The adipokine leptin influences many organ systems, including the skeleton.1 Ducy and colleagues2 proposed that leptin negatively regulates bone mass via a hypothalamic relay (Fig. 1A). These investigators reported that leptin signaling–deficient mice having either a loss of function mutation in the gene for leptin (ob/ob mice) or a loss of function mutation in the long form of the leptin receptor (db/db mice) have high bone mass and that hypothalamic administration of leptin decreases bone mass in ob/ob mice by inhibiting bone formation through a pathway involving increased sympathetic signaling.3 Their findings suggested that blocking the skeletal effects of leptin would increase bone mass by stimulating bone formation and thus provide a novel intervention for treatment of osteoporosis.4, 5 However, in spite of initial enthusiasm for neuronal control of bone remodeling balance by leptin,6 the concept as originally articulated failed to explain the previously reported osteopenic skeletal phenotype of leptin receptor–deficient db/db mice and fa/fa rats,7, 8 the growth-promoting effects of leptin on the skeleton of leptin-deficient ob/ob mice, or the protective effect of peripheral administration of the hormone in bone loss models.9–11 In reconciliation, Burguera and colleagues10 proposed that the skeletal actions of peripheral leptin were bone anabolic whereas its central actions were catabolic (Fig. 1B). This dual-pathway model obtained support when Hamrick12 documented that ob/ob mice do not have high bone mass globally as initially reported2 but exhibit a mosaic skeletal phenotype; compared to wild-type (WT) mice, ob/ob mice have higher cancellous bone mass in lumbar vertebra, but lower cancellous bone mass in distal femur metaphysis, lower cortical and total bone mass, and decreased bone length. Thus, the balance between peripheral and hypothalamic actions of the hormone was hypothesized to be responsible for the complex skeletal phenotype observed in ob/ob mice. However, it remained unclear why bone accrual at the periosteum, a compartment richly innervated with fibers of sympathetic origin,13 was not increased in leptin signaling–deficient rodents.14 Also, in conspicuous conflict with the concept of neuronal control of bone metabolism, neonatal sympathectomy failed to increase bone mass15 as predicted by this model. Even more problematic was the failure of either the Ducy or the Burguera/Hamrick models to predict the skeletal response of leptin-deficient ob/ob mice to increased hypothalamic leptin, delivered via leptin gene therapy. Hypothalamic leptin gene therapy actually increased bone length and total bone mass in growing ob/ob mice to values that did not differ from WT mice.14 More recently, serum osteocalcin, a biochemical marker of bone formation, was shown to be increased in ob/ob mice following hypothalamic leptin gene therapy16 as well as after direct administration of the hormone into the hypothalamus.17 Other studies have also failed to replicate a high cancellous bone mass in lumbar vertebra of db/db mice.18
Overall, the aforementioned findings are in clear conflict with the hypothesis that the central nervous system–mediated actions of leptin are antiosteogenic. They also question the concept that peripheral leptin is solely responsible for the putative bone anabolic effects of the hormone. We therefore sought a unifying hypothesis to explain the complex skeletal changes observed in leptin-deficient mice. To accomplish this aim, we reevaluated the effect of leptin signaling on bone formation and resorption in ob/ob mice. We also evaluated whether the physiological skeletal effects of leptin on bone metabolism are mediated through peripheral or central pathways. Our findings described in this manuscript provide strong evidence that leptin acts physiologically, primarily through peripheral pathways, to enhance bone growth and maturation by increasing osteoblast number and activity, and osteoclast activity (Fig. 1C).
Subjects and Methods
Leptin-deficient ob/ob mice, leptin receptor-deficient db/db mice, and their homozygous WT littermates (Jackson Laboratory, Bar Harbor, ME, USA) were used in the experiments. The mice were housed individually in a temperature (21–23°C) and light-controlled (12-hour light-dark cycle) room under normal (experiments 1 and 3) or specific pathogen-free (experiment 2) conditions. Food and water were provided ad libitum to all animals. Body weight and food consumption were monitored weekly for the duration of studies. The mice were maintained in accordance with the NIH Guide for the Care and the Use of Laboratory Animals and the Institutional Animal Care and Use Committee approved the experimental protocols.
