Hormone-sensitive lipase-knockout mice maintain high bone density during aging

Abstract

We tested the hypothesis that the actions of hormone-sensitive lipase (HSL) affect the microenvironment of the bone marrow and that removal of HSL function by gene deletion maintains high bone mass in aging mice. We compared littermate control wild-type (WT) and HSL^−/− mice during aging for changes in serum biochemical values, trabecular bone density using micro-computed tomography, bone histomorphometry, and characteristics of primary bone marrow cells and preosteoblasts. There is a regulated expression of HSL and genes involved in lipid metabolism in the bone marrow during aging. HSL^−/− mice have increased serum levels of insulin and osteocalcin with decreased leptin levels. Compared with the marked adipocyte infiltration in WT bone marrow (65% by area) at 14 mo, HSL^−/− mice have fewer (16%, P<0.05) and smaller adipocytes in bone marrow. While peak bone density is similar, HSL^−/− mice maintain a higher bone density (bone volume/total volume 6.1%) with age than WT mice (2.6%, P<0.05). Primary osteoblasts from HSL^−/− mice show increased growth rates and higher osteogenic potential, manifested by increased expression of Runx2 (3.5-fold, P<0.05) and osteocalcin (4-fold, P<0.05). The absence of HSL directs cells within the bone marrow toward osteoblast differentiation and favors the maintenance of bone density with aging.—Shen, W.-J., Liu, L.-F., Patel, S., Kraemer, F. B. Hormone-sensitive lipase-knockout mice maintain high bone density during aging.

In addition to hematopoietic stem cells that are capable of differentiating into blood cell lineages, the bone marrow contains mesenchymal stem cells that can differentiate into connective tissues, including cartilage, adipocytes, osteoblasts, and fibroblasts (1–3). Many studies have shown that there is an inverse relationship between the differentiation of mesenchymal stem cells along either an adipocyte or osteoblast lineage (3). At the same time, there is apparently plasticity between adipocytes and osteoblasts, such that mesenchymal stem cells differentiated or undergoing differentiation into adipocytes can display a capacity to convert to an osteoblast phenotype under particular conditions (4–5).

Although a complete understanding of all of the factors that determine osteoblastogenesis is still lacking, many factors secreted by adipose tissue have been shown to influence osteogenesis (6–7). Coculture of mature adipocytes has been shown to inhibit the proliferation of osteoblasts (8). Free fatty acids (FFAs) are the predominant products released from adipocytes due to the hydrolysis of stored triacylglycerol (TAG) through the action of lipolysis. In addition to being metabolized as an energy source, FAs and their metabolites can affect gene expression through their actions as ligands for peroxisome proliferator-activated receptors (PPARs) and G-protein-coupled receptors (9–11). Activation of PPARγ promotes adipogenesis and inhibits osteoblastogenesis (12–13), whereas haploinsufficiency of PPARγ promotes osteoblastogenesis (14). FAs and their metabolites can also act as second messengers in the transduction of external signals to many biochemical pathways; hence, they exert an effect on directing mesenchymal stem cell differentiation toward osteoblastic or adipogenic pathways (15–16). In addition to lipid and steroid mediators, many adipokines secreted by adipose tissue can influence osteoblastogenesis; among these are leptin, adiponectin, TNF-α, IL-6, adipsin, complement factor 3, and growth factors (7, 17).

The release of FAs from adipose tissue is mediated through the coordination of several droplet-associated proteins, such as perilipin, and intracellular lipases, such as adipose triglyceride lipase (ATGL) and hormone-sensitive lipase (HSL) (18–20). HSL activity is regulated by reversible phosphorylation, and the enzyme displays broad substrate specificity, being able to hydrolyze cholesteryl esters, TAG, diacylglycerol (DAG), and 1- or 3-monoacylglycerol, as well as other lipid substrates, such as retinyl esters (21–22). The activity against diacylglycerol is ∼10-fold and 5-fold higher than against TAG and monoacylglycerol. Inactivation of HSL results in the complete absence of neutral cholesteryl ester hydrolase (NCEH) activity in both white adipose tissue (WAT) and brown adipose tissue (BAT); however, TAG lipase activity in WAT was reduced by 60% and remains similar to wild-type (WT) mice in BAT. There is complete absence of stimulated glycerol, but not FA, release in HSL^−/− mice, and a marked increase in DAG accumulation in adipose tissue. HSL^−/− mice are resistant to a high-fat diet-induced obesity with alterations in adipose-specific genes and adipokine production. Here, we report changes observed in bone density in HSL^−/− mice compared with their WT littermates. Further studies show that bone marrow mesenchymal stem cells from HSL^−/− mice have higher growth rates with increased potential for osteoblastogenesis.

