• Bcl-2;
  • osteoblast;
  • transgenic mouse;
  • apoptosis;
  • gender;
  • sex


  1. Top of page
  2. Abstract
  7. Acknowledgements

Transgenic mice (Col2.3Bcl-2) with osteoblast-targeted human Bcl-2 expression were established. Phenotypically, these mice were smaller than their wildtype littermates and showed differential effects of the transgene on bone parameters and osteoblast activity dependent on sex. The net effect was an abrogation of sex differences normally observed in wildtype mice and an inhibition of bone loss with age. Ex vivo osteoblast cultures showed that the transgene had no effect on osteoblast proliferation, but decreased bone formation. Estrogen was shown to stimulate endogenous Bcl-2 message levels. These studies suggest a link between Bcl-2 and sex regulation of bone development and age-related bone loss.

Introduction: Whereas Bcl-2 has been shown to be an important regulator of apoptosis in development, differentiation, and disease, its role in bone homeostasis and development is not well understood. We have previously showed that the induction of glucocorticoid-induced apoptosis occurred through a dose-dependent decrease in Bcl-2. Estrogen prevented glucocorticoid-induced osteoblast apoptosis in vivo and in vitro by preventing the decrease in Bcl-2 in osteoblasts. Therefore, Bcl-2 may be an important regulator of bone growth through mechanisms that control osteoblast longevity and function.

Materials and Methods: Col2.3Bcl-2 mice were developed carrying a 2.3-kb region of the type I collagen promoter driving 1.8 kb of human Bcl-2 (hBcl-2). Tissue specific expression of hBcl-2 in immunoassays validated the transgenic animal model. Histomorphometry and DXA were performed. Proliferation, mineralization, and glucocorticoid-induced apoptosis were examined in ex vivo cultures of osteoblasts. The effect of estrogen on mouse Bcl-2 in ex vivo osteoblast cultures was assayed by RT-PCR and Q-PCR.

Results and Conclusions: Two Col2.3Bcl-2 (tg/+) founder lines were established and appeared normal except that they were smaller than their nontransgenic wildtype (+/+) littermates at 1, 2, and 6 months of age, with the greatest differences at 2 months. Immunohistochemistry showed hBcl-2 in osteoblasts at the growth plate and cortical surfaces. Nontransgenic littermates were negative. Western blots revealed hBcl-2 only in type I collagen-expressing tissues. Histomorphometry of 2-month-old mice showed a significant decrease in tg/+ calvaria width with no significant differences in femoral trabecular area or cortical width compared with +/+. However, tg/+ males had significantly more trabecular bone than tg/+ females. Female +/+ mice showed increased bone turnover with elevated osteoblast and osteoclast parameters compared with +/+ males. Col2.3Bcl-2 mice did not show such significant differences between sexes. Male tg/+ mice had a 76.5 ± 1.5% increase in ObS/BS with no significant differences in bone formation rate (BFR) or mineral apposition rate (MAR) compared with male +/+ mice. Transgenic females had a significant 48.4 ± 0.1% and 20.1 ± 5.8% decrease in BFR and MAR, respectively, compared with +/+ females. Osteoclast and osteocyte parameters were unchanged. By 6 months, femurs from female and male +/+ mice had lost a significant amount of their percent of trabecular bone compared with 2-month-old mice. There was little to no change in femoral bone in the tg/+ mice with age. Ex vivo cultures of osteoblasts from +/+ and Col2.3Bcl-2 mice showed a decrease in mineralization, no effect on proliferation, and an inhibition of glucocorticoid-induced apoptosis in Col2.3Bcl-2 cultures. Estrogen was shown to increase mouse Bcl-2 transcript levels in osteoblast cultures of wildtype mice, supporting a role for Bcl-2 in the sex-related differences in bone phenotype regulated by estrogen. Therefore, Bcl-2 differentially affected bone phenotype in male and female transgenic mice, altered bone cell activity associated with sex-related differences, and decreased bone formation, suggesting that apoptosis is necessary for mineralization. In addition, Bcl-2 targeted to mature osteoblasts seemed to delay bone development, producing a smaller transgenic mouse compared with wildtype littermates. These studies suggest that expression of Bcl-2 in osteoblasts is important in regulating bone mass in development and in the normal aging process of bone.