Experiment 1: Effect of Subcutaneous Leptin Administration on the Skeletal Phenotype of ob/ob Mice
Twelve-week-old female ob/ob mice were randomized by weight into two groups; ob/ob (n = 5) or ob/ob + leptin (n = 8). An additional control group consisted of WT littermates (n = 9). Leptin (40 µg/mouse) or vehicle was administered daily by subcutaneous injection for 3 weeks and the mice euthanized at 15 weeks of age. The mice were injected with the fluorochrome declomycin (15 mg/kg; Sigma Chemical, St. Louis, MO, USA) 9 days prior to euthanasia and the fluorochrome calcein (15 mg/kg; Sigma Chemical) 4 and 1 days prior to euthanasia to label mineralizing bone. The 9- and 1-day labels were used to measure endocortical bone formation and longitudinal growth and the 4- and 1-day labels were used to measure cancellous bone formation.
Experiment 2: Effect of Hypothalamic Leptin Gene Therapy on Osteoblasts in ob/ob Mice
Eight to 10-week-old male ob/ob mice were randomized by weight into two groups; ob/ob + recombinant adenoassociated virus-leptin (rAAV-Lep, n = 7) or control vector–encoding green fluorescent protein (rAAV-GFP, n = 7). An additional control group consisted of untreated WT mice (n = 3).
The rAAV-Lep and rAAV-GFP vectors were constructed and packaged as described.19 Mice were injected in the third ventricle of the hypothalamus with either rAAV-GFP (9 × 107 particles) or rAAV-Lep (9 × 107 particles) as described14, 20 and maintained for 15 weeks. Food intake, body weight, hypothalamic leptin mRNA expression, and hormone levels in response to rAAV-Lep therapy in these mice have been detailed.14, 20
Experiment 3: Effect of Peripheral Leptin on Osteoblasts
This study evaluated the effect of replacing WT bone marrow (BM) cells with BM cells from db/db mice. BM cells from db/db mice are incapable of responding to direct actions of leptin due to a loss of function mutation in the long form of the leptin receptor. Eight-week-old female WT mice were randomized by weight into three groups: untreated WT (n = 7), lethally irradiated WT BM recipients + WT donor BM (WT → WT) (n = 8), lethally irradiated WT BM recipients + db/db donor BM (db → WT) (n = 8). An additional control group consisted of untreated db/db mice (n = 7). For BM transplantation, BM cells were isolated from femurs and tibias of WT or db/db donor mice by flushing the bones with PBS and made into single-cell suspensions. After lysing of red blood cells, BM cells were resuspended to 5 × 107 cells/mL for BM transplantation. Transplant recipients were lethally irradiated at 9 Gy using a cobalt 60 irradiator source (Radiation Center, Oregon State University) and reconstituted with 1 × 107 donor BM cells by injection (200 µL) in the tail vein. The mice were maintained for 9 weeks following BM engraftment, labeled with calcein (15 mg/kg) 4 and 1 days prior to euthanasia, and euthanized at 17 weeks of age. Additionally, a group of WT mice (n = 4) were engrafted with BM from GFP mice to track cell repopulation in different tissues.
Tissue Collection and Analyses
Following euthanasia, 5th lumbar vertebrae and femora were excised for histomorphometry. In experiments 1 and 3, dissected abdominal white adipose tissue (WAT) was weighed. In experiment 3, blood was obtained at necropsy by cardiac puncture.
Serum osteocalcin was measured using mouse Gla-osteocalcin High Sensitive EIA kit from Clontech (Mountain View, CA, USA) and serum C terminal telopeptide of type I collagen (CTX) was measured using the mouse CTX ELISA kit from Life Sciences Advanced Technologies (St. Petersburg, FL, USA).
