al,⁵⁴ in a study of 19 female patients with a history of AN, determined bone mass at age 21 years and found, despite body weight recovery, that femur bone mass was lower than that of the control group. In Zipfel’s study,⁴⁴ monitoring of spine BMD showed bone gain after body weight recovery with a decrease in the percentage of osteopenic and osteoporotic female patients (35% to 13% and 54 to 21%, respectively),but a large proportion of the patients continued to have low BMD values. However, in a recent study, Wentz et al55 did not confirm those results and found no significant difference in BMD between the patients with a history of AN (11 years previously on average) and the controls.

Overall, specific bone-targeting treatments are of little efficacy in AN. As indicated previously, the primary objective is to achieve body weight recovery, which has a proven beneficial impact on bone. Various studies, including a recent one by us, have shown that hormonal treatment is not effective.⁵⁶ In contrast, achieving a BMI greater than 19 kg/m² and resumption of menstrual periods results in bone gains.

In conclusion, bone loss in female patients with AN is rapid and severe and is associated with a substantial fracture risk, the mechanism of which has yet to be fully elucidated and is probably multifactorial. Early screening is necessary and BMD must be determined as soon as AN is diagnosed. _

1. Kessler R, Berglund P, Delmer O, et al. Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Arch Gen Psychiatry. 2005;62:593-602.
2. Schweiger U, Deuschle M, Korner A, et al. Low lumbar bone mineral density in patients with major depression. Am J Psychiatry. 1994;151:1691-1693.
3. Coelho R, Silva C, Maia A, et al. Bone mineral density and depression: a community study of women. J Psychosom Res. 1999;46:29-35.
4. Halbreich U, Rojansky N, Palter S, et al. Decreased bone mineral density in medicated psychiatric patients. Psychosom Med. 1995;57:485-491.
5. Jacka F, Pasco J, Henry MJ, et al. Depression and bone mineral density in a community setting of perimenopausal women: Geelong Osteoporosis Study. Menopause. 2005;12:88-91.
6. Kahl KG, Rudolf S, Stoeckelhuber BM, et al. Bone mineral density, markers of bone turnover, and cytokines in young women with borderline personality disorder with and without comorbid major depressive disorder. Am J Psychiatry. 2005;162:168-174.
7. Michelson D, Stratakis C, Hill L, et al. Bone mineral density in women with depression. N Engl J Med. 1996;335:1176-1181.
8. Robbins J, Hirsch C,Whitmer R, Cauley J, Harris T. The association of bone mineral density and depression in an older population. J Am Geriatr Soc. 2001;49: 732-736.
9. Schweiger U, Weber B, Deuschle M, Heuser I. Lumbar bone mineral density in patients with major depression: evidence of increased bone loss at followup. Am J Psychiatry. 2000;157:118-120.
10. Vrkljan M, Thaller V, Lovricević I, et al .Depressive disorder as a possible risk factor for osteoporosis. Coll Antropol. 2001;25:485-492.
11. Wong S, Lau E, Lynn H, et al. Depression and bone mineral density: is there a relationship in elderly Asian men? Results from Mr. Os (Hong Kong). Osteoporos Int. 2005;16:610-615.
12. Yazici K, Akinci A, Sutcu A, Ozcakar L. Bone mineral density in premenopausal women with major depressive disorder. Psychiatry Res. 2003;117:271-275.
13. Altindag O, Altindag A, Asoglu M, Gunes M, Soran N, Deveci Y. Relation of cortisol levels and bone mineral density among premenopausal women with major depression. Int J Clin Pract. 2007;61:416-420.
14. Mussolino M, Jonas B, Looker A. Depression and bone mineral density in young adults: results from NHANES III. Psychosom Med. 2004;66:533-537.
15. Søgaard AJ, Joakimsen RM, Tverdal A, Fønnebø V, Magnus JH, Berntsen GK. Long-term mental distress, bone mineral density and non-vertebral factures: the Tromso Study. Osteoporos Int. 2005;16:887-897.
