Failure to Generate Bone Marrow Adipocytes Does Not Protect Mice from Ovariectomy-Induced Osteopenia

Abstract

A reciprocal association between bone marrow fat and bone mass has been reported in ovariectomized rodents, suggesting that bone marrow adipogenesis has a negative effect on bone growth and turnover balance. Mice with loss of function mutations in kit receptor (kit^W/W-v) have no bone marrow adipocytes in tibia or lumbar vertebra. We therefore tested the hypothesis that marrow fat contributes to development of osteopenia by comparing the skeletal response to ovariectomy (ovx) in growing wild type (WT) and bone marrow adipocyte-deficient kit^W/W-v mice. Mice were ovx at 4 weeks of age and sacrificed 4 or 10 weeks post-surgery. Body composition was measured at necropsy by dual-energy X-ray absorptiometry. Cortical (tibia) and cancellous (tibia and lumbar vertebra) bone architecture were evaluated by microcomputed tomography. Bone marrow adipocyte size and density, osteoblast- and osteoclast-lined bone perimeters, and bone formation were determined by histomorphometry. Ovx resulted in an increase in total body fat mass at 10 weeks post-ovx in both genotypes, but the response was attenuated in the in kit^W/W-v mice. Adipocytes were present in bone marrow of tibia and lumbar vertebra in WT mice and bone marrow adiposity increased following ovx. In contrast, marrow adipocytes were not detected in either intact or ovx kit^W/W-v mice. However, ovx in WT and kit^W/W-v mice resulted in statistically indistinguishable changes in cortical and cancellous bone mass, cortical and cancellous bone formation rate, and cancellous osteoblast and osteoclast-lined bone perimeters. In conclusion, our findings do not support a causal role for increased bone marrow fat as a mediator of ovx-induced osteopenia in mice.

1. INTRODUCTION

The physiological significance of bone marrow fat is currently a topic of debate [1–4]. Bone marrow fat is an important depot for storage of triglycerides. Fat depots generally act as energy reserves in which fatty acids are taken up by adipocytes from lipoproteins and subsequently released into the general circulation following lipolysis. Although bone marrow fat is highly labile [5], bone marrow adipocytes increase in number and sequester and esterify fatty acids during severe caloric restriction, suggesting that the bone marrow fat depot does not function as a typical energy reservoir for peripheral tissues [6].

Age-related reciprocal changes in bone marrow adiposity and bone mass are accelerated in humans following menopause, during disuse, and during treatment with glucocorticoids, and in animal models for pathological bone loss [7–12]. These findings have fueled speculation that increased marrow fat may lead to bone loss. Proposed mechanisms for the putative detrimental effects of marrow fat on bone metabolism fall into two general, though non-mutually exclusive, categories: 1) differentiation of stromal cells into adipocytes occurring at the expense of osteoblast differentiation, and 2) adipocyte production of factors (adipokines) which antagonize bone formation and/or increase bone resorption [13].

It is well established that osteoblasts and adipocytes are derived from a common bone marrow progenitor [14]. A deficiency in the transcription factor PPARγ reduces marrow fat and enhances osteogenesis in mice [15]. In contrast, osteoblast-targeted overexpression of PPARγ exaggerates ovariectomy (ovx)-induced osteopenia in mice [16], while PPARγ agonists are associated with bone loss in insulin-resistant humans [17]. Taken together, these findings suggest increased adipocyte differentiation curtails the differentiation of osteoblasts.

Estrogen plays a key role in skeletal growth and maturation and in the maintenance of bone balance in adults [18]. As in humans, gonadal insufficiency in female rodents is associated with increased peripheral fat accumulation, increased bone marrow adiposity, accelerated bone turnover, and osteopenia [19–23]. However, in spite of a strong inverse association between bone marrow fat and bone mass, it has proven difficult to directly test the hypothesis that excess fat accumulation in bone marrow is responsible for ovx-induced osteopenia. In this regard, we recently reported that kit receptor-deficient kit^W/W-v mice have no bone marrow adipocytes [24] but are able to produce osteoblasts and osteoclasts. We therefore reasoned that if bone marrow adipose tissue contributes to the detrimental skeletal effects of estrogen deficiency, kit^W/W-v mice would be resistant to ovx-induced osteopenia. We tested this hypothesis by determining the effect of ovx on body composition, bone mass and architecture, and dynamic and cellular indices of bone turnover in growing WT and kit^W/W-v mice.

