Editorial: Leptin—Central or Peripheral to the Regulation of Bone Metabolism?

In this issue of Endocrinology, Kveiborg et al. (1) examine the possible role of leptin in mediating the osteosclerosis they had previously observed in transgenic mice overexpressing ΔFosB, a naturally occurring splice variant of Fos B (2). ΔFosB is a member of the AP-1 (activator protein-1) family of basic leucine zipper transcription factors that are involved in the early responses of cells to a number of transmembrane signaling agents, including PTH (3). Mice overexpressing ΔFosB have marked reductions in serum leptin levels (2), and, based on the evidence reviewed below, it was possible that the osteosclerosis in these animals was related to this abnormality, as opposed to an intrinsic functional alteration in osteoblastic cells. The authors demonstrate fairly convincingly that restoration of circulating leptin levels in the ΔFosB transgenic mice fails to reduce bone formation parameters, consistent with an intrinsic change in osteoblast function in these animals. These findings provide further evidence that manipulating ΔFosB levels in osteoblasts may be a fruitful approach to novel anabolic approaches to osteoporosis. They also highlight, however, a fairly spirited controversy regarding the possible role of leptin in regulating bone metabolism. This revolves around the issue of whether the major effects of leptin on bone are mediated via central (i.e. through the central nervous system) or peripheral actions. As with most scientific controversies (as long as both sides are arguing based on sound data), potential resolutions may provide further insights that could perhaps have been overlooked by the opposing camps. It had been recognized for a long time that obesity is protective against the development of bone loss and osteoporotic fractures. A number of factors, including mechanical loading, increased aromatization of androgens to estrogens by adipose tissue (4), and increased insulin levels (5), had been proposed as mediating this protective effect of obesity on the skeleton, with evidence to support each of these factors. However, the identification of the ob gene (6), and the discovery that the protein encoded by this gene, leptin, was a hormone produced by adipocytes that regulated body weight (7), raised the possibility that leptin may also be a possible mediator of this protective effect. Consistent with this, some (8, 9), but not all (10), observational studies found that circulating leptin levels were positively associated with bone mass at various sites, particularly in postmenopausal women. Moreover, leptin administration reduced ovariectomy-induced bone loss in rats (11) and led to an increase in femoral length, total body bone area, bone mineral content, and bone density in leptin-deficient (ob/ob) mice (12). Additionally, in human bone marrow stromal cells (which were bipotential, with the capacity to differentiate either to osteoblasts or to adipocytes), leptin enhanced osteoblastic and inhibited adipocytic differentiation (13). These cells also expressed (by RT-PCR and by Western immunoblot analysis) both the short and long forms of the leptin receptor (OB-R). Interestingly, human bone marrow adipocytes, like extramedullary adipocytes, produce relatively large amounts of leptin (14), suggesting that leptin may, in fact, be an important paracrine signaling molecule between adipocytes and preosteoblastic cells in the bone marrow microenvironment. More recently, leptin has also been shown to be produced by primary cultures of human osteoblasts, to promote mineralization of these cells (15, 16), and to act as a growth factor on the chondrocytes of skeletal growth centers (17). Finally, leptin appears to inhibit osteoclast generation from both peripheral blood mononuclear cells as well as spleen cells, related (at least in part) to an increase in osteoprotegerin and decrease in receptor activator of nuclear factor-κB ligand production by supporting stromal cells (11, 16, 18). Collectively, these findings were easy to incorporate into conventional thinking about the relationship between fat mass and bone mass. More fat meant more leptin, and leptin appeared to 1) be positively associated with bone mass; 2) have beneficial effects on bone in animal studies; and 3) enhance osteoblast differentiation/function and reduce osteoclast development in vitro. This would have been a relatively straightforward story had it ended there. However, concurrently with these developments on the peripheral actions of leptin on bone, Ducy et al. (19) were pursing a novel and, in fact, somewhat surprising observation. Both the leptin-deficient (ob/ob) and -resistant (db/db) mouse have hypogonadism and hypercortisolism (19). Combined with the inability to either produce or respond to leptin, these abnormalities would be predicted to result in marked skeletal osteopenia; however, the reverse was observed. These animals had profound increases in bone mass associated with an increase in bone formation rate that appeared to be due not to an increase in osteoblast number but to increased osteoblastic activity. Moreover, intracerebroventricular infusion of leptin in both ob/ob and wild-type mice resulted in bone loss. Thus, these studies convincingly demonstrated that leptin inhibited bone formation through a central mechanism, likely mediated by the hypothalamus (19). These were truly remarkable findings, even if most of us were at a loss to reconcile them with the accumulating data indicating that leptin also appeared to have peripheral, positive effects on bone. At first glance, these negative central effects vs. the positive peripheral effects of leptin on bone make no sense; nature is generally quite parsimonious, and it seemed as though evolution had somehow misfired in coming up with these conflicting pathways that may, in the end, result in no net effect on bone. One approach is to argue that the data on one side or the other are wrong. This appears, however, to be unlikely. The findings of Ducy et al. (19) are elegant and convincing; on the other hand, the accumulating evidence from multiple laboratories that leptin does have direct actions on osteoblastic and chondrocytic cells (13, 15–18), attenuates ovariectomy-induced bone loss in rats (11), and is associated with bone mass in humans (8, 9) is also hard to dismiss. Can both be right, and if so, why would nature have gone through all this trouble? The solution, perhaps, is provided by an eloquent perspective delineated by Flier (20) that may help in understanding the true role of leptin in a number of systems. In it, he stresses that leptin should not be regarded as an antiobesity hormone but rather as a “signal of energy deficiency and as an integrator of neuroendocrine function.” According to this view, an important, or perhaps the major, function of leptin is as a “starvation” signal modulating a number of neuroendocrine responses to caloric restriction, as was likely to