16. AmsterdamJ, HooperM. Bonemineral density inmajor depression. Prog Neuro- Psychopharmacol Biol Psychiatry. 1998;22:267-277.
17. Kavuncu V, Kuloglu M, Kaya A, Sahin S, Atmaca M, Firidin B. Bone metabolism and bone mineral density in premenopausal women with mild depression. Yonsei Med J. 2002;43:101-108.
18. Reginster JY, Deroisy R, Paul I, Hansenne M, Ansseau M. Depressive vulnerability is not an independent risk factor for osteoporosis in postmenopausal women. Maturitas. 1999;33:133-137.
19. Whooley M, Cauley J, Zmuda J, Haney E, Glynn NW. Depressive symptoms and bone mineral density in men. J Geriatr Psychiatry Neurol. 2004;17:88-92.
20. Whooley M, Kip K, Cauley J, Ensrud K, Nevitt M, Browner W. Depression, falls and risk of fracture in older women. Arch Intern Med. 1999;159:484-490.
21. Yazici A, Bagis S, Tot S, Sahin G, Yazici K, Erdogan C. Bone mineral density in premenopausal women with major depression. Joint Bone Spine. 2005;72:540- 543.
22. Wu Q, Magnus JH, Liu J, Bencaz AF. Depression and low bone mineral density: a meta-analysis of epidemiologic studies. Osteoporos Int. 2009;20:1309-1320.
23. Kahl KG, Greggersen W, Rudolf S, et al. Bone mineral density, bone turnover, and osteoprotegerin in depressed women with and without borderline personality disorder. Psychosom Med. 2006;68:669-674.
24. Petronijevic M, Petronijevic N, Ivkovic M, et al. Low bone mineral density and high bone metabolism turnover in premenopausal women with unipolar depression. Bone. 2008;42:582-590.
25. Greendale G, Unger J, Rowe J, Seeman TE. The relation between cortisol excretion and fractures in healthy older people: results from the MacArthur studies. J Am Geriatr Soc. 1999;47:799-803.
26. Forsén L, Meyer HE, Søgaard AJ, Naess S, Schei B, Edna TH. Mental distress and risk of hip fracture: do broken hearts lead to broken bones? J Epidemiol Commun Health. 1999;53:343-347.
27. Mussolino M. Depression and hip fracture: the NHANES I. Epidemiologic followup study. Public Health Rep. 2005;120:71-75.
28. Korpelainen R, Korpelainen J, Heikkinen J, Väänänen K, Keinänen-Kiukaanniemi S. Lifelong risk factors for osteoporosis and fractures in elderly women with low body mass index-a population-based study. Bone. 2006;39:385-391.
29. Eskandari F, Martinez PE, Torvik S, et al; Premenopausal, OsteoporosisWomen, Alendronate, Depression (POWER) Study Group. Low bone mass in premenopausal women with depression. Arch Intern Med. 2007;167, 2329-2336.
30. Marques-Deak A, Cizza A, Sternberg E. Brain-immune interactions and disease susceptibility. Mol Psychiatry. 2005;10:239-250.
31. Cauley J, Fullman R, Stone K, et al. Factors associated with the lumbar spine and proximal femur bone mineral density in older men. Osteoporos Int. 2005; 16:1525-1537.
32. Diem S, Blackwell T, Stone K, et al. Use of antidepressants and rates of hip bone loss in older women. Arch Intern Med. 2007;167:1240-1245.
33. Van den Brand MWM, Samson MM, Pouwels S, van Staa TP, Thio B, Cooper C et al. Use of anti-depressants and the risk of fracture of the hip or femur. Osteoporos Int. 2009;20:1705-1713.
34. Yirmiya, R, Goshen I, Bajayo A, et al. Depression induces bone loss through stimulation of the sympathetic nervous system. Proc Natl Acad Sci U S A. 2006;103:16876-16881
35. Grinspoon S, Thomas E, Pitts S, et al. Prevalence and predictive factors for regional osteoporosis in women with anorexia nervosa. Ann Intern Med. 2000; 133:790-794.
36. Heaney RP, Abrams S, Dawson-Hughes B, et al. Peak bone mass. Osteoporos Int. 2000;11:985-1009.
37. Grinspoon S, Miller K, Coyle C, Krempin J, Armstrong C, Pitts S, Herzog D, Klibanski A. Severity of osteopenia in oestrogen-deficient women with anorexia nervosa and hypothalamic amenorrhea. J Clin Endocrinol Metab. 1999;84: 2049-2055.