2. MATERIALS AND METHODS

The experimental protocol was approved by the Institutional Animal Care and Use Committee. Animals were maintained under standard conditions with a 12-h light, 12-h dark cycle in accordance with the NIH Guide for the Care and Use of Laboratory Animals. Food (standard chow) and water were provided ad libitum.

Two experiments were conducted in 4-week-old ovary-intact and ovx female WBB6F1/J-kit^W/kit^W-v (kit^W/W-v) and wild type (WT) WBB6F1/J littermates purchased from the Jackson Laboratory (Bar Harbor, ME, USA). In the first experiment, WT and kit^W/W-v mice were maintained for 10 weeks following surgery and sacrificed at 14 weeks of age to evaluate the long-term skeletal response to ovx. In the second, follow-up experiment, the mice were maintained for 4 weeks following surgery and sacrificed at 8 weeks of age to assess short-term skeletal response to ovx. Specifically, the follow up study was designed to evaluate whether lack of adipocytes influences the effects of ovx on cancellous bone formation. The fluorochrome calcein (Sigma, St Louis, MO) was administered at 20 mg/kg 7 and 1 days prior to sacrifice (14-week-old mice) or 4 and 1 days prior to sacrifice (8-week-old mice). The mice were anesthetized with 2–3% isofluorane delivered in oxygen and body composition determined via dual energy x-ray absorptiometry (DXA) using a PIXImus small animal densitometer (Lunar Corp., Madison, WI, USA). Death was induced by cardiac excision, and tibiae (both experiments) and lumbar vertebrae (experiment 2 only) were removed and stored in 70% ethanol for micro-computed tomography (μCT) and histomorphometric analysis. Tibial length was measured using digital calipers (Little Machine Shop, Pasadena, CA).

2.1 Micro-computed Tomography

μCT was used for nondestructive three-dimensional evaluation of bone architecture. Tibiae and lumbar vertebrae were scanned using a Scanco μCT40 scanner (Scanco Medical AG, Basserdorf, Switzerland) at a voxel size of 12 × 12 × 12 μm and evaluated at a threshold of 265 (gray scale, 0–1000). The threshold value was determined empirically. Total tibia (cancellous plus cortical bone) was evaluated followed by evaluation of cortical bone in the tibial diaphysis and cancellous bone in the proximal tibial metaphysis. For cortical bone, 20 slices (0.24 mm) of bone 0.7 mm proximal to the tibio-fibular junction were measured and cross-sectional tissue volume (cortical and marrow volume, mm³), cortical volume (mm³), marrow volume (mm³), and cortical thickness (μm) determined [25]. For the tibial metaphysis, 60 slices (0.72 mm) of bone beginning 0.75 mm distal to the growth plate were measured. Analysis of the lumbar vertebra included the entire region of secondary spongiosa between the cranial and caudal growth plates. Cancellous bone measurements in tibia and lumbar vertebra included cancellous bone volume/tissue volume (%), connectivity density (1/mm³), trabecular number (mm⁻¹), trabecular thickness (μm) and trabecular separation (μm) [26].

2.2 Histomorphometry

The histological methods used here have been described in detail [27]. In brief, the proximal tibial metaphysis and lumbar vertebra were dehydrated in graded increases of ethanol and xylene and embedded undecalcified in methyl methacrylate. Sections (4 μm thick) were cut with vertical bed microtomes (Leica/Jung 2065 and 2165) and affixed to slides precoated with a 1% gelatin solution. Sections were stained with toluidine blue or acid phosphatase with a toluidine blue counterstain for cellular measurements or mounted unstained for measurement of fluorochromes. Osteoblasts and adipocytes were measured in the toluidine blue-stained sections and osteoclasts were measured in acid phosphatase/toluidine blue-stained sections. All cell data were collected under visible light and all fluorochrome data were collected under ultraviolet light using the OsteoMeasure System (OsteoMetrics, Inc., Atlanta, GA). The sampling site for the proximal tibial metaphysis was located 0.25–1.25 mm distal to the growth plate, included secondary spongiosa only, and averaged 1.2 ± 0.1 mm² (mean ± SE). The sampling site in lumbar vertebra included the entire cancellous compartment and averaged 1.75 ± 0.05 mm². Both sampling sites excluded cortical margins.