Lumbar vertebrae and femora were prepared for histomorphometric evaluation as described.21 Sections were stained according to the Von Kossa method with a tetrachrome counter stain (Polysciences, Warrington, PA, USA) for assessment of bone area and cell-based measurements. Sections were stained with toluidine blue (Sigma Chemical) for assessment of cartilage area and bone marrow cell density. Unstained sections were used for assessment of dynamic histomorphometry. Histomorphometric data were collected using the OsteoMeasure System (OsteoMetrics, Inc., Atlanta, GA, USA). Static cancellous bone endpoints included bone area (bone area/tissue area, %) and the derived architectural indices of trabecular number (mm−1), trabecular thickness (µm), and trabecular separation (µm). Osteoblast and osteoclast perimeters were also measured and expressed as percent of total bone perimeter. Fluorochrome-based indices of bone formation included mineralizing perimeter (percentage of bone with double-label + ½ single-label) and mineral apposition rate (mean distance between two fluorochrome markers that comprise a double-label divided by interlabel time, µm/d). Bone formation rate was calculated using a bone perimeter referent (µm2/µm/year). All data are reported using standard nomenclature.22
GFP-specific PCR was performed to detect the presence of GFP donor cells in BM transplant recipients. Genomic DNA was isolated from various tissues following BM transplantation in recipient mice using Qiagen DNeasy Blood and Tissue kit (Qiagen, Valencia, CA, USA). Purified genomic DNA (100 ng per reaction) was used to detect the presence of GFP by PCR using GFP-specific primers as published by Jackson Laboratory (IMR0872: AAGTTCATCTGCACCACCG; IMR1416: TCCTTGAAGAAGATGGTGCG) that yield a 173–base pair (bp) PCR product.
A one-way ANOVA followed by a Bonferroni post-hoc test was used to evaluate differences among treatment groups. If ANOVA assumptions of homogeneity of variance were not met, a Kruskal-Wallis followed by a Tamhane post hoc test was used. Statistical analysis was performed using SPSS 19.0 (SPSS Inc., Chicago, IL, USA). Differences were considered significant at p < 0.05. All data are expressed as mean ± SE.
Subcutaneous Administration of Leptin Stimulates Bone Formation
Experiment 1 was performed to assess the effects of leptin deficiency and subcutaneous leptin replacement on measures of energy metabolism and indices of bone formation and resorption in leptin-deficient ob/ob mice. As expected, food consumption, body weight, and abdominal white adipose tissue weight were higher in ob/ob mice than in WT mice and subcutaneous administration of leptin lowered these measures (Table 1). In addition, ob/ob mice had greater cancellous bone area in lumbar vertebra but normal cancellous bone area in distal femur metaphysis and epiphysis compared to WT mice (Table 1).
Table 1. Effects of Leptin Treatment on Food Consumption, Body Weight, and WAT Weight and Cancellous Bone Histomorphometry in Lumbar Vertebra and Distal Femur Metaphysis and Epiphysis, and Cortical Bone Histomorphometry in the Midshaft Femur Diaphysis
The effects of leptin deficiency and leptin replacement on bone formation are shown in Fig. 2 and Table 1. Cancellous bone formation rate (Fig. 2A–C) and osteoblast-lined cancellous bone perimeter (Fig. 2H–J) were lower in lumbar vertebra, and distal femur metaphysis and epiphysis in ob/ob compared to WT mice. Lower mineralizing perimeter in lumbar vertebra and lower mineralizing perimeter and mineral apposition rate in the femur metaphysis and epiphysis (Table 1) contributed to the lower bone formation rate at the respective skeletal sites. Femoral endocortical mineralizing perimeter (Table 1) and osteoblast-lined bone perimeter (Fig. 2K) and longitudinal growth rate (Fig. 2O) were also lower in ob/ob compared to WT mice. Subcutaneous administration of leptin to ob/ob mice for 3 weeks resulted in higher bone formation rate (Fig. 2A–D), mineralizing perimeter (Table 1), mineral apposition rate (Table 1), osteoblast-lined bone perimeter (Fig. 2H–K), and longitudinal growth rate (Fig. 2O) compared to vehicle-treated ob/ob mice and, in the majority of cases, compared to WT mice. Representative photomicrographs illustrating dynamic and static bone histomorphometry in WT, ob/ob, and leptin-treated ob/ob mice are shown in Fig. 2E–G, L–N, and P–R.