38. Biller BMK, Saxe V, Herzog DB, Rosenthal DI, Holzman S, Klibanski A. Mechanisms of osteoporosis in adult and adolescents women with anorexia nervosa. J Clin Endocrinol Metab. 1989;68:548-554.
39. Soyka LA, Grinspoon S, Levitsky LL, Herzog DB, Klibanski A. The effects of anorexia nervosa on bone metabolism in female adolescents. J Clin Endocrinol Metab. 1999;84:4489-4496.
40. Soyka LA, Misra M, Frenchman A, et al. Abnormal bone mineral accrual in adolescent girls with anorexia nervosa. J Clin Endocrinol Metab. 2002;87:4177-4185.
41. Zipfel S, Beumont PJ, Russel J, HerzogW. Osteoporosis in eating disorders. Eur Eat Disord Rev. 2000;8:108-116.
42. Bolton JGF, Patel S. Osteoporosis in anorexia nervosa. J Psychosom Res. 2001; 50:177-178.
43. Lucas AR, Melton LJ, Crowson CS, O’fallonWM. Long-term fracture risk among women with anorexia nervosa : a population-based cohort study. Mayo Clin Proc. 1999;74:972-977.
44. Zipfel S, Seibel MJ, Lowe B, Beumont PJ, Kasperk C, Herzog W. Osteoporosis in eating disorders : a follow-up study of patients with anorexia and bulimia nervosa. J Clin Endocrinol Metab. 2001;86:5227-5233.
45. Seeman E, Kerlsson MK, Duan Y. On exposure to anorexia nervosa, the temporal variation in axial and appendicular skeletal development predisposes to site-specific deficits in bone size and density: a cross-sectional study. J Bone Miner Res. 2000;15:2259-2265.
46. Legroux-Gérot I, Vigneau J, D’Herbomez M, et al. Evaluation of bone loss and mechanisms in anorexia nervosa. Calcif Tissue Int. 2007;81:174-182.
47. Hotta M, Shibasaki T, Sato K, Demura H. The importance of body weight history in the occurrence and recovery of osteoporosis in patients with anorexia nervosa : evaluation by dual X-ray absorptiometry and bone metabolic markers. Eur J Endocrinol. 1998;139:276-283.
48. Stoving RK, Veldhuis JD, Flyvbjerg A, et al. Jointly amplified basal and pulsatile Growth Hormone (GH) secretion and increased process irregularity in women with anorexia nervosa : indirect evidence for disruption of feedback regulation within the GH-Insulin-Like Growth Factor I axis. J Clin Endocrinol Metab. 1999: 84:2056-2063.
49. Scacchi M, Pincelli AI, Caumo A, et al. Spontaneous nocturnal growth hormone secretion in anorexia nervosa. J Clin Endocrinol Metab. 1997;82:3225-3229.
50. Counts DR, Gwirtsman H, Carlsson LMS, Lesem M, Cutler GB Jr. The effect of anorexia nervosa and refeeding on growth hormone-binding protein, the insulinlike growth factor (IGFs), and the IGF-binding proteins. J Clin Endocrinol Metab. 1992;75:762-767.
51. Hotta M, Fukuda I, Sato K, Hizuka N, Shibasaki T, Takano K. The relationship between bone tunover and body weight, serum insulin-like growth factor (IGF) I, and serum IGF-binding protein levels in patients with anorexia nervosa. J Clin Endocrinol Metab. 2000;85:200-206.
52. Legroux-Gérot I, Vignau J, Biver E, et al. Anorexia nervosa, osteoporosis and circulating leptin: the missing link. Osteoporos Int. 2010 Jan 6. [Epub ahead of print] DOI 10.1007/s00198-009-1120-x.
53. Audi L, Vargas DM, Gussinyé M, Yeste D, Marti G, Carrascosa A. Clinical and biochemical determinants of bone metabolism and bone mass in adolescent female patients with anorexia nervosa. Pediatric Res. 2002;51:497-504.
54. Hartman D, Crisp A, Rooney B, Rackow C, Atkinson R, Patel S. Bone density of women who have recovered from anorexia nervosa. Int J Eat Disord. 2000; 28:107-112.
55. Wentz E, Mellström D, Gillberg C, Sundh V, Gillberg I C, Rastam M. Bone density 11 years after anorexia nervosa onset in a controlled study of 39 cases. J Eat Disord. 2003;34:314-318.
56. Legroux-Gerot I, Vignau J, Collier F, Cortet B. Factors influencing changes in bone mineral density in patients with anorexia nervosa-related osteoporosis: the effect of hormone replacement therapy. Calcif Tissue Int. 2008; 83: 315-323.