Tissue area, bone area, osteoblast perimeter (osteoblast perimeter/bone perimeter, %) and osteoclast perimeter (osteoclast perimeter/bone perimeter, %) were measured. Adipocyte number and area were also measured and expressed as bone marrow adiposity (% tissue area occupied by adipocytes), adipocyte density (number/mm²) and mean adipocyte size (μm²). Adipocytes were identified morphologically as large circular or oval-shaped cells bordered by a prominent cell membrane and lacking cytoplasmic staining due to alcohol extraction of intracellular lipids during processing. This method was previously validated by fat extraction and analysis [5]. Fluorochrome-based indices of bone formation included: 1) mineralizing perimeter (cancellous bone perimeter covered with double plus half single fluorochrome label normalized to bone perimeter, %), 2) mineral apposition rate (the distance between two fluorochrome markers that comprise a double label divided by interval (3 days for cancellous bone) between label administration, μm/day), and 3) bone formation rate (calculated by multiplying mineralizing perimeter by mineral apposition rate normalized to bone perimeter, μm²/μm/yr).

For cortical bone measurements in the tibia, undecalcified cross-sections (150 μm) were cut at the tibiofibular synostosis using a low-speed diamond-edge saw (Isomet; Buehler, Lake Bluff, IL, USA) and ground on a roughened glass plate to ~20 μm. Fluorochrome-based indices of cortical bone formation were measured in the diaphysis (periosteal or endocortical perimeters) on one cross section/bone and included: 1) mineralizing perimeter, 2) mineral apposition rate (6 day labeling interval for cortical bone), and 3) bone formation rate. Mineralizing perimeter, mineral apposition rate and bone formation rate were calculated as for cancellous bone.

2.3 Statistics

Two-way ANOVA was used to evaluate the respective effects of genotype, surgery and genotype x surgery interactions (SPSS 19.0, SPSS Inc., Chicago, IL). Differences were considered significant at P < 0.05. All data are reported as mean ± SE.

3. RESULTS

3.1 Long-term Effects of Ovx on the Skeleton of WT and Bone Marrow Adipocyte-Deficient Mice

The effects of genotype (WT vs. kit^W/W-v) and surgery (sham-operated ovary-intact vs. ovx) on body composition in 14-week-old mice 10 weeks following surgery are shown in Figure 1. Total body mass, fat mass, and fat expressed as % body mass was higher in WT than kit^W/W-v mice, and higher in ovx than ovary-intact animals. A significant genotype by surgery interaction was not detected for total body weight. However, the interaction was significant for fat mass and % fat mass. Specifically, compared to WT mice, kit^W/W-v mice exhibited a reduced increment in fat mass following ovx.

Figure 1.

The effects of genotype (WT vs. kit^W/W-v) and surgery (ovary-intact vs. ovx) on total body mass (a), fat mass (b), and fat mass expressed as % body mass (c) in 14-week-old mice 10 weeks following surgery. Values are mean ± SE. ^aP<0.05 compared to ovary-intact within genotype, ^a*P<0.1 compared to ovary-intact within genotype.

The effects of genotype and surgery on total tibia bone volume, tibial length, and cortical bone architecture in the tibial diaphysis are shown in Table 1. Tibiae were longer in WT than kit^W/W-v mice and longer in ovx mice than in ovary-intact mice. Tibial bone volume was also higher in WT than kit^W/W-v mice but lower in ovx mice than in ovary-intact mice. Significant genotype by surgery interactions were not detected for either tibial length or total tibial bone volume. Cross-sectional tissue volume, cortical volume, marrow volume, and cortical thickness were higher in WT mice than in kit^W/W-v mice. Surgery had no effect on cross-sectional tissue volume. However, cortical volume and thickness were lower and medullary volume was higher in ovx mice compared to ovary-intact mice. A significant genotype by surgery interaction was not detected for any of the cortical endpoints evaluated.

Table 1.

The effect of kit receptor deficiency and ovariectomy on total tibia bone volume and length and cortical bone architecture proximal to the tibio-fibular junction in 14-week-old mice at 10 weeks following surgery.