  1. Top of page
  2. Abstract
  7. Acknowledgements

CELL RENEWAL AND apoptosis occurs during bone development and in adulthood when bone mass peaks but continues to be remodeled through cycles of bone resorption and formation. Apoptosis is induced through two different pathways: the first is through the mitochondrial or organelle pathway and the second is by a mitochondria-independent mechanism such as through the activation of cell surface death receptors (i.e., TNF receptor family). Both pathways eventually activate caspases, which participate in apoptotic cell death. Within the mitochondrial pathway, apoptosis is regulated by a number of positive and negative control proteins. The best known of these control proteins are members of the Bcl-2 family.(1–4) There are at least 15 Bcl-2 family members, which can be either pro- (e.g., Bax, Bak, Bcl-Xs, Bad, and Bid) or anti-apoptotic (e.g., Bcl-2, Bcl-XL, and Bcl-w). These proteins are critical in development, tissue homeostasis, and protection against pathogens. Bcl-2 can protect cells from γ and UV irradiation, cytokine withdrawal, dexamethasone, and cytotoxic drugs. Bcl-2 is localized to the cytoplasmic face of the mitochondria outer membrane, endoplasmic reticulum, and the nuclear envelope. It is involved in ion and small protein flux across the membrane through the COOH-terminal hydrophobic domain. Bcl-2 also prevents the release of cytochrome c from mitochondria and can bind caspases directly or though adapter molecules.

A number of in vitro and in vivo studies have examined the role of Bcl-2. Overexpression of Bcl-2 targeted to specific tissues such as monocytes,(5) neurons,(6, 7) gonads, adrenal glands,(6) and liver(8) has been used successfully to protect cells from apoptosis in those tissues with no gross abnormalities. In MCF-7 breast cancer cells and neurons, estrogen is able to increase Bcl-2 levels and prevent apoptosis.(9–14) Bcl-2 deficient mice (−/−) appeared normal at birth, but by several weeks of age they had a smaller body, as well as a smaller spleen and thymus size, decreased lymphocyte numbers, and immature facial features, and 50% died by 6 weeks because of renal failure.(15–17) In another study, the bone phenotype was analyzed.(18) By 35 days of age, the Bcl-2−/− showed differences in long bone length, increased osteoclast number, and disorganized collagen. We have shown a role of Bcl-2 in glucocorticoid-induced apoptosis of bone.(19) The induction of glucocorticoid-induced apoptosis occurred through a dose-dependent decrease in the Bcl-2/Bax ratio. Estrogen prevented glucocorticoid-induced osteoblast apoptosis in vivo and in vitro by preventing the decrease in Bcl-2 in osteoblasts. In conclusion, Bcl-2 may be an important regulator of bone growth and development through mechanisms that control osteoblast longevity and function.

To better understand the contribution of apoptosis to bone remodeling and bone mass, we established a mouse model with osteoblast-targeted transgene expression of human Bcl-2 (hBcl-2). In Col2.3Bcl-2 mice, the hBcl-2 transcript is under the control of a 2.3-kb type I collagen a1 (Col1a1) promoter fragment in an outbred strain of CD-1 mice. This promoter becomes active in osteoblasts that are synthesizing type I collagen. To understand the phenotypic changes associated with osteoblast overexpression of Bcl-2, we examined physical characteristics, transgene expression, and static and dynamic bone histomorphometry by sex. Data from 2-month-old transgenic and wildtype littermates were the main focus of this study because of our finding that hBcl-2 was strongly expressed in active osteoblasts and the growth of the mice differed so that the greatest difference in growth between tg/+ and +/+ mice was observed at 2 months. Our goal was to determine what role apoptosis/Bcl-2 has in bone formation. The tg/+ mice bred normally and had no apparent abnormalities, except that heterozygote transgenic mice were significantly smaller than the corresponding sex-matched wildtype littermates at 2 months. Histomorphometric evaluation of Col2.3Bcl-2 mice revealed a loss of differences in osteoblast parameters and activity observed between male and female wildtype littermates. The effect of human Bcl-2 was persistent at 6 months, which led to a prevention of age-induced bone loss in the trabecular bones of the femurs of transgenic mice. Our study is the first to show that differences in bone phenotype related to the sex of the animal are affected by Bcl-2 overexpression. In addition, ex vivo osteoblast cultures and in vivo data show that prevention of apoptosis by expression of Bcl-2 affects bone formation.