Keywords: psychiatric disorder; depression; anorexia nervosa; hyperadrenocorticism; estrogen; osteoporosis

Download this issue

Maria Luisa BRANDI, PhD
Gerard KARSENTY, MD, PhD
Metabolic Bone Diseases
Department of Internal Medicine
University of Medicine
Medical School, Florence
ITALY

The association between diabetes and bone health has long been a matter of debate. Both type 1 diabetes and type 2 diabetes have been linked to increased risk of fractures, with bonemineral density being decreased in type 1 diabetes and increased in type 2 diabetes. Insulin has an anabolic effect on bone, and the qualitatively different effects of type 1 and type 2 diabetes on bone mass are consistent with the opposite insulin-secretory states (hypoinsulinemia vs hyperinsulinemia). The existence of an elevated fracture risk in type 2 diabetes, despite the underlying hyperinsulinemia, has led to speculation about differences in bone quality between type 1 diabetes and type 2 diabetes. This could be explained by the fact that increased blood glucose levels are associated with increased urinary calcium loss, resulting in a negative calcium balance. There is also speculation about the role of the resistance to parathyroid hormone observed in diabetes, and its effect on calcium and bone turnover. Also, collagen glycosylation may alter bone biomechanical competence. Falls associated with diabetes-related comorbidities are another possible cause of low-trauma fractures. Adequate glycemic control and prevention of diabetic complications are the mainstay of therapy to lower fracture risk, with the caveat that thiazolidinediones increase fracture risk in postmenopausal women with type 2 diabetes. In conclusion, bone health is an important consideration in diabetes, and caution should be exercised in prescribing thiazolidinediones to postmenopausal women with low bonemass and patients with prior fragility fracture. This article reviews the current state of knowledge on the association between diabetes and bone health.

More than 180 million people worldwide suffer from type 2 diabetes, a disease that more than doubles the risk of death, mainly from cardiovascular disease.¹ Interestingly, the medical literature provides evidence of a convergence between diabetes, a metabolic disease, and potential mechanisms accounting for osteoporosis. Skeletal involvement in diabetes was first suggested more than 80 years ago, prompted by radiological findings of retarded bone development and bone atrophy in children with type 1 diabetes.² In 2007, a systematicmeta-analysis in women with type 2 diabetes reported that, although there was no significant increase in vertebral or distal forearm fractures, hip fracture risk was elevated 1.7- fold.³ Furthermore, it is now recognized that diabetes and hip fractures share common risk factors. Nevertheless, despite a large body of accumulated data on the skeletal effects of diabetes, many questions remain unresolved, with biochemical and imaging studies producing conflicting findings. This is likely to be due in large part to the complex pathophysiology of diabetes, the diversity of skeletal sites examined, the multitude of techniques used for measuring bone mass, and variations in the duration, severity, and treatment of diabetes in the different studies. This paper reviews our current understanding of the pathogenetic bases of bone disease in diabetes.

Figure 1. Current model of the leptindependent regulation of bone mass.

Pathophysiology

_ The biological relevance of bone remodeling
There is a constant turnover of bone through bone remodeling, via a biphasic process occurring throughout the skeleton over a period of approximately 3 months.⁴ It includes destruction (resorption) of preexisting bone, a function exerted by a specialized bone-specific cell, the osteoclast, followed by de novo bone formation, a function exerted by another bonespecific cell, the osteoblast. Normally, resorption and formation of bone occur not only sequentially, but in a balanced manner in order to maintain bone mass nearly constant during most of adulthood. This qualifies bone remodeling as a true homeostatic function controlled by cytokines acting locally and hormones acting systemically.

Maintenance of constant bone mass is the aspect of bone remodeling we are most familiar with, because osteoporosis, the most frequent bone disorder, is a bone-remodeling disease.⁵ Osteoporosis results from an increase in bone resorption exceeding bone formation.⁶ Bone remodeling can be studied by means of biological markers in serum and urine, or bone mineral density (BMD). BMD is a strong predictor of fracture risk, but bone mineral quantity is only one component of bone strength, and various disorders, including diabetes, can be associated with poor bone quality.

The relatively recent observation of a convergence between bone and energy homeostasis suggests that energy metabolism and bone mass are regulated by the same hormones, such as leptin (Figure 1),⁷ adiponectin,⁸ neuropeptide Y,⁹ and substance P.¹⁰ A remarkable feature of most types of hormonal regulation is that they are controlled by feedback loops, such that the cells targeted by a hormone send signals influencing the hormone-producing cells. When applied to skeletal biology, the concept of feedback regulation suggests that bone cells exert an endocrine function.