	WT		Kit W/W-v		ANOVA (P<)
	Intact	Ovx	Intact	Ovx	Genotype	Surgery	Interaction
Total Tibia
Bone length (mm)	17.56 ± 0.10	17.96 ± 0.04	17.31 ± 0.12	17.73 ± 0.08	0.003	0.001	NS
Bone volume (mm3)	16.35 ± 0.24	14.82 ± 0.23	13.99 ± 0.39	12.93 ± 0.34	0.001	0.001	NS
Tibial Diaphysis (cortical bone)
Tissue volume (mm3)	0.181 ± 0.002	0.177 ± 0.002	0.162 ± 0.002	0.159 ± 0.002	0.001	NS	NS
Cortical volume (mm3)	0.137 ± 0.002	0.128 ± 0.002	0.123 ± 0.002	0.119 ± 0.002	0.001	0.001	NS
Marrow volume (mm3)	0.044 ± 0.001	0.048 ± 0.001	0.039 ± 0.001	0.041 ± 0.002	0.001	0.03	NS
Cortical thickness (μm)	257 ± 4	238 ± 2	241 ± 2	231 ± 4	0.001	0.001	NS

Data are mean ± SE

The effects of genotype and surgery on cancellous bone measurements in the proximal tibial metaphysis and dynamic cortical bone measurements in the tibial diaphysis are shown in Figure 2. Cancellous bone volume/tissue volume was higher in WT than kit^W/W-v mice and lower in ovx mice than in ovary-intact mice. These differences can be readily appreciated in images depicted in Figure 3 (a–d). A significant genotype by surgery interaction was not detected for cancellous bone volume/tissue volume. Osteoblast and osteoclast perimeters were higher in kit^W/W-v than WT mice and higher in ovx mice than in ovary-intact mice. However, significant genotype by surgery interactions were not detected for either endpoint evaluated.

Figure 2.

The effects of genotype (WT vs. kit^W/W-v) and surgery (ovary-intact vs. ovx) on cancellous bone volume/tissue volume (BV/TV) (a), osteoblast perimeter (Ob.Pm/B.Pm) (b), osteoclast perimeter (Oc.Pm/B.Pm) (c), endocortical mineralizing perimeter/bone perimeter (M.Pm/B.Pm) (d), mineral apposition rate (MAR) (e), and bone formation rate/bone perimeter (BFR/B.Pm) (f), periosteal M.Pm/B.Pm (g), MAR (h), and BFR/B.Pm (i), and on bone marrow adiposity (j), adipocyte density (k), and adipocyte size (l) in the proximal tibial metaphysis of 14-week-old mice at 10 weeks following surgery. Values are mean ± SE. ^aP<0.05 compared to ovary-intact within genotype. ^a*P<0.1 compared to ovary-intact within genotype.

Figure 3.

Representative microcomputed tomography images of cancellous bone in the proximal tibial metaphyses from intact WT, ovx WT, intact kit^W/W-v and ovx kit^W/W-v mice (a–d). Please note the lower bone volume in ovx compared to intact mice. Also shown are photomicrographs of histological sections from intact WT, ovx WT, intact kit^W/W-v and ovx kit^W/W-v mice showing adipocytes demarcated by an asterisk (e–h) in toluidine blue-stained sections from the proximal tibial metaphysis. Please note the absence of adipocytes in the kit^W/W-v mice.

Differences in endocortical and periosteal mineralizing perimeter, mineral apposition rate, or bone formation rate in the tibial diaphysis were not detected with genotype. Endocortical and periosteal bone formation rate was higher in ovx than ovary-intact mice. This increase was due predominantly to a greater increase in mineralizing perimeter at both envelopes. Significant genotype by surgery interactions were not detected for any endocortical or periosteal dynamic endpoints measured.

Bone marrow adipocytes were common in proximal tibia of WT mice. Bone marrow adiposity (% tissue area with fat) and adipocyte size were higher in ovx than in ovary-intact WT mice. Also, there was a strong tendency (P=0.1) for adipocyte density to be higher in ovx than ovary-intact WT mice. Adipocytes were not detected in either ovary-intact or ovx kit^W/W-v mice. The genotype and ovx-induced differences in bone marrow adiposity can be appreciated in images depicted in Figure 3 (g–h).