MATERIALS AND METHODS

Animals

HSL^−/− mice were generated by homologous recombination, as described previously (23), and backcrossed 5 times with C57/BL6J mice. Mice were maintained in the animal facility at the Veterans Affairs (VA) Palo Alto Health Care System on a 12-h light-dark cycle. All procedures were in accordance with institution guidelines and approved by the institutional animal care and use committee of VA Palo Alto Health Care System. For breeding experiments, mice heterozygous for the deleted HSL allele were used to generate homozygous HSL^−/− mice and HSL^+/+ (WT) littermates. Genotyping was performed by a single-step PCR using three primers, as described previously (23). Only female mice were used in these studies.

Serum biochemical analysis

Serum obtained from mice was analyzed for triglyceride, FFAs, and glucose using kits from Stanbio (Boerne, TX, USA). Insulin, osteocalcin, leptin, and adiponectin were measured using mouse adipokine and bone panel multiplex assays (Millipore, Bedford, MA, USA) and detected by the Luminex xMAP method (Luminex 200; Millipore). Insulin, adiponectin, leptin, and osteocalcin were measured in 6-mo and 14-mo-old mice in the fed state.

Micro-computed tomography (micro-CT) analysis of trabecular bone density

Female WT and HSL^−/− mice were sedated using isoflurane. The femurs of WT and HSL^−/− animals were measured by micro-CT, using a Scanco vivaCT 40 μCT scanner (Scanco Medical AG, Basserdorf, Switzerland). Scanning was started from the point where the growth plate bridge across the middle of the metaphysis ends. A total of 160 consecutive 10-μm-thick sections were scanned, and 100 sections were analyzed, encompassing a 1.0-mm length of the secondary spongiosa. Cortical bone was excluded from the region of interest with semiautomatically drawn contours. The segmentation values were kept constant at 1.0/1/276. Relative bone volume/total volume (BV/TV), trabecular number (TbN), trabecular thickness (TbTh), and trabecular separation (TbSp) were calculated by measuring 3-dimensional distances directly in the trabecular network and taking the mean over all voxels.

Bone histomorphometry

All animals were subcutaneously injected with calcein (10 mg/kg) at 12 and 5 d before necropsy. The right proximal femurs were dehydrated in ethanol, embedded undecalcified in methylmethacrylate, and sectioned longitudinally with a Leica/Jung 2255 microtome (Leica Microsystems, Buffalo Grove, IL, USA) at 4- and 8-μm-thick sections. The 4-μm sections were stained with hematoxylin and eosin (H&E) for collection of bone mass and architecture data by light microscopy, whereas the 8-μm sections were left unstained for measurements of fluorochrome-based indices. Static and dynamic histomorphometry were performed using a semiautomatic BioQuant imaging analysis system (R&M Biometrics, Nashville, TN, USA) linked to a microscope equipped with transmitted and fluorescence light. A counting window, measuring 8 mm² and containing only trabecular bone and bone marrow, was created for the histomorphometric analysis. Static measurements included total tissue area (TAr), bone area (BAr), and bone perimeter (BPm). Dynamic measurements included single-labeled perimeter (sLPm) and double-labeled perimeter (dLPm) and interlabel width (IrLWi). These indices were used to calculate BV/TV, mineralizing surface/bone surface (MS/BS), and mineral apposition rate (MAR). Finally, surface-based bone formation rates [bone formation rate/bone surface (BFR/BS)] were calculated by multiplying mineralizing surface (single-labeled surface/2 + double-labeled surface) with MAR, according to Parfitt et al. (24). Additional histological sections were stained with Goldner trichrome for visualization and counting of osteoblasts and osteoclasts by BioQuant Image Analysis Corp. (Nashville, TN, USA). The numbers of osteoblasts and osteoclasts were normalized per millimeter of bone perimeter.