This was recently demonstrated by the finding that the skeleton exerts an endocrine regulation of glucose homeostasis through the “secretion” of osteocalcin, one of the very few osteoblast- specific proteins, which improves glucose homeostasis by favoring β-cell proliferation and insulin secretion (Figure 2, page 366).¹¹ Teleologically, the proliferation function of osteocalcin may have arisen during evolution to maintain the size of the pancreatic islets constant in periods of food deprivation.

_ Bone phenotypes in type 1 and type 2 diabetes
Type 1 diabetes, also called insulin-dependent diabetes mellitus, is characterized by little or no insulin production and hyperglycemia. Improved glucose monitoring, insulin delivery methods, and pharmacologic treatments are increasing patient lifespan. However, as a result, there is a parallel increase in the risk of complications due to extended exposure to diabetes. Attention has been recently focused on diabetic bone pathology, as type 1 diabetes was found to be clearly associated with bone loss and suppressed bone formation. As reported by McCabe comparing type 1 diabetic patients and healthy age-matched subjects, it is estimated that more than 50% of type 1 diabetic patients have bone loss and almost 20% of patients aged 20 to 56 meet the criteria for osteoporosis.¹² Quite logically in this connection, type 1 diabetes has been shown to be a risk factor for delayed fracture healing.¹³ Bone loss can begin as early as at onset of diabetes in children, but there are reports of children with type 1 diabetes who do not exhibit bone loss.^14,15 Bone loss occurs predominantly in the appendicular skeleton. A concern is that existing bone loss in type 1 diabetic patients could compound the fracture risk associated with conditions such as menopause and aging.

Figure 2. Regulation of energy metabolism by the skeleton.

The mechanisms contributing to type 1 diabetic bone loss are unknown, but several theories have been put forward. Analysis of type 1 diabetic bone remodeling serum markers suggests that bone turnover is unaltered or decreased, while bone formation is decreased, as indicated by reduced serum levels of osteocalcin and histomorphometric studies.^16,17 The potential contributors to type 1 diabetic bone phenotypes are listed in Table I.

Type 2 diabetes, also called non–insulin-dependent diabetes mellitus, develops when cells become resistant to insulin signaling, and accounts for more than 90% of diabetes cases. Diet, obesity, and reduced physical activity are several of the factors that are thought to contribute to the development of type 2 diabetes. Available data concerning an association between reduced BMD and type 2 diabetes are equivocal. Type 2 diabetes mellitus in the literature has been reported to be associated with increased,¹⁸ unchanged,¹⁹ or decreased²⁰ BMD. However, most large-scale epidemiological studies indicate normal or above-normal BMD.²¹ Possible contributing factors to the higher BMD of type 2 diabetes mellitus are listed in Table II.

Table I. potential contributors of the bone phenotypes in type 1
diabetes mellitus.

Table II. Potential contributors of high BMD in type 2 diabetes
mellitus.

_ Risk of fracture in type 1 and type 2 diabetes mellitus
The most convincing evidence that osteoporosis is a complication of diabetesmellitus comes fromepidemiological studies that have shown an increased risk of fragility fractures. Diabetes and hip fracture share common risk factors (eg, physical inactivity, advanced age); in contrast, obesity, a risk factor for diabetes, is associated with a lower risk of fractures, and any apparent modification in fracture risk by diabetes is likely to reflect a confounding effect of these and other extraneous factors.

Investigations into fracture risk in type 1 diabetes have yielded inconsistent results, with increased incidence of hip fracture being reported in some studies, but not in others.22-28 A recent meta-analysis in patients with type 1 diabetes mellitus reported that this population is at six- to sevenfold higher risk of hip fracture than nondiabetic individuals.3 Cross-sectional and prospective studies have shown type 1 diabetes to confer an increased risk of fragility fracture at other sites, in both men and women.26,29,30

Even though a recentmeta-analysis involving a total of 836000 participants concluded that hip fracture risk was elevated 1.7- fold in women with type 2 diabetes mellitus,3 some studies have reported either no increase in hip fractures27,31 or risks restricted to patients with a higher duration of disease.22,32,33 The reports of increased fracture risk are somewhat unexpected because 2-dimensional areal BMD is normal or elevated in persons with type 2 diabetes,21,34 and this implies that diabetic individuals are at decreased risk of fracture. Moreover, the meta-analysis found no significant increase in vertebral or distal forearm fractures in these patients.³ At present, there is no clear explanation for this apparent contradiction. An increased risk of falling in diabetic patients35 could account for the elevated hip fracture risk in the face of normal or elevated BMD. A possible explanation for increased bone fragility in diabetes mellitus is the accumulation of advanced glycation end products within bone collagen, leading to increased stiffness of the collagen network.36 Increased blood glucose levels could also have direct deleterious effects on bone cells,37 with consequences on bone biomechanical competence.38 Moreover, adipose tissue (usually increased in type 2 diabetes mellitus) produces cytokines, namely, adipokines, such as leptin, resistin, and adiponectin, which may negatively modulate BMD.39 Figure 3 depicts the potential mechanisms contributing to fracture susceptibility in diabetes mellitus.