3.2 Short-term Effects of Ovx on the Skeleton of WT and Bone Marrow Adipocyte-Deficient Mice

The effects of genotype and surgery on body composition in 8-week-old mice 4 weeks following surgery are shown in Figure 4. Total body mass, fat mass, and fat expressed as % body mass was higher in WT than kit^W/W-v mice. Total body mass was higher in ovx compared to ovary-intact animals. While differences in fat mass were not detected with surgery, fat expressed as % body mass was lower in ovx compared to intact mice at 4 weeks post-surgery. Significant genotype by surgery interactions were not detected for total body mass, fat mass or fat expressed as % body mass.

Figure 4.

The effects of genotype (WT vs. kit^W/W-v) and surgery (ovary-intact vs. ovx) on total body mass (a), fat mass (b), and fat mass expressed as % body mass (c) in 8-week-old mice 4 weeks following surgery. Values are mean ± SE.

The effects of genotype and ovx on tibial length, total tibial bone volume, and tibial cortical and cancellous bone architecture are shown in Table 2. Tibial length did not differ between WT and kit^W/W-v mice. Tibiae were longer in ovx mice than in ovary-intact mice. Tibial bone volume was greater in WT than kit^W/W-v mice and lower in ovx mice than in ovary-intact mice. A significant genotype by surgery interaction was not detected for either tibial length or total tibial bone volume. Diaphyseal cross-sectional tissue volume, cortical volume, and marrow volume were higher in WT mice than in kit^W/W-v mice. Differences in cortical thickness were not detected with genotype. Cortical bone volume and cortical thickness were lower and medullary volume was higher in ovx mice compared to ovary-intact mice. Significant genotype by surgery interactions were not detected for any of the cortical endpoints evaluated.

Table 2.

The effect of kit receptor deficiency and ovariectomy on total tibia bone volume and length, cortical bone architecture proximal to the tibio-fibular junction, and cancellous bone volume and architecture in the proximal tibial metaphysis and 3rd lumbar vertebra in 8-week-old mice at 4 weeks following surgery.

	WT		Kit W/W-v		ANOVA (P<)
	Intact	Ovx	Intact	Ovx	Genotype	Surgery	Interaction
Total Tibia
Bone length (mm)	17.10 ± 0.08	17.58 ± 0.10	17.00 ± 0.09	17.37 ± 0.14	NS	0.001	NS
Bone volume (mm3)	14.78 ± 0.42	12.80 ± 0.26	13.23 ± 0.29	12.51 ± 0.29	0.01	0.001	NS
Tibial Diaphysis (cortical bone)
Tissue volume (mm3)	0.154 ± 0.004	0.156 ± 0.002	0.140 ± 0.002	0.138 ± 0.002	0.001	NS	NS
Cortical volume (mm3)	0.118 ± 0.002	0.114 ± 0.002	0.110 ± 0.002	0.102 ± 0.002	0.001	0.01	NS
Marrow volume (mm3)	0.036 ± 0.002	0.042 ± 0.002	0.030 ± 0.002	0.036 ± 0.002	0.001	0.001	NS
Cortical thickness (μm)	260 ± 5	245 ± 2	259 ± 4	235 ± 6	NS	0.001	NS
Proximal Tibial Metaphysis (cancellous bone)
Bone volume/Tissue volume (%)	6.75 ± 0.35	1.99 ± 0.21	6.74 ± 1.28	1.94 ± 0.30	NS	0.001	NS
Connectivity density (1/mm3)	50.62 ± 7.40	6.48 ± 1.29	43.66 ± 15.85	5.05 ± 1.31	NS	0.001	NS
Trabecular number (1/mm)	3.69 ± 0.07	2.87 ± 0.13	3.93± 0.32	2.68 ± 0.10	NS	0.001	NS
Trabecular thickness (μm)	40 ± 1	35 ± 2	37 ± 2	30 ± 2	0.04	0.001	NS
Trabecular spacing (μm)	276 ± 6	366 ± 16	263 ± 18	380 ± 16	NS	0.001	NS
3rd Lumbar Vertebra (cancellous bone)
Bone volume/Tissue volume (%)	12.36 ± 0.74	8.20 ± 0.31	13.68 ± 0.43	8.93 ± 0.42	0.05	0.001	NS
Connectivity density (1/mm3)	156.70 ± 13.90	90.73 ± 5.91	147.63 ± 7.70	90.82 ± 7.01	NS	0.001	NS
Trabecular number (1/mm)	4.47 ± 0.18	3.92 ± 0.06	4.37 ± 0.08	3.68 ± 0.06	NS	0.001	NS
Trabecular thickness (μm)	40 ± 1	36 ± 0	44 ± 1	38 ± 1	0.001	0.001	NS
Trabecular spacing (μm)	228 ± 9	259 ± 4	235 ± 4	277 ± 4	0.04	0.001	NS