Isolation, culture, and differentiation of primary bone marrow cells

Femurs, tibiae, and humeri from 4- to 5-mo-old, female WT and HSL^−/− mice were dissected and rinsed with 75% ethanol and diethylpyrocarbonate water to eliminate any surrounding soft tissue. Bone marrows were flushed out with PBS containing 1% FA-free BSA using a 25-gauge needle and were pooled and treated with red blood cell lysis buffer (Sigma, St. Louis, MO, USA) for 5 min before washing 3 times with PBS with 1% FA-free BSA. For purification of bone marrow adipocytes, the cells were then centrifuged at 3000 rpm for 5 min, and the floating adipocytes were collected and washed another 2–3 times with PBS with 1% FA-free BSA. Adipose cell size was determined by measuring the diameter of 100 adipose cells/animal by microscopy of paraformaldehide fixed cells. The bone marrow stromal cells (BMSCs) remaining after adipocyte isolation were then collected. BMSCs were plated in α-minimum essential medium (α-MEM) with 10% FBS and grown until confluent before being trypsinized and plated in 96 wells for assay of cell growth rate or in a 24-well plate for studies of differentiation. Osteoblastogenesis was induced with growth medium containing 10 mM β-glycerophosphate and 50 mM ascorbic acid for up to 14 d with change of medium every 3 d. Adipogenesis was induced with insulin, dexamethasone, and isobutylmethylxanthene, as described previously (25). For assay of cell growth rates, secondary cultures of BMSCs from 5-mo-old WT and HSL^−/− mice were seeded at 2000 cells/well in 96-well plates. Medium was changed every other day. Alamar blue dye (AbD Serotec, Raleigh, NC, USA), which is an indicator of the cellular redox activity, was used to measure cell numbers each day, starting at 2 d after the seeding. Cells were labeled for 8 h, and the increase in the redox activity, as indicated by the increase in fluorescence emitted at 590 nm, was read.

Isolation of preosteoblasts

Calvariae from WT and HSL^−/− mice of indicated ages were dissected, cleaned, and rinsed with PBS buffer before cutting into small pieces and digesting with collagenase type I (1 mg/ml) at 37°C with rotation at 150 rpm for 1 h. Cells were then passed through a cell strainer (70 μm) before washing 3 times with PBS with 1% FA-free BSA and plated in α-MEM with 10% FBS and grown until confluent before harvesting.

Alkaline phosphatase (ALP) assay

Secondary cultures of BMSCs from 4-mo-old mice were plated onto 24-well plates and treated with 0.3-mM ascorbic acid phosphate and 10 mM β-glycerophosphate starting at d 2 until confluence. The cells were removed from the wells, sonicated in 10 mM Tris/0.1% Triton X-100 (pH 7.0), and assayed for ALP using p-nitrophenolphosphate as the substrate. Protein concentration was determined by the bicichoninic acid (BCA) method. Data are expressed as millimoles of p-nitrophenol per minute produced at 37°C normalized by the protein content.

RNA isolation and quantitative real-time PCR analysis

For RNA isolation, bone marrow samples were homogenized in 1 ml of TRIzol reagent using a power homogenizer (Ultra-Turrax T25; Labortechnik, Göttingen, Germany), and total RNA was isolated using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). After the ethanol precipitation step, total RNA isolated from bone marrow adipocytes was dissolved in 30 μl RNase-free water and reamplified to aRNA and then converted to cDNA for real-time PCR analysis. Real-time PCR was performed with the cDNA prepared as above using an ABI Prism 8500 System using SYBR green master mix reagent (Applied Biosystems, Foster City, CA, USA), and the primer pairs used were as described previously (26–27). The relative mass of specific RNA was calculated by the comparative cycle of threshold detection method, according to the manufacturer's instructions. Three independent sets of real-time PCR were performed using different RNA preparations from bone marrow samples. Genes examined included HSL, perilipin (Plin1), adipocyte differentiation-related protein (Plin2), TIP47 (Plin3), FSP27, S3–12 (Plin4), ATGL, CGI-58, fatty acid binding protein 4 (FABP4), PPARγ, adiponectin, leptin, Runx2, osteocalcin, osterix, and ALP.

Statistics

Data are expressed as means ± se. Statistical analyses were performed by 2-way ANOVA or by Mann-Whitney U tests for nonparametric data using Prism 5 for Mac OS X (GraphPad Software, La Jolla, CA, USA). Differences between groups were considered statistically significant at values of P < 0.05.