Effects of antidiabetic agents on bone

Oral antidiabetic drugs are commonly used to improve glycemic control, but there are concerns that some may increase the risk of cardiovascular events.⁴⁰ Moreover, several epidemiological studies have investigated the effects of antihyperglycemic treatment on fracture risk in diabetes. In the largest of these, in which all individuals diagnosed with fracture in Denmark in 2000 were matched with controls, it was reported that metformin and sulfonylurea treatments were associated with reduced incidences of fracture, while insulin was associated with a nonsignificant trend toward reduced risk of hip, forearm, and spine fractures.²⁶

Conversely, recent evidence suggests that the thiazolidinediones, first introduced for the treatment of type 2 diabetes mellitus in 1999, may affect the skeleton, with an increase in fracture risk in women randomized to rosiglitazone versus those randomized to metformin or glyburide monotherapy.⁴¹ In this study, fracture events were not increased in men and did not increase with time.⁴¹ These results were also confirmed in preliminary data from another study.⁴²

Interestingly, pioglitazone, the other currently available thiazolidinedione, may have similar skeletal effects, with the majority of fractures occurring at nonvertebral sites, including the lower limb and distal upper limb.⁴³

As these findings support the hypothesis of a class effect of thiazolidinediones in increasing fracture risk in women with type 2 diabetes mellitus, letters to health care providers have been issued by the manufacturers.^44,45 However, doubts still exist about the clinical relevance of this phenomenon, and more studies are needed to address a number of still pending questions,^21,46 such as the precise mechanism of action of these agents (Figure 4).

Figure 3. Potential mechanisms contributing to low bone mass and increased fracture susceptibility in diabetes mellitus.

Physicians should carefully check for the existence of risk factors for osteoporosis and fractures in their patients before putting them on thiazolidinedione treatment, and an adequate clinical follow-up of treated patients is strongly recommended.

Figure 4. Potential mechanisms for bone fracture
with thiazolidinediones (TZDs).

Future prospects

The prevalence of diabetes mellitus is increasing rapidly in the population, with the implication that adverse outcomes of the condition are likely to grow in importance as well. Considerable concern has been expressed about fracture risk in these patients. Although fractures may now be prevented thanks to the availability of effective treatments, no clear rationale exists for treating patients with type 2 diabetes with antifracture agents able to increase BMD, and our knowledge base is not strong enough for a more effectively tailored prophylaxis to be designed for this group. Additional research is needed to better define the determinants of bone strength in diabetic individuals, including the abnormal properties of bone that might respond to treatment of diabetes itself. Conversely, the differences between type 1-diabetic- and age-associated bone loss stress the importance of selecting condition-specific individualized treatments for osteoporosis. Because in type 1 diabetes the bone defect results predominantly from a decrease in bone formation, anabolic therapies appear likely to be the most effective treatment.

Future studies should contribute to a more thorough understanding of the mechanisms of diabetic bone loss, enabling the development of newer and more effective drugs. Optimizing therapies that prevent bone loss or restore bone density will allow diabetic patients to live longer, with strong healthy bones. _