Data are mean ± SE

Genotype had minimal effect on cancellous bone architecture in the proximal tibial metaphysis, with only trabecular thickness being higher in the 8-week-old WT compared to kit^W/W-v mice. Ovx resulted in lower bone volume/tissue volume, connectivity density, trabecular number and trabecular thickness, and higher trabecular spacing. Significant genotype by surgery interactions were not detected for any of the cancellous endpoints evaluated.

The effects of genotype and ovx on dynamic histomorphometry in proximal tibial metaphysis are shown in Figure 5. Significant differences between WT and kit^W/W-v mice were not observed. Ovx resulted in greater mineralizing perimeter, mineral apposition rate, and bone formation rate. Significant genotype by surgery interactions were not detected for any of the dynamic endpoints evaluated. Bone marrow adipocytes were common in proximal tibia of WT mice. Bone marrow adiposity was higher in ovx WT mice than in intact WT mice. The increase in marrow adiposity was due predominantly to an increase in adipocyte density. Adipocytes were not detected in either intact or ovx kit^W/W-v mice.

Figure 5.

The effects of genotype (WT vs. kit^W/W-v) and surgery (ovary-intact vs. ovx) on cancellous bone mineralizing perimeter/bone perimeter (M.Pm/B.Pm) (a), mineral apposition rate (MAR) (b), bone formation rate/bone perimeter (BFR/B.Pm) (c), bone marrow adiposity (d), adipocyte density (e), and adipocyte size (f) in the proximal tibial metaphysis and M.Pm/B.Pm (g), MAR (h), and BFR/B.Pm (i) in lumbar vertebra of 8-week-old mice 4 weeks following surgery. Values are mean ± SE. ^aP<0.05 compared to ovary-intact (intact) within genotype.

The effects of genotype and ovx on vertebral cancellous bone architecture and fluorochrome-based indices of bone turnover are shown in Table 2 and Figure 5, respectively. Vertebral cancellous bone volume/tissue volume, trabecular thickness, and trabecular spacing were greater in kit^W/W-v mice than in WT mice (Table 2). Ovx resulted in lower bone volume/tissue volume, connectivity density, trabecular number, trabecular thickness, and greater trabecular spacing. Differences in dynamic bone measurements were not detected with genotype. Mineralizing perimeter and bone formation rate were higher in ovx compared to ovary-intact mice (Figure 5). Differences in mineral apposition rate were not detected with surgery, nor was there a genotype by surgery interaction for any of the endpoints measured. Adipocytes were rare in lumbar vertebrae of ovary-intact WT mice (0.2 ± 0.1 adipocytes/mm²) but more numerous in ovx WT mice (1.7 ± 0.2 adipocytes/mm², P<0.001). Adipocytes were not present in lumbar vertebrae of either ovary-intact or ovx kit^W/W-v mice.

4. DISCUSSION

Ovx in WT mice resulted in increased body mass and total body fat mass and decreased cancellous and cortical bone mass. The observed cortical and cancellous osteopenia was associated with increased bone marrow fat and increased cellular and dynamic indices of bone turnover. Similar ovx-induced changes in bone mass, architecture, and indices of bone turnover following ovx were observed in kit^W/W-v mice in spite of absence of bone marrow adipocytes and an attenuated increase in total body fat mass.