  1. Top of page
  2. Abstract
  7. Acknowledgements

Col2.3Bcl-2 construct and establishment of the transgenic mice

The 1.8-kb human Bcl-2 (hBcl-2) fragment(20) kindly provided by Jos Vomen, Duke University, was cloned downstream of the rat 2.3-kb Col1a1 promoter (Col2.3)(21–23) and upstream of the bovine growth hormone polyadenylation (bGH-PA) signal by methods similar to those previously described.(24, 25) The orientation and fidelity of hBcl-2 insert was confirmed by forward and reverse sequencing and comparing the sequence data to NCBI accession M14745. The 4.28-kb linear DNA fragment (Fig. 1) containing the Col2.3Bcl-2 fusion gene was excised from the flanking vector with ClaI and used for pronuclear injection of CD-1 (Charles River Laboratories, Wilmington, MA, USA) mouse embryos to generate the transgenic mice. Animals were treated humanely according to the guidelines proposed in the Declaration of Helsinki. Transgenic mice were originally identified by slot-blot hybridization using a ClaI-Pa probe. Subsequent transgenic mice were identified using PCR of ear punch isolated genomic DNA with 5′-tgaagtcaacatgcctgcc and 3′-ctctaaaggtgcggcttcct primers that produce a 670-bp product specific to the 3′ untranslated region (UTR) of hBcl-2. All transgenic mice used in this study were heterozygous for the transgene.

thumbnail image

Figure FIG. 1.. Col2.3Bcl-2 DNA construct. hBcl-2 was targeted to mature osteoblasts by using the 2.3-kb region of the rat 2.3-kb Col1a1 promoter (Col2.3). This promoter drives the transcription of the 1.8-kb human Bcl-2 fragment and the bGH-PA signal. The hBcl-2 insert contains the 5′ UTR within base pairs 1–31, the hBcl-2 cDNA within bases 32–750, and a portion of the 3′ UTR contained within base pairs 751–1860. The hBcl-2 insert was sequenced and was found to be identical to NCBI accession M14745.

Download figure to PowerPoint

Description of mouse physical phenotype

At death, transgenic and wildtype littermates were weighed (g), and body length was measured (mm; tip of nose to tail end) at time of death. Similar results were found with measurements from the tip of the nose to end of the body. After dissection, femur lengths were measured with a micrometer (mm).


Static and dynamic histomorphometry were used to evaluate differences between bones from wildtype and transgenic mice. For dynamic histomorphometry, mice received 10 mg/kg body weight of calcein in 2% NaHCO3 7 days before death, and 2 days before death, 90 mg/kg body weight of xylenol orange administered intraperitoneally. Femurs were dissected free of tissue and fixed in 70% ethanol for 3–5 days. The femurs were dehydrated in increasing concentrations of ethanol, cleared in xylene, and embedded undecalcified in methyl methacrylate. Five-micrometer-thick longitudinal serial sections were cut. Sections taken from the middle of the femur, where the central vein is located, were stained with modified Masson trichrome stain.(26) Osteoblasts were identified as cuboidal cells lining the trabecular perimeter. Osteoclasts were identified as multinucleated cells on the trabecular bone surface. Osteocyte number per bone area were evaluated in the trabeculae of femurs. For static measurements, femurs and calvaria were dissected free of surrounding tissue and fixed in 4% paraformaldehyde in PBS, pH 7.4, at 4°C for 2–5 days. The bones were decalcified in EDTA/NH3OH for an additional 2–5 days and dehydrated in progressive concentrations of ethanol, cleared in xylene, and embedded in paraffin. The embedded bones were sectioned longitudinally and stained for TRACP and counterstained with hematoxylin. Osteoclasts were identified by TRACP staining and by their characteristic multinucleated morphology. Histomorphometric measurements were made in a blinded, nonbiased manner using the BioQuant or OsteoMeasure computerized image analysis system (BIO-QUANT; R & M Biometrics, Nashville, TN, USA; OsteoMetrics, Atlanta, GA, USA) interfaced with a Nikon E400 microscope (Nikon, Melville, NY, USA). The terminology and units used are those recommended by the Histomorphometry Nomenclature Committee of the American Society for Bone and Mineral Research.(27) All measurements were confined to the secondary spongiosa and restricted to an area between 400 and 2000 μm distal to the growth plate-metaphyseal junction of the distal femur. Cortical widths were measured between 1200 and 3200 μm distal to the growth plate-metaphyseal junction on both sides of the distal femur, and the mean was determined from multiple measurements at set distances along the femur.


Chrom-alum-coated slides with 5-μm-thick femur or calvaria sections from wildtype (CD-1) and transgenic (Col2.3 Bcl-2) mice were baked for 1 h at 56°C, deparaffinized in xylene, and rehydrated in decreasing percentages of ethanol in deionized water. Antigen retrieval was performed using 4N HCl for 10 minutes at 37°C in a humid container. Endogenous peroxidase activity was blocked by incubating 10 minutes with 3% hydrogen peroxide. Nonspecific protein was blocked by incubation with Power Block (BioGenex, San Ramon, CA, USA) for 30 minutes at room temperature. Sections were incubated with humidity at room temperature for 1 h with 4 μg/ml of mouse anti-human Bcl-2 antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) in 50 mM Tris HCl, pH 7.4, 1% BSA. Secondary antibody and substrate staining were done as per manufacturer's instructions from the Envision+ kit (DAKO, Carpinteria, CA, USA) and developed with either DAB or AEC. The slides were counterstained with Harris modified hematoxylin and coverslipped in Crystal/Mount aqueous-based mounting solution (Biomeda, Foster City, CA, USA).