The effects of leptin deficiency and leptin replacement on bone resorption are shown in Fig. 3A–H. Osteoclast-lined bone perimeter, an index of bone resorption, was higher in the femoral epiphysis of ob/ob compared to WT mice (Fig. 3C). However, cartilage remnants, indicative of reduced bone resorption, were relatively uncommon in WT but greatly increased in the femoral epiphysis in ob/ob (79-fold increase) and db/db (189-fold increase) mice (Fig. 3E–H). This unexpected finding is indicative of a defect in osteoclast number and/or activity. Measurement of serum biochemical markers of bone turnover revealed that CTX, a marker of bone resorption, and osteocalcin, a marker of bone formation, were reduced in db/db mice compared to age-matched WT mice by 99% and 69%, respectively (Fig. 3I, J). Significant differences in osteoclast-lined bone perimeter between WT and ob/ob mice were not detected in lumbar vertebra or femoral metaphysis, but tended to be higher in the femoral diaphysis of the mutant mice (Fig. 3A, B, D). Similar to the epiphysis, cartilage remnants were detected in the lumbar vertebra and femoral metaphysis and diaphysis (data not shown). Administration of leptin had no significant effect on osteoclast perimeter in lumbar vertebra, femoral metaphysis, or femoral epiphysis. However, endocortical osteoclast-lined bone perimeter was lower in leptin-treated ob/ob mice compared to vehicle-treated ob/ob mice (Fig. 3D).
Hypothalamic Leptin Is Not Antiosteogenic
Experiment 2 was performed to assess the effects of central (hypothalamic) leptin via rAAV-Lep gene therapy on bone in ob/ob mice. The effects of hypothalamic leptin gene therapy on bone histomorphometry are shown Fig. 4. rAAV-GFP-treated ob/ob mice had higher vertebral cancellous bone area and trabecular number, did not differ in trabecular thickness, and exhibited lower trabecular spacing compared to WT mice (Fig. 4A–D). Osteoblast-lined bone perimeter was lower in rAAV-GFP ob/ob mice than in WT mice (Fig. 4E). Significant differences in osteoclast-lined bone perimeter were not detected among treatment groups (Fig. 4F). Selectively increasing leptin levels in the hypothalamus via rAAV gene therapy for 15 weeks restored vertebral bone mass, architecture, and osteoblast-lined bone perimeter to values that did not differ from WT mice (Fig. 4A–E).
Antagonism of Peripheral Leptin Inhibits Bone Formation
Experiment 3 was performed to assess the differential effects of peripheral leptin on measures of energy metabolism and bone formation in WT mice. The effects of BM transplantation of GFP-positive cells into WT mice on tissue distribution are shown in Fig. 5A. BM transplantation resulted in GFP-positive cells in spleen, liver, kidney, pancreas, and abdominal white adipose tissue. GFP-positive cells were not detected in brown adipose tissue or the central nervous system (brain and hypothalamus).
The effects of transplantation of WT and db/db BM cells on food intake, body weight, abdominal white adipose tissue weight, and blood glucose levels are shown in Fig. 5B–E. As expected, food intake, body weight, abdominal white adipose tissue weight, and serum glucose were higher in db/db mice than in WT mice. BM transplantation resulted in lower mean food intake and body weight compared to WT mice. However, despite replacement of peripheral tissues with db/db-derived cells, significant differences in food intake, body weight, abdominal white adipose tissue, or blood glucose were not detected between WT → WT mice and db → WT mice.
The effects of high-dose radiation and bone marrow transplantation on bone marrow cell density in the distal femur diaphysis are shown in Fig. 6. The drastic reduction in bone marrow cell density 3 days following irradiation is illustrated in Fig. 6A. The mouse shown did not receive bone marrow transplantation. The histological appearance of bone marrow of irradiated WT mice 9 weeks following transplantation with either bone marrow from WT or db/db mice was indistinguishable from normal (Fig. 6C–F). Quantitative measurement of bone marrow cell density is shown in Fig. 6B and confirms that marrow transplantation restored WT bone marrow cell density in WT → WT mice and db → WT mice. Also, there was no difference in bone marrow cell density between WT and db/db mice.
