Although obesity is a major risk factor for many diseases and has become a severe burden on healthcare costs in Western societies, it may have at least one beneficial effect, that of preventing osteoporosis, itself another major health problem. Indeed, it is well established that body weight is one of the strongest predictors of bone mass in both sexes.(1–5) However, the mechanisms that explain this relationship remain unclear and there is still ongoing controversy as to whether body fat mass (BFM) or lean mass is a better predictor of bone mineral density (BMD).(5–9)
The relationship between fat mass and bone mass also appears to be gender specific. Reid et al.(10) found no association in men, but they and others(9–11) did find a significant relationship between BFM and BMD in women. Early after menopause, a higher body mass index (BMI) is related to a slower bone loss rate induced by estrogen deficiency.(12,13) In a group of elderly female hip fracture patients, BFM has been shown to be a better explanatory factor for hip fracture than lean mass or body weight.(14) Likewise, in the prospective cohort study EPIDOS, a decrease of 1 SD in BFM was associated with a 30% increase in hip fracture risk, whereas lean mass was not significantly associated with fracture risk.(15)
Obesity-induced mechanical loading certainly is a contributing factor in the positive relationship between BFM and BMD. But this protective effect has been observed even at non-weight-bearing bone sites.(11) Thus, factors other than skeletal loading are likely important in mediating the relationship between BFM and BMD. Indeed, several lines of evidence have suggested that the effects of BFM on BMD may be mediated by hormonal factors, including estrogens and insulin. Adipose tissue produces estrogens by aromatization of androgens,(16) which represents the principal source of estrogens after menopause.(17) This peripheral estrogen production could affect bone turnover and BMD, especially after menopause.(18,19) However, BFM and BMD remain positively and strongly correlated in women, even after adjusting for serum estrogen levels.(20,21) It remains controversial as to whether BFM is better associated with BMD before(22,23) or after menopause.(6,24)
Serum insulin levels also have been shown to be related to BMD,(25) and hyperinsulinemia has been proposed as a potential explanation for the association between obesity and BMD in women, perhaps because of direct mitogenic effects of insulin on osteoblasts.(26,27) However, patients with non-insulin-dependent diabetes mellitus have been reported to have increased, decreased, or unchanged BMD (reviewed by Thomas et al.(28)). Hence, other humoral or metabolic factors are almost certainly involved in the mechanisms of fat-induced bone protection.
Recently, leptin has emerged as a potential candidate responsible for protective effects of fat on bone tissue. Leptin, the product of the ob gene, is secreted mainly by white adipose tissue(29) and is correlated strongly with BFM.(30–32) This 14-kDa protein is a representative marker of the body's energetic status and raises considerable interest. In fact, in addition to its effects on the central nervous system,(33,34) recent studies provide new evidence of its role in many different metabolic pathways. Besides insulin action, to which leptin is related strongly, leptin may alter numerous endocrine functions, for example, hormonal secretion from the anterior pituitary,(35) ovaries,(36) testes,(37) thyroid, and adrenal glands.(38) Experimental data suggest that the central loop of leptin action regulating appetite and these peripheral effects on endocrine glands are not related.(39) In addition to these effects, leptin may act as a growth factor, with the ability of directly enhancing development of hemopoietic precursor cells,(40–44) myoblast-like cells,(45,46) or lung cells.(47) Furthermore, it can modulate jejunum epithelial cell activity,(48) and even more importantly, it exerts angiogenic effects on vascular endothelial cells.(49) Conversely, leptin suppresses specific biochemical reactions that contribute to adipocyte differentiation and lipid accumulation.(50,51) Leptin acts through highly specific receptors,(52) which have been described on hemopoietic precursor cells,(40) in placenta and several fetal tissues,(53) white and brown adipocytes,(54) in lung tissue,(49) and the jejunum.(47)
Are bone cells potential targets for leptin?