RESULTS

BMSCs express HSL and genes involved in lipid metabolism in a regulated pattern during aging

To investigate whether HSL and other genes involved in lipid metabolism are expressed in bone marrow, RNA and protein were extracted from bone marrow pooled from femurs, tibiae, and humeri. As shown in Fig. 1A (top), the protein band that is immunoreactive with anti-HSL antibody was detected in the bone marrow at a level ∼60 times less than that of adipose tissue. Measurement of HSL by RT-PCR showed that bone marrow cells express 62.5-fold less HSL mRNA than gonadal fat. Further examination by RT-PCR showed that bone marrow also expresses many lipid droplet-associated proteins, including Plin1, Plin2, Plin3, Plin4 and FSP27, as well as ATGL and CGI-58 (data not shown). Since adipocytes constitute a small percentage of the cells in bone marrow of young animals, we isolated bone marrow adipocytes from WT animals at 6 and 14 mo of age and analyzed the expression of genes encoding lipid droplet-associated proteins, as well as other genes related to lipid metabolism. As shown in Fig. 1B, the expression levels of HSL, Plin1, Plin2, FABP4, PPARγ, and leptin were significantly increased in 14-mo-old compared to 6-mo-old WT mice. The changes in expression of genes involved in osteogenesis were also examined in BMSCs during aging (Fig. 1C); Runx2, ALP, and osteocalcin showed significantly lower expression levels at 14 mo when compared with 6-mo-old animals. HSL^−/− mice have been shown to have less adiposity and decreased adipose expression of many genes involved in lipid metabolism. Figure 1D compares the expression levels of genes in WT and HSL^−/− bone marrow adipocytes at 14 mo. Similar to previous observations of lower expression of several adipose-specific genes in peripheral adipose tissue of HSL^−/− mice (27), expression of Plin1, FABP4, and leptin were significantly lower in bone marrow adipocytes from HSL^−/− than from WT mice, but adiponectin was significantly higher. The size distribution of adipocytes isolated from bone marrow increased with age in both WT and HSL^−/− mice (Fig. 1E, F); however, there was a greater population of small cells observed in HSL^−/− than in WT mice at both 6 mo (Fig. 1E) and 14 mo (Fig. 1F) of age.

Figure 1.

Bone marrow cells express HSL and genes involved in lipid metabolism in a regulated pattern during aging. Bone marrow from tibiae, femurs, and humeri of 5-mo-old female C57/BL6J mice were pooled and analyzed. A) Western blot analysis of HSL protein. Total cell extracts were prepared by sonication of the cells; 10 μg total protein from perigonadal adipose tissue (lane 1) and 25 μg total protein from bone marrow cells (lane 2) were separated on SDS-PAGE, and subjected to immunoblot analysis using anti-HSL antibody and anti-β-actin for loading control. B, C) RT-PCR analysis of expression of genes involved in lipid metabolism in bone marrow adipocytes isolated from WT mice during aging (B) and genes involved in osteogenesis (C) in BMSCs isolated from WT mice during aging. Data are expressed as means ± sd of 3 independent RNA isolations and RT-PCR experiments. D) RT-PCR analysis of expression of genes involved in lipid metabolism in bone marrow adipocytes isolated from WT and HSL^−/− (KO) mice at 14 mo of age. RNA was reamplified and reverse transcribed for RT-PCR analysis of gene expression. Data are means ± se of 2 independent RNA isolations and 2 independent reamplifications and RT-PCR experiments. E, F) Size distribution of adipocytes isolated from bone marrow from WT and HSL^−/− mice at 6 mo (E) and 14 mo (F) of age. OCN, osteocalcin. *P < 0.05; **P < 0.01.

HSL^−/− mice have higher serum insulin, osteocalcin, and lower leptin levels during aging

Changes in serum biochemical values are shown in Table 1. Serum glucose, triglyceride, and cholesterol concentrations were generally similar in WT and HSL^−/− mice at all ages studied, though serum cholesterol was significantly lower in HSL^−/− mice at 4 mo of age, and serum triglycerides were significantly lower in HSL^−/− mice at 14 mo of age. Circulating FFA concentrations were lower in HSL^−/− mice at 4 and 6–8 mo of age, but similar to WT at 14 mo. Serum levels of insulin, adiponectin, leptin, and osteocalcin are shown in Fig. 2. Serum insulin levels did not change with age in WT mice, whereas there was a trend for insulin levels to increase with age in HSL^−/− mice, with significantly higher insulin levels in HSL^−/− mice compared to WT mice at 14 mo. Serum adiponectin increased in WT mice until 4–6 mo and then declined in older animals. In contrast, adiponectin levels were significantly lower in HSL^−/− mice at 4–8 mo of age. Serum leptin increased with age in WT mice but was significantly lower in the HSL^−/− mice at 8 and 14 mo. Serum osteocalcin levels were highest in 2-mo-old mice and decreased with age in both WT and HSL^−/− mice; however, osteocalcin was significantly higher in HSL^−/− mice from 4 to 14 mo of age.