1. World Health Organization. Diabetes. 2008. www.who.int/mediacentre/fact sheets/fs312/ en/index.htlm.
2. Morrison LB, Bogan IK. Bone development in diabetic children: a roentgen study. Am J Med Sci. 1927;174:313-318.
3. Janghorbani M, Van Dam RM, Willett WC, Hu FB. Systematic review of type 1 and type 2 diabetes mellitus and risk of fracture. Am J Epidemiol. 2007;166: 495-505.
4. Rodan GA, Martin TJ. Therapeutic approaches to bone diseases. Science. 2000;289:1508-1514.
5. Cooper C, Melton LJI. Magnitude and impact of osteoporosis and fractures. In: Marcus R, Feldman D, Kelsey J, eds. Osteoporosis. san Diego, Calif: Academic Press; 2001:419-434.
6. Raisz LG. Clinical practice. Screening for osteoporosis. N Engl J Med. 2005; 353:164-171.
7. Karsenty G. Convergence between bone and energy homeostases: leptin regulation of bone mass. Cell Metab. 2006;4:341-348.
8. Kanazawa I, Yamaguchi T, Yano S, Yamauchi M, Yamamoto M, Sugimoto T. Adiponectin and AMP kinase activator stimulate proliferation, differentiation and mineralization of osteoblastic MC3T3-E1 cells. BMC Cell Biol. 2007;8:51-62.
9. Allison SJ, Baldock PA, Enriquez RF, et al. Critical interplay between neuropeptide Y and sex steroid pathways in bone and adipose tissue homeostasis. J Bone Miner Res. 2009;24:294-304.
10. Wang L, Zhao R, Shi X, et al. Substance P stimulates bone marrow stromal cell osteogenic activity, osteoclast differentiation, and resorption activity in vitro. Bone. 2009;45:309-320.
11. Lieben L, Callewaert F, Bouillon R. Bone and metabolism: a complex crosstalk. Horm Res. 2009;71(suppl 1):134-138.
12. McCabe LR. Understanding the pathology and mechanisms of type 1 diabetic bone loss. J Cell Biochem. 2007;102:1343-1357.
13. White CB, Turner NS, Lee GC, Haidukewych GJ. Open ankle fractures in patients with diabetes mellitus. Clin Orthop. 2003;414:37-44.
14. Bechtold S, Dirlenbach I, Raile K, Noelle V, Bonfig W, Schwarz HP. Early manifestation of type 1 diabetes in children is a risk factor for changed bone geometry. Data using peripheral quantitative computed tomography. Pediatrics. 2006; 118:e627-e634.
15. Valerio G, Del Puente A, Esposito-Del Puente A, Buono P, Mozzillo E, Francese A. The lumbar bone mineral density is affected by long-term poor metabolic control in adolescents with type 1 diabetes mellitus. Horm Res. 2002;58: 266-272.
16. Lu H, Kraut D, Gerstenfeld LC, Graves DT. Diabetes interferes with the bone formation by affecting the expression of transcription factors that regulate osteoblast differentiation. Endocrinology. 2003;144:346-352.
17. Thrailkill KM, Liu L, Wahl EC, Bunn RC, et al. Bone formation is impaired in a model of type 1 diabetes. Diabetes. 2005;54:2875-2881.
18. van Daele PL, Stork RP, Burger H, etal. Bone density in non-insulin-dependent diabetes mellitus. The Rotterdam Study. Ann Intern Med. 1995;122:409-414.
19. Wakasugi M, Wakao R, Tawata M, Gan N, Koizumi K, Onaya T. Bone mineral density measured by dual energy x-ray absorptiometry in patients with non-insulin- dependent diabetes mellitus. Bone. 1993;14:29-33.
20. Ishida H, Seino Y, Matsukura S, et al. Diabetic osteopenia and circulating levels of vitamin D metabolism in Type 2 (noninsulin-dependent) diabetes. Metabolism. 1985;34:797-801.
21. Adami S. Bone health in diabetes: considerations for clinical management. Curr Med Res Opin. 2009;5:1057-1072.
22. Forsen L, Meyer HE, Midthjell K, Edna TH. Diabetes mellitus and the incidence of hip fracture: results from the Nord-Trondelag Health Survey. Diabetologia. 1999;42:920-925.
23. Janghorbani M, Hu FB, Willett WC, Li TY, Manson JE, Logroscino G, Rexrode KM. Prospective study of type 1 and 2 diabetes and risk of stroke subtypes: The Nurse’s Health Study. Diabetes Care. 2007;30:1730.1735.
24. Miao J, Brismar K, Nyren O, Ugarph-Morawski A, Ye W. Elevated hip fracture risk in type 1 diabetic patients: a population-based cohort study in Sweden. Diabetes Care. 2005;28:2850-2855.
25. Nicodemus KK, Folsom AR. Type 1 and type 2 diabetes and incident hip fractures in postmenopausal women. Diabetes Care. 2001;24:1192-1197.