The effects of transplantation of WT and db/db BM cells on bone formation and bone resorption are shown in Fig. 7 and Table 2. Bone formation rate was lower in femoral metaphysis, epiphysis, and endocortex, and tended to be lower in lumbar vertebra in db/db compared to WT mice (Fig. 7A–D). The lower bone formation in db/db mice at the skeletal sites measured (femur diaphysis, metaphysis, and epiphysis) was in part due to a lower mineralizing perimeter (Table 2). Similarly, osteoblast-lined bone perimeter was lower at each skeletal site in db/db mice. A lower mineral apposition rate in db/db mice contributed to the lower bone formation rate in femur diaphysis. Significant differences in osteoclast-lined bone perimeter were not detected between WT and db/db mice at any of the four skeletal sites measured. Bone area was higher in the lumbar vertebra of db/db mice compared to WT mice but significant differences between these mice were not detected in distal femur metaphysis and epiphysis.
Table 2. Effects of Transplantation of Bone Marrow From WT Mice or Leptin Receptor-Deficient db/db Mice Into Lethally Irradiated WT Mice on Cancellous Bone Histomorphometry in Lumbar Vertebra and Distal Femur Metaphysis and Epiphysis, and Cortical Bone Histomorphometry in the Midshaft Femur Diaphysis
Significant differences in bone formation rate were not detected between WT mice and WT mice that received WT BM cells (WT → WT) (Fig. 7A–D). In contrast, replacing WT BM with db/db BM in db → WT mice resulted in a bone formation rate that did not differ from untreated db/db mice. Representative photomicrographs illustrating dynamic bone histomorphometry in WT, WT → WT, db → WT, and db/db mice are shown in Fig. 7E–H. Significant differences in mineralizing perimeter were not detected between WT mice and WT mice that received WT BM cells (WT → WT) (Table 2). In contrast, replacing WT BM with db/db BM in db → WT mice resulted in mineralizing perimeter that did not differ from untreated db/db mice. Reduced mineral apposition rate contributed to the reduction in bone formation rate in femur diaphysis of db → WT mice. Osteoblast-lined bone perimeter was lower following bone marrow transplantation in WT → WT mice in femur diaphysis and lower at all sites in db → WT mice. Finally, marrow transplantation had no effect on osteoclast-lined bone perimeter at any of the four skeletal sites examined.
Bone formation was measured in 4-month-old leptin-deficient ob/ob mice, leptin receptor-deficient db/db mice, and WT littermate mice by dynamic bone histomorphometry in lumbar vertebra and femur. Our choice of regions of interest was based on their clinical and biomechanical significance. The sampling sites are representative of the axial (lumbar vertebra) and appendicular (femur) skeleton, cortical (femoral diaphysis), and cancellous (lumbar vertebra and femoral metaphysis and epiphysis) bone, and varying levels of weight bearing (femoral diaphysis and epiphysis as representative of high weight bearing sites and lumbar vertebra and femoral metaphysis as representative of low weight bearing sites). Compared to WT mice, bone formation rate was consistently lower in ob/ob and db/db mice. The markedly lower value for osteoblast-lined bone perimeter and mineralizing perimeter in leptin-deficient ob/ob mice provides strong evidence that leptin increases bone formation, in part, by increasing osteoblast number. Indeed, subcutaneous replacement of leptin in ob/ob mice resulted in a site-independent increase in bone formation and osteoblast-lined perimeter. Further, positive effects of subcutaneous leptin replacement on mineral apposition rate in ob/ob mice suggest that the hormone also increases osteoblast activity.
In contrast to our findings, Ducy and colleagues2 reported that leptin-deficient ob/ob mice have increased bone formation due to a higher mineral apposition rate. We considered the possibility that age, gender, or cohort differences may explain the discrepancy in results. To rule out the possibility of gender-specific effects, we analyzed male as well as female ob/ob mice. To rule out age as a factor, we have analyzed mice at 1.5, 2, 4, 6.5, and 10 months of age. Depending on the study, we measured osteoblast-lined bone perimeter, dynamic histomorphometry, and/or serum osteocalcin. Histomorphometric analysis was performed at a variety of skeletal sites representative of cortical and cancellous bone and representative of the axial and appendicular skeleton. To date, we have analyzed ob/ob or db/db mice from 11 separate cohorts. In no case was bone formation elevated in the mutant mice. Indeed, administration of leptin to ob/ob mice dramatically increased bone formation at all four representative skeletal sites and hypothalamic leptin gene therapy increased osteoblast-lined bone perimeter and serum osteocalcin levels. Taken together, this comprehensive body of published14, 16, 23–25 and unpublished work has failed to identify any condition in which leptin suppresses bone formation. As such, we have no explanation for the discrepancy in the results, although it may be relevant that, based on the work of Erben,26 the long fluorochrome labeling interval used by Ducy and colleagues2 to estimate bone formation would be expected to lead to unreliable results due to excessive label escape error.