Bone marrow stromal cells have the ability of differentiating into osteoblasts as well as adipocytes, chondroblasts, and myoblasts.(55) Thomas et al.,(56) using a human stromal cell line, and others(57–60) showed that these cells might respond to leptin action because they expressed both short and long forms of leptin receptors. This expression might be dependent on the differentiation stage, because Ducy et al.(61) found no expression of these receptors in mature murine osteoblasts. Both leptin and its receptors also were found in murine fetal cartilage/bone template(62) as well as in the growth plate, especially in chondrocytes close to the vascular invasion of this tissue.(63) These data are of great interest regarding the recently shown angiogenic properties of leptin.(49)
Do we have evidence that leptin affects bone metabolism?
First, several in vitro studies suggest that stromal cells respond to leptin stimulation. Takahashi et al.(64) showed that leptin induced MAPK-dependent cell proliferation in the mouse embryonic cell line C3H10T1/2, a multipotential stromal cell line. These data recently were confirmed on a different osteoblastic cell line.(58) These effects could be differentiation stage-dependent, because Thomas et al.(56) observed no effect on proliferation in the human marrow stromal cell line hMS2-12, progenitor cells that probably are more mature than the C3H10T1/2 cells. Conversely, leptin administration increased osteoblastic differentiation of these hMS2-12 cells, leading to an increase in the mineralization of the extracellular matrix.(56) Similar results were observed in other cell models, with an increased number of mineralized nodules in rat cell culture treated with leptin(65) and in long-term human osteoblast cultures.(60)
Leptin could be one of several factors that modulate the reciprocal differentiation of stromal cells between osteoblastic and adipocytic pathways, because it has the ability to inhibit adipogenesis in a negative feedback loop.(50,56) The cellular mechanisms, which mediate these effects, remain unclear. The leptin-induced activation of the MAPK cascade may be critical because these biochemical events stimulate both osteoblastic differentiation(66) and phosphorylation of peroxisome proliferator-activated receptor (PPARγ), which has been shown to inhibit adipogenesis.(67)
In addition to direct positive effects on the osteoblastic differentiation of stromal cells, leptin also may modulate bone remodeling. Indeed, it has been shown recently in human stromal cells that leptin inhibited the expression of RANKL, the major downstream cytokine controlling osteoclastogenesis.(68) Furthermore, it stimulated the expression of osteoprotegerin (OPG), its decoy receptor, by stromal cells(68) or mononuclear cells.(69) Leptin also was able to induce the expression and secretion of the interleukin (IL)-1 receptor antagonist by human monocytes,(70) which could blunt an IL-1-related increase in bone turnover, another key event in estrogen deficiency-induced bone loss.
Although leptin may exert direct effects on preosteoblastic stromal cells, one has to keep in mind that its effects on different endocrine systems could indirectly alter bone metabolism as well. In fact, leptin has been shown to modulate the growth hormone /insulin-like growth factor pathway,(71) the pituitary-thyroid axis,(72) as well as the pituitary-adrenal,(38) and gonadal axis.(36) The relative contribution of these effects to the resultant action of leptin on bone remains largely undetermined.
Second, animal studies strongly suggest that leptin alters bone metabolism, but conclusions on the nature of these effects remain highly controversial so far. Logically, several studies have explored the putative effects of leptin on bone metabolism using animals either deficient in leptin, that is, the ob/ob mice, or with mutation in the leptin-receptor, that is, the fa/fa Zucker rats and the db/db mice. Foldes et al.(73) showed that the obese fa/fa rats had a lower bone mass with shorter and lighter femurs, as compared with their normal littermates (NLMs). The decrease in bone mass was associated with an increase in bone resorption activity. Trabecular microarchitecture was also altered with an increase in trabecular number but a decrease in trabecular thickness and no change in trabecular bone volume. A recent report on fa/fa rats presented different data with a significant increase in trabecular bone volume resulting in a significant increase in biomechanical strength of long bones.(74) The reasons for these discrepancies remain unclear but the association of thicker cortices and shorter bones indeed may explain an improved mechanical strength.