DXA analysis

All analysis was conducted using the Lunar PIXImus densitometer (Lunar Corp., Madison, WI, USA) calibrated with a phantom of defined value. Two-month old Col2.3Bcl-2 and CD-1 wildtype mice were killed and frozen at −20°C. After thawing, the animals were placed in the same direction with the whole body over the imaging area of the densitometer (excluding portions of the skull). For each analysis, the skull was defined as a region of exclusion and individual femurs were separately analyzed as regions of interest. Total and femoral BMD (mg/cm2) and BMC (g) were determined by the image analysis software provided with the instrument.

Western blot analysis

At death, 11 tissue types (described below) per mouse were collected, snap frozen, and stored at −70°C. The tissues were homogenized in 0.5 mM DTT, 0.1 mM EDTA, 10 mM Tris-HCl buffer (pH 7.2), 250 mM sucrose, leupeptin (10 μg/ml), aprotinin (10 μg/ml), 1 mM phenylmethylsulphonylfluoride (PMSF), and 0.1% Triton X-100 by polytron on ice. For cell culture experiments, the adherent cells were washed twice with ice-cold PBS and lysed in ice-cold RIPA buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, protease inhibitor cocktail [Sigma], 1% Triton x-100, 1% sodium deoxycholate, 0.1% SDS, and 0.004% sodium azide) by shaking at 4°C for 15 minutes. The homogenates were centrifuged 10,000g for 20 minutes at 4°C. The supernatant protein concentration was determined by BCA assay according to the manufacturer's instruction (Pierce, Rockford, IL, USA). Proteins (50 μg) were separated by a 10–20% gradient SDS-PAGE system and transferred to nitrocellulose membrane. The blots were blocked with 5% (wt/vol) nonfat milk in TTBS (20 mM Tris, 150 mM NaCl, pH 7.5, and 0.05% Tween-20) for 1 h at room temperature and incubated with 1:100 monoclonal mouse anti-human Bcl-2 antibody (sc509) or rabbit anti-mouse actin antibody (sc1615; Santa Cruz Biotechnology) overnight at 4°C in blocking buffer, washed three times, followed by 1:10,000 horseradish peroxidase conjugated goat anti-mouse or anti-rabbit secondary antibody. The protein bands were visualized using an Immun-Star HRP substrate kit (Bio-Rad Laboratories, Hercules, CA, USA) according to manufacturer's instructions.

TUNEL staining for apoptosis

Two month-old tg/+ and +/+ mice were treated with 1 mg/kg body weight (BW) of dexamethasone or saline vehicle each day for 7 days and killed. In situ hybridization of apoptotic cells by TUNEL assay was performed on paraformaldehyde-fixed, decalcified, paraffin-embedded femurs and calvaria isolated from wildtype and Col2.3Bcl-2 transgenic mice. Five-micrometer bone sections were cut onto chrom-alum-coated slides, baked at 56°C for 1 h, deparaffinized, and rehydrated. Apoptotic cells were identified using TACS TdT kit (R&D Systems, Minneapolis, MN, USA) using DAB as a substrate. The slides were counterstained with hematoxylin or methyl green. Quantification of percent TUNEL+ osteoblasts was determined in a blinded nonbiased manner and involved counting the number of stained and nonstained osteoblasts within three separate regions: (1) trabecular osteoblasts (primary and secondary spongiosa), (2) periosteal cells, and (3) endosteal osteoblasts (metaphysis and diaphysis). Two sections/slide from each bone were counted in three separate experiments.

In +/+ and tg/+ osteoblast cultures, apoptosis was also examined by TUNEL. Because of the extensive number of apoptotic cells in the nodules that mineralize and the multilayering of cells in the nodules, individual apoptotic cells could not be accurately counted. Instead, the area of TUNEL+ brown-stained cells in the nodules was visualized and traced using OsteoMeasure (see Histomorphometry section). Total nodule area and the area of apoptotic nuclei were measured to calculate the percent of apoptotic nodule in a blinded, nonbiased manner for tg/+ and +/+ osteoblast-like cultures.