Gonadal hormones, and in particular estrogen, are important physiological regulators of bone growth, maturation, and turnover balance. Changes in the circulating level of estrogen directly modulate the differentiation and activity of osteoblasts, osteoclasts, and chondrocytes [18]. Also, changes in estrogen levels have species-, skeletal compartment-, and age-specific effects on bone architecture and turnover. The normal increase in estrogen levels at puberty accelerates epiphyseal closure and promotes accumulation of cancellous bone at mechanically loaded skeletal sites [18]. Not surprisingly, estrogen acts in consort with anabolic hormones during growth to determine peak bone mass, and with reproductive hormones to regulate mineral availability during pregnancy and lactation [28, 29]. In humans, bone metabolism is disrupted by natural and surgery-induced menopause. The disruption results in accelerated but often unbalanced bone turnover which can lead to rapid local bone loss [18].

Osteoblasts and adipocytes are derived from stromal cell progenitors and it has been inferred that adipocyte differentiation within the bone marrow reduces, or even precludes, concurrent osteoblast differentiation leading to bone loss [7, 10, 30–36]. Our observation that bone marrow adipocyte-deficient kit^W/W-v mice have increased osteoblast perimeter is consistent with the view that low levels of adipogenesis increase osteoblast differentiation. However, results from the current study showing cancellous osteopenia in 14-week-old kit^W/W-v mice provide evidence that absence of bone marrow adipocytes is not sufficient to lead to a net increase in bone mass. In the case of kit^W/W-v mice, increased osteoblast perimeter in the proximal tibia metaphysis of 14-week-old mice was accompanied by increased osteoclast perimeter and net bone loss.

The above observations do not exclude the possibility that factors produced by adipocytes contribute to the negative bone balance observed at some skeletal sites in ovx rodents. Adipocytes produce factors, including TNF α and IL-6, known to inhibit bone formation and increase bone resorption [37]. On the other hand, not all adipokines are antiosteogenic; adipocytes also produce factors such as leptin and IGF-I which can increase bone formation [5, 38–40].

In the present study, ovx resulted in an increase in tibia length but a reduction in overall tibial bone volume. In spite of increases in bone formation, marrow volume increased and cancellous bone volume decreased. It is well established that bone mass and architectural changes following ovx are not uniform throughout the skeleton. Specifically, osteopenia only occurs at sites where increases in bone formation are inadequate to match the increase in bone resorption [41]. For example, ovx results in increased bone turnover in both the proximal metaphysis and epiphysis of the tibia in rats yet rapid bone loss occurs in the metaphysis whereas cancellous bone in the epiphysis is spared. Even in the metaphysis, bone is selectively lost from sites experiencing low strain energy levels [41]. At the periosteum, a skeletal compartment with minimal bone resorption in rodents, bone formation predominates following ovx, at times leading to a net increase in cortical bone mass [42, 43]. Adipokines produced by bone marrow adipocytes could contribute to osteopenia by altering bone growth or the local balance between bone formation and resorption in favor of the latter. We evaluated these possibilities by investigating the skeletal response to ovx in kit^W/W-v mice. Kit^W/W-v mice have no marrow adipocytes but exhibited ovx-induced skeletal changes strikingly similar to those in WT mice. We interpret this finding as strong evidence that an increase in bone marrow adiposity following ovx is not required for the dramatic skeletal responses to ovx.

In addition to an increase in bone marrow adiposity, ovx normally results in increases in body weight and peripheral fat. Rather than through local production, it is possible that the negative skeletal effects of ovx are mediated through adipokines produced in extramedullary fat depots that circulate to skeletal sites. However, ovx kit^W/W-v mice, while exhibiting similar cancellous osteopenia as WT mice, exhibited a smaller long-term increase in total body fat mass. Also, cortical and cancellous osteopenia was well established during the initial 4 weeks following ovx with no corresponding increase in total body fat mass during this time. Thus, this possibility is not supported by our data.

Kit signaling is required for the ovx-induced increase in bone marrow adiposity in mice. In addition to an absence of bone marrow adipocytes, the kit receptor-deficient kit^W/W-v mice have multiple abnormalities in fat metabolism. These abnormalities include hypertriglyceridemia and hypercholesterolemia [44, 45]. Chylomicrons, very low density lipoprotein, and intermediate density lipoprotein are also reported to be elevated. Additionally, the mutant mice have reduced lipoprotein lipase activity. Taken together, these findings suggest that kit signaling plays an important but not necessarily direct role in fat metabolism. Kit is required for differentiation and survival of mast cells and recent studies suggest that mast cells play an important role in obesity and diabetes by regulating adipocyte differentiation and function [46, 47]. Interestingly, receptor tyrosine kinase inhibitors targeting kit signaling have been reported to reduce blood glucose levels in patients with chronic myeloid leukemia and decrease body weight in rodents fed a high fat diet [48, 49].