occur during times of famine. Thus, leptin levels fall rapidly, within hours of caloric restriction in rodents (21). These changes in leptin levels appear to be causally related to a number of starvation-induced changes in neuroendocrine function, including activation of the hypothalamic-pituitary-adrenal axis (resulting in hypercortisolism) and suppression of the thyroid and reproductive axes (Ref. 20 and Fig. 1). At least in rodents, starvation also causes a suppression of both GH and IGF-I production (22), whereas in humans, it causes GH to rise but suppresses IGF-I (23). Proposed scheme for central vs. peripheral effects of leptin on bone. Under conditions of caloric restriction (top panel), peripheral leptin levels fall rapidly, resulting in a pari passu reduction in CSF leptin levels. This, in turn, activates a number of adaptive neuroendocrine responses that are critical for the survival of the organism. Many of these, including an activation of the hypothalamic-pituitary-adrenal (HPA) axis, suppression of the gonadal axis, and a reduction in IGF-I levels, would result in bone loss (indicated by a minus sign). This appears to be offset, at least partially, by a centrally mediated increase in bone formation (BF) triggered also by the decrease in CSF leptin levels (solid line, plus sign). The low peripheral leptin levels under these conditions likely have little or no direct effect on bone (dashed line). In contrast, under conditions of high caloric intake, peripheral leptin levels rise. These high peripheral levels likely do result in direct beneficial effects on bone (solid line), enhancing the ability of the skeleton to support the increase in body weight. The high peripheral levels, however, translate into only a modest increase in leptin action in the central nervous system due to the development of central leptin resistance (related, in part, to decreased leptin transport into the CSF). This limits the adverse central effects of leptin on bone (dashed line), allowing the beneficial peripheral effects of leptin to predominate (solid line). In addition, restoration of CSF leptin levels reverses the neuroendocrine changes associated with caloric restriction, resulting in normal (NL) cortisol, gonadal steroid, and IGF-I levels, with consequent beneficial effects on bone. These “adaptive” responses to starvation, likely mediated by the fall in circulating [and, hence, cerebrospinal fluid (CSF)] leptin levels are clearly important to the survival of the organism: hypercortisolism results in increased glucose production and activation of a stress response that, in the past, was likely critical to the search for food; hypogonadism ensures that reproduction will not occur when food is scarce (rather than leaving it to the capricious judgment of the individual); suppression of the thyroid axis likely contributes to a fall in metabolic rate; and suppression of IGF-I limits increases in body/muscle size when calories are scarce. Collectively, these changes, many of which are triggered by the fall in leptin, help the organism survive though periods of scarcity, which is the natural state for most wild animals and, for that matter, for humans before our current civilization. However, they are potentially devastating to the skeleton. Thus, hypercortisolemia, hypogonadism, and decreased IGF-I production all would be predicted to lead to marked bone loss. It appears that nature may have attempted to counteract this through a central mechanism that increases bone formation as a consequence of falling leptin levels (Fig. 1). In this context, the fall in leptin and the central control of bone formation may be viewed as perhaps the last defense of the skeleton during periods of caloric restriction and the concomitant adaptive neuroendocrine responses that are necessary for survival but are deleterious to bone mass. Even this defense may break down, however, under conditions of pathological starvation as occurs, for example, in anorexia nervosa, which is often associated with marked osteopenia (24). What, then, occurs when the organism survives starvation and food again is plentiful? Based on the “thrifty genotype” hypothesis, this should now be a time to store energy for when times are not so good again. It has been suggested that under conditions of increasing energy intake and rising fat stores, it would be to the organism’s disadvantage to decrease appetite and forfeit this opportunity to store up for hard times (20). Under this scenario, as leptin levels rise, there develops a central resistance to leptin that appears, at least in part, to be due to markedly reduced transport of leptin from the blood into the CSF (25). This would, therefore, tend to limit the reduction in appetite and likely also blunts the central inhibition of bone formation with rising leptin levels. Concurrently, however, the positive peripheral effects of leptin on bone would then come into play (Fig. 1). This makes perfect sense from the perspective of the organism, because as fat mass rises, there is a need to support this with an increase in skeletal mass: for this, the central inhibitory effect of leptin on bone formation is blunted and the peripheral beneficial effects of leptin on bone formation and resorption become predominant. This formulation also predicts that the skeletal response to exogenous, systemic leptin administration may be quite variable, depending on the balance between the central inhibitory and peripheral stimulatory effects of leptin in a given individual. Although the scheme summarized in Fig. 1 is plausible and attractive in terms of reconciling the seemingly disparate observations on the central and peripheral actions of leptin on bone, it is a hypothesis that requires further testing. Clearly, major gaps remain in our understanding of how changes in energy intake, body weight, and fat mass regulate bone metabolism. For example, it is at present unclear just how the hypothalamus alters bone formation. In addition, are there species differences in this model? That this may be the case is suggested by the observations that peripheral leptin administration decreases food intake in rodents (26), but fails to do so (at least acutely) in rhesus monkeys (27), perhaps due to a more drastic limitation in transport of leptin into the CSF. Finally, it appears that neuropeptide Y and hypothalamic Y2 receptors, which are involved in appetite control, also regulate bone formation via a central mechanism (28). Thus, this complex story is likely just starting to unfold. What seems clear, however, is that it would be unwise to dismiss leptin as unimportant to bone. The bulk of the evidence would, in fact, suggest that leptin may be one of the key factors linking energy intake to bone metabolism via both central and peripheral mechanisms. I would like to thank Drs. B. Lawrence Riggs and James Levine for helpful comments and suggestions. This work was supported, in part, by NIH Grants AG-04875 and AR-27065. Cerebrospinal fluid