Our leptin replacement study, in which the hormone was administered subcutaneously, does not distinguish between direct peripheral and indirect central actions of leptin on bone formation because peripherally administered leptin crosses the blood brain barrier.27 To address the possibility that central leptin has a unique effect on osteoblasts, we performed hypothalamic leptin gene therapy. Selectively increasing leptin levels in the hypothalamus of ob/ob mice by gene therapy normalized bone mass and architecture to WT values, and similar to subcutaneous administration of the hormone, increased osteoblast-lined bone perimeter. This finding provides strong evidence that peripheral and central leptin do not have opposing actions on osteoblasts. The experiments described above suggest that leptin signaling increases bone formation throughout the skeleton. This conclusion is further supported by the observation that serum osteocalcin, a biochemical marker that reflects overall bone formation, is higher in WT than db/db or ob/ob mice.16, 18 However, it is not clear from the present or previous studies whether the physiological actions of leptin are mediated through central or peripheral pathways.
Leptin receptors have been reported in a variety of cells, including those of the osteoblast lineage.28–34 Also, leptin has been shown to have direct effects on primary osteoblasts to increase mineralization of bone matrix in cell culture.28, 32, 34–41 It would therefore be surprising if peripheral leptin did not have actions on the skeleton. It is well established that peripheral leptin levels are much higher than central levels.27 Thus, normal circulating levels of leptin would be anticipated to act preferentially through peripheral pathways. In contrast, much higher circulating levels of leptin (or direct administration of the hormone into the hypothalamus) may be required to modulate bone metabolism centrally.
To verify the physiological role of peripheral leptin signaling on bone formation, we evaluated long-duration changes in bone formation following BM transplantation of leptin receptor-deficient db/db donor BM cells into lethally irradiated WT recipient mice (db → WT). Selectively antagonizing peripheral leptin signaling by replacing WT BM with db/db BM in db → WT mice resulted in drastically lower bone formation in WT mice, to values that did not differ from untreated db/db mice. Furthermore, the reduced bone formation was largely due to a decrease in osteoblast-lined bone perimeter. Reduced bone formation was not due to irradiation and transplantation artifacts because WT mice that received WT BM cells (WT → WT) had normal bone formation. Also, high-dose irradiation results in a drastic reduction in BM cell density,42 which was restored in the present studies to normal levels following transplantation with BM cells derived from either WT or db/db mice.
BM transplantation in growing mice resulted in replacement of BM cells in peripheral organs, with no evidence for replacement of cells in the central nervous system. Although we cannot be absolutely certain that no db/db donor cells were transplanted into the brain, cell tracking using GFP cells indicates that their number, if present, would be diminishingly small. It seems unlikely that incorporation of a very small number of leptin receptor–deficient cells in the continued presence of much larger numbers of leptin receptor–positive cells would impact leptin signaling. This conclusion is further supported by our observation that transplantation of db/db donor cells into WT mice had no effect on hypothalamic regulation of energy metabolism.
The effects of leptin on bone metabolism do not appear to be directly coupled to the well-established actions of the hormone on energy metabolism. In agreement with numerous prior studies, both subcutaneous administration of leptin and hypothalamic leptin gene therapy20 decreased food consumption, body weight gain, and abdominal white adipose tissue weight. In contrast, decreasing peripheral leptin signaling via BM transplantation of leptin receptor–deficient db/db BM cells into WT mice had no impact on these endpoints. These results clearly demonstrate that central leptin signaling was not altered by irradiation and BM transplantation.