Stem cell factor (kit ligand), the endogenous ligand for kit receptor, has long been recognized to regulate early hematopoietic linage decision [50]. Kit^W/W-v mice have a point mutation in the tyrosine kinase domain that inactivates the kit receptor in one allele and a mutation in the second allele that results in kit receptor deficiency at the cell surface [51]. As a consequence, activation of kit signaling by kit ligand is greatly reduced in kit receptor-expressing cells in kit^W/W-v mice. The skeletal phenotype of kit^W/W-v mice has not been described in detail. In the present study, growing mice and mice approaching skeletal maturity [52, 53] were evaluated. Our results suggest that kit receptor deficiency results in cortical as well as premature age-related cancellous osteopenia. The lower cortical bone mass in the mutant mice appears to be largely due to a reduction in bone growth. The lower cancellous bone volume in the proximal tibial metaphysis of 14-week-old mice was associated with increases in osteoblast and osteoclast perimeters, suggesting that kit signaling in mice also functions to regulate bone balance in adults.

The skeletal abnormalities observed in kit^W/W-v mice are similar to those reported in kit^Sl/Sl-d mice [52]. Kit^Sl/Sl-d mice produce a protein similar to secreted kit ligand but are unable to produce membrane-associated kit ligand. Secreted kit ligand binds to and activates kit receptor but is unable to duplicate all of the actions of the membrane-associated ligand [51, 54, 55]. As a consequence, kit^Sl/Sl-d mice are sterile, anemic, osteopenic, have elevated osteoclast perimeter and have no marrow adipocytes [24]. Thus, activation of kit signaling through kit receptor by membrane-associated kit ligand appears necessary for differentiation of a variety of marrow-derived cells, including adipocytes.

Kit^W/W-v mice are also mast cell-deficient [44]. Mast cells produce an impressive variety of factors capable of influencing bone metabolism, including histamine, TNF-α, serotonin, and platelet-derived growth factor alpha [56, 57]. Mastocytosis is associated with abnormal bone metabolism, decreased bone mineral density, and increased fracture risk [58–62]. Ovx results in increased mast cell density in bone marrow of rats, leading to speculation that mast cells may contribute to the etiology of osteoporosis [63]. The present study, demonstrating comparable changes in WT and kit^W/W-v mice following ovx, does not support an important role for mast cells in mediating the skeletal response to ovx in mice. However, it should be noted that, in contrast to humans and rats, mast cells are rarely present in bone marrow of mice. Furthermore, there is evidence that the presence of mast cells within bone marrow contributes to species differences in bone metabolism [64].

5. CONCLUSIONS

In summary, we replicated the previously reported inverse relationship between changes in bone mass and bone marrow adiposity in WT mice following ovx. Specifically, cortical and cancellous bone mass was decreased in ovx compared to ovary-intact WT mice while bone marrow adiposity was increased. We confirmed the absence of adipocytes in bone marrow of kit^W/W-v mice and showed that marrow adipocytes were not induced by ovx in the kit-deficient animals. Ovx resulted in increases in bone turnover and statistically indistinguishable changes in bone architecture in WT and bone marrow adipocyte-deficient kit^W/W-v mice. These findings suggest that bone marrow adipose tissue is not required for development of cortical or cancellous osteopenia following ovx in growing mice.

*HIGHLIGHTS.

• Compared to WT mice, kit receptor-deficient kit^W/W-v mice had lower body fat and lacked bone marrow adipocytes.

• Adipocytes were present in bone marrow of WT mice and bone marrow adiposity increased following ovariectomy.

• Adipocytes were not detected in ovariectomized kit^W/W-v mice.

• Significant differences in bone response to ovariectomy were not detected between WT and kit^W/W-v mice.

• Our findings do not support a causal role for increased bone marrow fat as a mediator of ovariectomy-induced osteopenia in mice.