Two studies evaluating the effects of peripheral leptin administration in ob/ob mice observed a stimulatory effect on bone growth(59) with a dramatic increase in cortical bone formation(75) in treated animals as compared with their controls. Indeed, intraperitoneal leptin administration to 4-week-old ob/ob male mice for 3 weeks reversed the defect in bone growth and osteopenia, as measured by DXA and peripheral quantitative computerized tomography (pQCT), despite a 40% decrease in food intake and a 14% decrease in body weight.(59) No effect was observed in the lean littermates.
In contrast, Ducy et al. reported that 3-month and 6-month old ob/ob mice and db/db mice had higher bone mass associated with a higher mineral apposition rate as compared with their wild-type littermates, despite hypogonadism and hypercortisolism secondary to leptin deficiency.(61) Furthermore, they showed that intracerebroventricular (icv) administration of leptin to 4-month old ob/ob female mice for 4 weeks normalized their bone mass. Therefore, these authors suggested that the increase in BMD observed in obese humans might be secondary to leptin resistance, a phenomenon believed to be critical in human obesity development.(76) This leptin resistance has been related first to a defect of leptin transport across the blood-brain barrier(77) into the central nervous system, as suggested by a more efficient leptin uptake (measured as the cerebro-spinal fluid [CSF]/plasma leptin ratio) up to four to six times greater in lean compared with obese individuals.(78–80) Second, a true hypothalamic resistance to leptin might occur with the onset of obesity, because there was no correlation between cerebrospinal fluid concentrations of leptin and its principal mediators neuropeptide Y (NPY) and melanocyte stimulating hormone (α-MSH) in obese women.(81) Furthermore, a progressive resistance to icv administration of leptin on food intake accompanied the development of high fat diet-induced obesity in C57Bl/6J mice.(82,83)
The reasons for discrepancies between different studies using the ob/ob mice model remain unclear. Age or sex might play a role but similar data were observed in male mice.(61) Actually, the most striking difference was the route of leptin administration. Ducy et al. reported no change in serum leptin levels after icv leptin infusion,(61) strongly suggesting a central regulation of bone metabolism through a cascade of neuroendocrine events triggered by leptin. Paradoxically, these authors observed the same effects in inhibiting bone formation with icv administration of the neuropeptide Y,(61) an orexigenic peptide negatively regulated by leptin at the hypothalamus level in the satiety control loop.(84) Recently, Baldock et al.(85) reported interesting but puzzling data confirming, first, the existence of a central nervous regulation of bone formation and, second, the negative effects of NPY on osteoblast activity. Indeed, they showed that mice deficient in Y2 receptors (one of the five NPY receptor types) had a twofold increase in trabecular bone volume. Furthermore, they showed that a selective deletion of hypothalamic Y2 receptors induced a similar increase in bone mass with a stimulated osteoblast activity and no change in osteoclast surfaces.