Table 1.

Serum biochemical values

Component	4 mo		6–8 mo		14 mo
	WT	HSL−/−	WT	HSL−/−	WT	HSL−/−
FFAs (mM)	1.7 ± 0.4	1.3 ± 0.2*	1.2 ± 0.3	1.0 ± 0.2*	1.5 ± 0.8	1.4 ± 0.5
Glucose (mg/dl)	115 ± 23	119 ± 22	104 ± 18	115 ± 8*	112 ± 27	83 ± 12
TG (mg/dl)	117 ± 25	94 ± 28	68 ± 18	62 ± 21	83 ± 9	46 ± 17*
TC (mg/dl)	125 ± 23	71 ± 11*	119 ± 41	116 ± 40	105 ± 32	107 ± 18

Serum from female mice (n=3–7) of the indicated ages was collected and assayed. TG, triglyceride; TC, total cholesterol.

^*P < 0.05 vs. age-matched WT.

Figure 2.

HSL^−/− mice have higher serum insulin and osteocalcin and lower leptin levels during aging. Serum samples from animals of various ages were collected and assayed for various analytes. A) Insulin. B) Adiponectin. C) Leptin. D) Osteocalcin. KO, HSL^−/−. Results are means ± sd of 3–7 animals/group. *P < 0.05 vs. age-matched WT.

HSL^−/− mice maintain higher bone density and less adipocyte infiltration during aging

In view of the decreased adipokines and increased osteocalcin in HSL^−/− mice, we measured the trabecular bone density of the femur using micro-CT scanning technology. As shown in Fig. 3A, there was no difference in peak bone density between WT and HSL^−/− mice, but after the peak (which occurred at ∼10–14 wk of age), bone density decreased with age in both WT and HSL^−/− animals; however, the rate of decline was different in WT and HSL^−/− mice. The curve fit shows that the bone density of WT mice decreased at a faster rate compared to HSL^−/− mice. The difference in bone mass is clearly revealed in Fig. 3B, where the reconstructed micro-CT images of trabecular bone clearly illustrate that HSL^−/− mice start to show slightly higher density than WT mice at 4 and 6 mo of age, and the differences in bone mass are obvious when the animals reach 14 mo of age.

Figure 3.

Micro-CT analysis of trabecular bone density. Right femurs of female WT and HSL^−/− mice of various ages were subjected to micro-CT analysis. A) Bone density, as represented by BV/TV, is plotted against the age of the animals; curve fit was obtained using the polynomial curve fitting function in the Prism statistical analysis package. B) Reconstruction of micro-CT images of the distal femur of WT and HSL^−/− (KO) mice at 4, 6, and 14 mo of age.

The difference in bone volume per tissue volume of WT and HSL^−/− mice observed by micro-CT was also confirmed when bone histomorphometry analysis was performed. For this purpose, animals were labeled with calcein in vivo at d 12 and 5 before the animals were sacrificed. Femurs were cleaned, dehydrated, and embedded in plastic to allow for longitudinal section of the distal femur for bone histomorphometry analysis and quantification of bone formation during the two labeling times. Table 2 summarizes the histomorphometry analysis. As the animals aged, the TbTh, TbN, BV/TV, MS/BS, and BFR/BS showed a trend to decrease, while TbS tended to increase in WT animals. Similar changes were observed in HSL^−/− mice, but to a lesser extent than in WT, such that TbN, BV/TV, and BFR/BS were significantly greater in HSL^−/− mice than WT mice at 14 mo (P<0.05). In addition, Goldner trichrome staining was performed to assess the number of osteoblasts and osteoclasts in the same section of bones from 4- and 14-mo-old WT and HSL^−/− mice (Table 2). No differences between WT and HSL^−/− mice in either osteoblast or osteoclast numbers per millimeter of bone perimeter were noted at 4 mo of age. However, there were significantly more osteoblasts per millimeter bone perimeter in 14-mo-old HSL^−/− mice (P<0.05) owing to a maintenance of osteoblast numbers, as opposed to the decline observed in WT mice. No differences in osteoclast numbers were observed at 14 mo.