Compared to WT mice, leptin-deficient ob/ob mice had normal cancellous bone volume in distal femur metaphysis and epiphysis, increased cancellous bone volume in lumbar vertebra, decreased cortical bone volume in the diaphysis, and shorter bones.14 The deficit in bone length in the mutants suggests that leptin is required for optimal bone growth.43 The present studies, demonstrating an impaired femur growth rate in ob/ob mice and accelerated bone growth in ob/ob mice treated with leptin, support this view. Based on our findings, we conclude that an overall higher rate of bone acquisition is at least partially responsible for the higher bone mass observed in leptin-replete compared to leptin signaling–deficient rodents.
Research on the regulatory action of leptin on bone metabolism has traditionally focused on bone formation. However, the present studies revealed that leptin is also important for normal bone resorption. The normal or elevated cancellous bone mass in leptin signaling–deficient mice, in spite of drastically reduced bone formation, implies a concomitant reduction in bone resorption. Histological analysis revealed that leptin signaling–deficient ob/ob and db/db mice expressed a skeletal phenotype consistent with mild osteopetrosis. Specifically, the mutant mice were less effective in replacing calcified cartilage with bone. However, the decrease in bone resorption in the mutant mice was not associated with a reduction in osteoclast-lined bone perimeter. This finding provides evidence that osteoclast activity was reduced. In support of this conclusion, there was a dramatic reduction in the bone resorption marker CTX in serum of db/db mice.
Based on densitometry and biomechanical evidence, Ealey and colleagues44 concluded that femoral mineral and matrix components were compromised in ob/ob and db/db mice. Recent studies, evaluating bone material properties, support these findings.18 Thus, pathological retention of cartilage due to impaired skeletal maturation provides a plausible mechanism for the poor bone quality observed in leptin signaling-deficient rodents.
The present data suggest that the “mosaic” skeletal phenotype observed in leptin signaling–deficient mice arises because leptin acts on growth plate cartilage cells, osteoblasts, and osteoclasts to enhance their number and/or activity. Based on this model, the skeletal changes observed at different skeletal sites in leptin signaling–deficient mice may share a common mechanism; leptin deficiency results in an overall decrease in bone turnover. However, as found with other major regulators of bone metabolism,45 regional changes in bone mass and architecture depend upon the local prevalence of osteoblasts, osteoclasts, and their precursors. As a consequence, decreased bone formation prevails to lower bone accrual in growing mice at skeletal sites where bone formation greatly exceeds bone resorption (eg, periosteum). In contrast, at cancellous bone sites undergoing a high rate of bone turnover, reduced bone resorption may preserve trabecular number, providing a template for addition of new bone. By this mechanism, leptin can result in local increases in bone mass in spite of an overall reduction in bone turnover.
In summary, the present studies contradict the widely held view that leptin, acting primarily through a hypothalamic relay, reduces bone mass accrual by suppressing bone formation. The results instead suggest that peripheral leptin is essential for normal bone resorption and enhances bone formation. Indeed, accelerated bone growth and turnover in ob/ob mice treated with leptin results in an overall increase in bone mass and a skeleton with improved bone quality. Thus, efforts aimed at increasing bone mass by antagonizing hypothalamic leptin signaling46 are based on an incorrect premise and may be counterproductive. Furthermore, the observed drastic reduction in bone formation following selective attenuation of peripheral leptin signaling in normal mice suggests that conclusions derived from studies involving direct administration of leptin into the hypothalamus may greatly exaggerate the importance of central regulation of bone mass by leptin.
All authors state that they have no conflicts of interest.
This work was supported by National Institute of Health grants AR 054609 and AR 060913 (to UTI) and DK 37273 (to SPK). We thank Dr. Scott Mann for irradiating the mice.
Authors' roles: Study design: RTT, SPK, and UTI. Implementation of studies: CPW, KAP, and SB. Data collection: RTT, CPW, KAP, and LBL. Data analysis: UTI. Drafting manuscript: RTT and UTI. Revising manuscript content: RTT, SPK, CPW, KAP, LBL, SB, and UTI. Approving final version: RTT, SPK, CPW, KAP, LBL, SB, and UTI. UTI takes responsibility for the integrity of the data analysis.