Next, recent studies, using normal rodent animals, indeed support the hypothesis that leptin effects on bone could depend on the site of administration and, therefore, the signaling pathway activated. Burguera et al.(68) showed that continuous subcutaneous administration of leptin to 6-month-old rats was able to prevent partially the rapid bone loss induced by estrogen deficiency, in accordance with leptin in vitro effects on OPG expression and osteoclastogenesis.(68,69) Using another model of rapid bone loss, the tail-suspended rat, characterized by a rapid increase in bone resorption and a sustained decrease in bone formation, Thomas et al.(86) recently showed that a similar administration of leptin prevented disuse-induced bone loss. In addition, systemic daily administration of leptin to sexually mature male mice significantly increased bone strength by >20%.(58)
Third, human cross-sectional studies provide data, which, however, are less conclusive. Although some authors(87–89) reported a positive association between serum leptin levels and BMD, others(90–92) failed to find such a relationship. In a large cohort of 692 subjects, Thomas et al.(87) observed a positive association between leptin and BMD in women but not in men, regardless of the bone site (weight-bearing or not) and menopausal status. Furthermore, adjusting for leptin significantly reduced the positive association between BFM and BMD in women. However, because leptin is almost exclusively produced by fat with a very strong association between serum leptin levels and fat mass,(30–32) the relationship between leptin and BMD disappeared after adjusting for fat mass in most(87,88,91) but not all(89) studies. Interestingly, Pasco et al.(93) recently reported a direct relationship between serum leptin level and bone mineral content (BMC) at the femur and lateral spine and with BMD at the lateral spine, even after adjusting for fat mass in a population of 214 healthy nonobese women aged 20–91 years old. These data, together with a previous study showing a positive association between serum leptin level and whole body bone area,(94) suggest that leptin might act on the periosteal envelope of cortical bone, thereby increasing bone size. Moreover, in a group of 139 postmenopausal women aged 48–78 years old, leptin levels were significantly lower in postmenopausal women with vertebral fractures than in those without fracture.(89) In this study, multiple regression analysis showed leptin but not fat mass, as a significant contributing factor to fracture presence.(89)
In addition, there is some evidence that leptin may act as a growth factor during fetal development. Leptin and its receptor are produced by human placenta(95) and umbilical cord leptin levels are associated positively with height(96) and weight(97,98) at birth in humans. This association was independent of insulin-like growth factor I cord levels.(98) Furthermore, leptin levels in umbilical cord blood were independently associated with intrauterine growth after adjusting for adiposity, placental weight, insulin, and maternal leptin levels.(99)
Finally, pathological evidence in humans also provides evidence that leptin could exert direct or indirect effects on bone metabolism. Osteoporosis is a very severe feature of anorexia nervosa (AN), which displays an endocrine profile including hypogonadism and hypercorticism and dramatically reduced serum levels of leptin.(100) The magnitude of bone loss is much higher in AN than in hypothalamic amenorrhea and is related to nutritional factors in addition to the magnitude and duration of estrogen deficiency; estrogen replacement is not sufficient to prevent fully such a bone loss.(101) Interestingly, Karlsson et al.(101) recently suggested in a cross-sectional study in women with AN that estrogen deficiency may reduce volumetric BMD, the true bone density, whereas reduced bone size observed in AN may be more related to nutritional disorders. Likewise, bulimic women with functional amenorrhea and significantly decreased leptin levels had a lower BMD despite normal body weight.(102)
Case-reports in other patient populations include the observation of Shirakura et al.,(103) who documented a significant increase in serum leptin levels in women with ossification of spinal ligaments, as compared with matched controls. Ozata et al.(104) reported, in a highly consanguineous extended pedigree with a missense leptin gene mutation, four homozygous patients exhibiting alterations in parathyroid hormone (PTH)-calcium function; one of them had a low BMD despite significant obesity. Last but not least, Farooqi et al.(105) evaluated the effects of a 1-year recombinant leptin therapy in a 9-year-old girl with congenital leptin deficiency and severe obesity. They observed a 15.6-kg decrease in fat mass and a 0.82-kg decrease in lean mass concurrent with a 0.15-kg increase in whole body BMD.
In summary, it remains difficult to draw a clear picture of leptin effects on bone metabolism, because data appear sometimes conflicting or contradictory. Beyond differences in models or experimental procedures, it is tempting, so far, to hypothesize that leptin exerts dual effects depending on the bone tissue or the signaling pathway. Thus, early in life, leptin could stimulate bone growth and bone size through angiogenic effects and osteogenic activity in immature cortical bone. Later, it may decrease bone remodeling in the mature skeleton when trabecular bone turnover is high. Central negative effects could counterbalance these peripheral and positive effects, the latter being predominant only when central leptin resistance occurs or else when serum leptin level rises above a certain threshold. Therefore, the sex-dependent specificity of the relationship between leptin and BMD in human studies could be at least in part due to serum leptin levels that are 2- to 3-fold higher in women than in men, independent of adiposity. While these hypotheses remain highly speculative and require further investigations, existing studies consistently support the role of leptin as a link between fat and bone.