Table 2.

Bone histomorphometry analyses

Parameter	4 mo		6–8 mo		14 mo
	WT	HSL−/−	WT	HSL−/−	WT	HSL−/−
TbTh (μm)	46.2 ± 9.5	67.6 ± 17.4	37.6 ± 9.9	47.6 ± 9.4	47.3 ± 5.0	57.7 ± 6.8
TbN (mm−)	1.2 ± 0.4	1.9 ± 1.0	1.1 ± 0.3	1.8 ± 0.6*	0.6 ± 0.1	1.0 ± 0.1*
TbSp (μm)	804.7 ± 225.6	581.3 ± 323.9*	959.6 ± 265.2	556.9 ± 165.9*	1738.3 ± 258.4	937.8 ± 51.7*
BV/TV (%)	5.7 ± 3.4	8.2 ± 1.0*	3.9 ± 1.2	8.6 ± 3.9*	2.6 ± 0.1	6.1 ± 0.8*
MAR (μm/d)	2.2 ± 0.4	2.0 ± 0.2	2.2 ± 0.1	2.2 ± 0.2	1.5 ± 0.4	2.9 ± 0.1*
MS/BS (%)	27.2 ± 10.4	24.2 ± 5.9	18.1 ± 5.7	28.9 ± 5.6*	13.6 ± 0.5	17.7 ± 2.0
BFR/BS (μm/d)	0.5 ± 0.2	0.5 ± 0.1	0.4 ± 0.1	0.6 ± 0.2	0.2 ± 0.1	0.5 ± 0.1*
N.Ob/B.pm (n/mm)	15.9 ± 0.8	15.8 ± 0.7	ND	ND	13.3 ± 1.4	15.6 ± 1.1*
N.Oc/B.pm (n/mm)	1.3 ± 0.1	1.1 ± 0.1	ND	ND	1.6 ± 0.5	1.3 ± 0.2

WT and HSL^−/− mice at the indicated ages were injected with calcein at 12 and 5 d before sacrifice. Right distal femur from individual mice was used to prepare histological slides for analysis using the Bioquant imaging system. Data presented are means ± sd from 3–5 animals/group. TbTh, trabecular thickness; TbN, trabecular number; TbSp, trabecular separation; BV/TV, bone volume/total volume; MAR, mineral apposition rate; MS/BS, mineralization surface/bone surface; BFR/BS, bone formation rate/bone surface; N.Ob/B.pm, number of osteoblasts per bone perimeter; N.Oc/B.pm, number of osteoclasts per bone perimeter; ND, not determined.

^*P < 0.05 vs. age-matched WT.

The architecture of the distal femur was examined after H&E staining of longitudinal sections. As shown in Fig. 4, there was very little adipocyte infiltration in the bones of 4- or 6-mo-old WT (∼8 and 9%, respectively) and HSL^−/− (∼7 and 8%, respectively) mice. However, by age 14 mo, adipocytes constituted ∼65% of the bone area in WT mice, yet there was only ∼16% adipocyte infiltration in the bone of HSL^−/− mice (P<0.05).

Figure 4.

Bone histomorphometry analysis. H&E staining of longitudinal sections of distal femurs from WT and HSL^−/− (KO) mice at 4, 6, and 14 mo of age.

BMSCs from HSL^−/− mice have increased growth rate and higher potential for osteoblastogenesis

In order to begin to address potential mechanisms underlying the differences in bone formation rates and bone density in older HSL^−/− mice vs. WT mice, we isolated primary BMSC from 5-mo-old animals and examined the proliferation of the cells in tissue culture. As shown in Fig. 5A, BMSC from HSL^−/− mice have an increased growth rate in the first 3 d in culture before plateauing, whereas cells from WT animals grow slower and reach a plateau at d 5. BMSCs have the potential to differentiate into adipocytes as well as osteoblasts. To examine whether the ablation of HSL changes the potential of the stromal cells for differentiation, we assessed the ability of BMSCs from WT and HSL^−/− mice to differentiate into osteoblasts, using ALP activity as a measure of osteoblast differentiation. Secondary cultures of BMSCs from 5-mo-old mice were grown in the presence of ascorbic acid and glycerophosphate until confluent, and ALP activity was assayed in the cell lysates. As shown in Fig. 5B, cells from HSL^−/− mice had almost twice the ALP activity as cells from WT mice (P<0.01).

Figure 5.