26. Vestergaard P, Rejnmark L. Mosekilde L. Relative fracture risk in patients with diabetes mellitus, and the impact of insulin and oral antidiabetic medication on relative fracture risk. Diabetologia. 2005;48:1292-1299.
27. Heath III H, Melton III LJ, Chu Cp. Diabetes mellitus and risk of skeletal fracture. 1980;303:567-570.
28. Melchior TM, Sorensen H, Torp-Pedersen C. Hip and distal arm fracture rates in peri- and post-menopausal insulin-treated diabetic females. J Intern Med. 1994;236:203-208.
29. Ahmed LA; Joakimsen RM, Berntsen GK, Fønnebø V, Joakimsen RM. Diabetes mellitus and the risk of non-vertebral fractures: the Tromsø study. Osteoporos Int. 2006;17:495-500.
30. Kelòsey JL, Browner WS, Seeley DG, Nevitt MC, Cummings SR. Risk factors for fractures of the distal forearm and proximal humerus. The Study of Osteoporotic Fractures Research Group. Am J Epidemiol. 1992;135:477-489.
31. Ivers RQ, Cumming RG, Mitchell P, Peduto AJ; Blue Mountains Eye Study Group. Diabetes and risk of fracture: the Blue Mountains Eye Study. Diabetes. 2001;24:1198-1203.
32. Leslie WD, Lix LM, Prior HJ, Derksen S, Metge C, O’Neil J. Biphasic fracture risk in diabetes: a population-based study. Bone. 2007;40:1595-1601.
33. de Liefde I, van der Klift M, de Laet CE, van Daele PL, Hofman A, Pols HA. Bone mineral density and fracture risk in type 2-diabetesmellitus: the Rotterdam Study. Osteoporos Int. 2005;16:1713-1720.
34. Hofbauer LC, Brueck CC, Singh SK, Dobnig H. Osteoporosis in patients with diabetes mellitus. J Bone Miner Res. 2007;22:1317-1328.
35. Schwartz AV, Sellmeyer DE. Women, type 2 diabetes and fracture risk. Curr Diab Rep. 2004;4:364-369.
36. Paul RG, Bailey AJ. Glycation of collagen: the basis of its central role in the late complications of ageing and diabetes. Int J Biochem Cell Biol. 1996;28:1297- 1310.
37. Gopalakrishnan V, VigneshbRC, Arunakaran J, Aruldhas MM, Srinivasan N. Effects of glucose and its modulation by insulin and estradiol on BMSC differentiation into osteoblastic lineages. Biochem Cell Biol. 2006;84:93-101.
38. Saito M, Fujii K, Soshi S, Tanaka T. Reductions in degree of mineralization and enzymatic collagen cross-links and increases in glycation-induced pentosidine in the femoral neck cortex in cases of femoral neck frature. Osteoporos Int. 2006;17:986-995.
39. Kanazawa I, Yamaguchi T, Yamamoto M, Yamauchi M, Yano S, Sugimoto T. Relationships between serum adiponectin levels versus bone mineral density, bone metabolic markers, and vertebral fractures in type 2 diabetes mellitus. Eur J Endocrinol. 2009;160:265-273.
40. Tzoulaki I, Molokhia M, Curcin V, et al. Risk of cardiovascular disease and all cause mortality among patients with type 2 diabetes prescribed oral antidiabetes drugs: retrospective cohort study using UK general practice research database. BMJ. 2009;339:b4731-b4739.
41. Kahn SE, Haffner S, Heise MA; ADOPT Study Group. Glycemic durability of rosiglitazone, metformin, or glyburide monotherapy. N Engl J Med. 2006;355:2427- 2443.
42. Home PD, Jones NP, Pocock SJ; RECORD Study Group. Rosiglitazone RECORD study: glucose control outcomes at 18 months. Diabet Med. 2007;24:626-634.
43. Dormandy JA, Charbonnel B, Eckland D; PROactive investigators. Secondary prevention of macrovascular events in patients with type 2 diabetes in the PROactive Study (PROspective PioglitAzone Clinical Trial In macroVascular Events): a randomised controlled trial. Lancet. 2005;366:1279-1289.
44. Takeda. Dear Healthcare Provider Letter. Observation of an increased incidence of fractures in female patients who received long-term treatments with Actos (pioglitazone HCl) tablets for type 2 diabetes mellitus. Available from http: //www.fda.gov/medwatch/safety/2007/Actosma0807.pdf.
45. GlaxoSmithKline. Dear Healthcare Provider Letter, re: clinical trial observation of an increased risk of fractures in female patients who received long-termtreatment with Avandia (rosiglitazone maleate) tablets for type 2 diabetes mellitus. Available from http://www.fda.gov/MedWaatch/safety/2007/Avandia_GSK_Ltr.pdf.
46. Falchetti A, Masi L, Brandi ML. Thiazolidinediones and bone. Clin Cases Miner Bone Metab. 2007;4:103-107.