Characterization of bone marrow mesenchymal stromal cells. A) Cell proliferation assay. Primary bone marrow mesenchymal stromal cells were isolated from 5-mo-old female WT and HSL^−/− mice. Cells were plated, and proliferation assay was performed using Alamar blue. B) BMSCs were induced to differentiate with β-glycerophosphate and ascorbate for 14 d. Cells were harvested, and ALP activity was assayed.

Runx2 and osteocalcin gene expression are elevated in primary preosteoblasts from HSL^−/− mice

Expression of genes involved in osteogenesis was evaluated in primary preosteoblasts isolated from calveriae of WT and HSL^−/− mice by RT-PCR. As shown in Fig. 6, levels of expression of Runx2 and osteocalcin were higher, whereas ALP and osterix expression were similar in HSL^−/− preosteoblasts at both 6 and 14 mo compared with WT preosteoblasts. The expression levels of Runx2 and osteocalcin were significantly lower in preosteoblasts isolated from 14-mo-old as compared with 6-mo-old WT mice; however, the expression levels of these genes were similar in preosteoblasts from 6- and 14-mo-old HSL^−/− mice.

Figure 6.

Characterization of primary preosteoblasts from WT and HSL^−/− mice. Relative level of RNA from primary preosteoblasts isolated from calveriae of 6- and 14-mo-old WT and HSL^−/− (KO) mice. Primary preosteoblasts were isolated from calveriae of WT and HSL^−/− mice of 6 and 14 mo of age; 3–5 mice/group. Total RNA was prepared and used for RT-PCR analysis. Data are means ± sd of the summary of 4 independent experiments. OCN, osteocalcin. *P < 0.05 vs. WT; ^†P < 0.05 vs. 6-mo-old groups.

DISCUSSION

Bone remodeling is a lifelong process and is closely associated with whole-body physiological conditions. Recent studies have revealed a close link between bone remodeling and metabolic homeostasis through a novel endocrine loop consisting of leptin, osteocalcin, and insulin (28–30). Apart from the well-demonstrated role of leptin in regulating appetite, energy expenditure, and insulin secretion (31–33), together with osteocalcin, leptin has been shown to influence bone remodeling through hypothalamic modulation of the sympathetic nervous system (34–36). At the same time, osteocalcin has been demonstrated to be a skeletal mediator of energy balance by promoting insulin secretion and increasing adipocyte insulin sensitivity (37–38). Meanwhile, insulin signaling has also been shown to promote the activation of osteoblasts and enhance production of osteocalcin in a feed-forward loop (30). Therefore, the three circulating hormones, leptin, insulin, and osteocalcin, form a novel endocrine loop linking pancreatic insulin secretion with adipocyte insulin sensitivity and skeletal remodeling, working coordinately to orchestrate whole-body energy balance. During aging, these processes undergo changes affecting many physiological pathways. In our studies, we demonstrated an increase in circulating leptin together with decreased osteocalcin during aging in WT mice. When HSL was ablated, there was an alteration in this pattern with aging, as leptin remained low, probably reflecting the reduced adiposity observed in HSL^−/− mice (39), and osteocalcin remained at high levels seen in young animals. These hormonal changes might well contribute to alterations in whole-body metabolism that have been observed previously in HSL^−/− mice (40), as well as to changes in bone remodeling as observed in the current studies.

Mesenchymal stem cells within bone marrow have the potential to differentiate into multiple cell types, including chondrocytes, adipocytes, and osteoblasts (1–2). Many studies have documented a strong inverse relationship between the differentiation of mesenchymal stem cells along adipocyte or osteoblast lineages (2, 41–42). There is also a very strong inverse relationship between bone mass and the quantity of adipose tissue found within the bone marrow with aging (43–44). Although all of the factors involved in determining the differentiation toward the various cell lineages are unknown, several studies have documented that adipocytes secrete a number of factors (17, 45), such as FFAs, interferons, prostaglandins (46–47) leptin (34), and adiponectin (48), as well as steroids, that can influence osteoblast differentiation (45). Several transcription factors, such as PPARγ, FosB, TAZ, Esr1, Msx2, C/EBPβ, and Id4, are known to be important for the determination/specification of the commitment of mesenchymal stem cells to osteoblast or adipocyte lineages (42, 49, 50). The expression of genes involved in adipogenesis, such as PPARγ2 and FABP4, have been shown to be increased, while genes involved in osteogenesis, such as Dlx5 and osteocalcin, are decreased in bone marrow cells of 20-mo-old, as compared to 6-mo-old mice (44). In the current work, we have shown that these changes occur earlier in the aging process in normal mice, as our data demonstrate that the expression of several genes involved in adipogenesis and lipid metabolism, such as PPARγ, HSL, Plin1, Plin2, and FABP4, as well as leptin, is increased in bone marrow adipocytes of 14-mo-old mice, while genes involved in osteogenesis, such as Runx2, osteocalcin, and ALP, are decreased in BMSCs at 14-mo-old as compared to 6-mo-old mice. These changes reflect the increased infiltration of adipocytes in bone marrow and a reduction in bone mass that occur in normal mice with age.

HSL is an intracellular lipase that is highly expressed in adipose tissue and is a major regulator of lipid homeostasis, controlling the hydrolysis of TAG and DAG in response to hormone stimulation (21, 51). When HSL is ablated, there is a significant decrease in adiposity, together with a decrease in the expression of adipose-specific genes (27), reflecting a disturbance in whole-body lipid homeostasis. Circulating levels of FFAs and adipocyte-derived hormones, such as adiponectin and leptin, are reduced in HSL^−/− mice compared with WT mice of the same age. In contrast, circulating osteocalcin levels are increased in HSL^−/− mice. Therefore, the hormonal and adipocytokine changes in HSL^−/− mice appear to produce an environment that favors osteogenesis. Indeed, HSL^−/− mice maintain higher bone density during aging as compared with WT mice, and this is reflected in greater bone volume, greater trabecular numbers, reduced trabecular separation, and higher rates of bone formation, as well as a greater number of osteoblasts. Furthermore, there is considerably less adipocyte infiltration in bone marrow with age in HSL^−/− mice, and the adipocytes found within the bone marrow are smaller than in WT mice. In recent studies examining gene expression profiles of bone marrow adipocytes during aging, we observed that genes involved in adipogenesis and lipolysis are up-regulated with aging (unpublished results), suggesting that bone marrow adipocytes are metabolically active. Metabolically active bone marrow adipocytes are likely to carry out hydrolysis of stored TAG, providing FAs, which drive preadipocytes to differentiate into lipid-filled adipocytes, resulting in an increased population of mature adipocytes in a feed-forward system. This scenario would be consistent with the increase in adipogenesis observed in the bone marrow with age and with the inverse relationship between the presence of bone marrow adipocytes and bone density with aging (43). However, this cycle appears to be attenuated by the ablation of HSL, since HSL^−/− mice display less adipocyte infiltration, as well as smaller adipocytes, within bone with aging, along with an increase in bone density. Although the changes in expression of HSL during aging in WT bone marrow adipocytes are relatively small (∼20% increase from 6 to 14 mo), the cumulative effect of disturbed lipid metabolism due to the ablation of HSL results in a dramatic difference in skeletal homeostasis as the animals age.

PPARγ is among many of the transcription factors that are involved in the control of the commitment of mesenchymal stem cells toward adipogenesis or osteogenesis, and mice with haploinsufficiency of PPARγ have increased bone density (14). We recently reported that adipocytes can modulate key metabolic functions of osteoblasts through the release of secretory products and that PPARγ plays a key role in mediating the effects of adipocytes on osteoblasts (52). In addition, we have reported that one of the mechanisms by which HSL modulates adipose metabolism is by providing intrinsic ligands or proligands for PPARγ and that this is defective in HSL^−/− cells (39). The expression of PPARγ is decreased in HSL^−/− mice, whereas genes associated with osteogenesis, such as Runx2 and osteocalcin, are increased in primary mouse preosteoblasts isolated from HSL^−/− mice, suggesting that HSL can modulate osteoblast function. Whether HSL is affecting osteogenesis in these cells indirectly through PPARγ remains to be further investigated. Aging is known to be associated with decreased bone mass, which is consistent with our demonstration that Runx2 and osteocalcin gene expression are decreased in WT preosteoblasts during aging. Interestingly, when HSL was ablated, not only was there high expression of Runx2 and osteocalcin in primary preosteoblasts, but the expression levels of these genes in primary preosteoblasts did not decline with age. In addition, BMSC from HSL^−/− mice showed an increased growth potential and increased potential for undergoing osteoblast differentiation. Therefore, taken together, our data suggest that the absence of HSL appears to direct cells within the bone marrow toward osteoblast differentiation and favors the maintenance of bone density with aging.