Ex vivo cell culture

Calvarial cells were isolated from 5- to 8-day-old wildtype and transgenic mice using a modification of the method described by Wong and Cohn.(28) Briefly, after removal of sutures, calvaria were subjected to four sequential 15-minute digestions in an enzyme mixture containing 0.05% trypsin (Gibco BRL, Grand Island, NY, USA) and 0.1% collagenase P (Roche Applied Science, Indianapolis, IN, USA) at 37°C on a rocking platform. Cell fractions two to four were collected. The fractions were pooled, centrifuged, resuspended in DMEM containing 10% FBS, 100 U/ml of penicillin, and 100 μg/ml of streptomycin (Gibco BRL), and filtered through a 70-μm cell strainer. Cells were plated at a density of 1 × 104 cells/cm2 in Primaria 6-well plates (Becton Dickinson Labware, Franklin Lakes, NJ, USA) in DMEM containing 10% FBS and antibiotics. Twenty-four hours later, medium was changed. On the sixth day of the culture, medium was changed to differentiation medium (α-MEM containing 10% FBS, 50 μg/ml of ascorbic acid, 4 mM β-glycerophosphate, and antibiotics) and the medium was changed every 2 days.

Generation of growth curve

Cell proliferation was assessed by measuring DNA content by fluorometric analysis. Freshly isolated cells were plated at a density of 1 × 104 cell/cm2 into 24-well plates in DMEM containing 10% FBS and antibiotics. On each of 9 days, medium from one plate was aspirated, and that plate was frozen at −70°C. After collecting all the plates for 9 days, the cells were disrupted with 0.01% SDS. Cell lysates were transferred to a 96-well plate, and 10 μg/ml H33258 was added to the cell lysate. DNA content was measured at 360 nm excitation and 460 nm emission by a microplate reader (Bio-TEK instruments, Winooski, VT, USA).

RT-PCR and quantitative real-time PCR

Primary isolate neonatal calvarial cells were plated at a density of 2 × 104 cells/cm2 in Primaria 6-well plates in DMEM containing 10% charcoal-stripped FBS and antibiotics. Twenty-four hours later, the cells were treated with varying concentrations of β-estradiol (Sigma) or 17α-estradiol (Sigma) as a negative control or media alone as an experimental control. The cells were cultured for 24 h and harvested. Total RNA was isolated by Tri Reagent (Molecular Research Center, Cincinnati, OH, USA), according to the manufacturer's protocol. Reverse transcriptase treatment of RNA was performed at 42°C for 80 minutes using 3 μg of total RNA in the presence of oligo (dT) primers (Gibco BRL), and BD Sprint Powerscript (BD Biosciences Clontech, Palo Alto, CA, USA). The reaction was terminated by heating at 70°C for 15 minutes. Amplification by PCR was performed on the iCycler (Bio-Rad Laboratories) with a set of primers to detect mouse Bcl-2 (m-Bcl-2) mRNA. The 5′ primer was CTGGCATCTTCTCCTTCCAG, and the 3′ primer was GACGGTAGCGACGAGAGAAG, with a predicted size of 183 bp. Four microliters of the RT RNA mixture was used as templates. After 95°C for 3 minutes, 32 cycles were run at 94°C for 30 s (denaturation), 54°C for 30 s (annealing), and 72°C for 20 s (extension), followed by 72°C for 5 minutes. GAPDH (Clontech Laboratory; predicted size, 452 bp) was amplified as an internal control. The resulting products were fractionated by electrophoresis on a 2% agarose gel and visualized by ethidium bromide staining.

For quantitative real-time PCR, total RNA was isolated as described above. First-strand cDNA for real-time PCR was synthesized from 3 μg of total RNA using oligo (dT) primers (Gibco BRL), and BD Sprint Powerscript (BD Biosciences) at 42°C for 80 minutes. The quantitative real-time PCR was performed in an iCycler (Bio-Rad Laboratories) using iQ SYBR Green Supermix (Bio-Rad Laboratories). Reactions were performed in 10 μl of a mixture containing 4 μl cDNA dilution, 1 μl primers, and 5 μl iQ SYBR Green Supermix containing iTaq DNA polymerase, reaction buffer, dNTP mixture, and the SYBR Green I. 5′ and 3′ primers are the same as listed above. PCR conditions were as follows: 1 cycle of 15 minutes at 95°C, followed by 40 cycles of 95°C for 15 s, 55°C for 30 s, and 72°C for 20 s, 1 cycle of 95°C for 1 minute, and finally 55°C for 1 minute. PCR amplification was performed in triplicate and repeated in three independent experiments. The expression level was normalized against ribosomal 18S using the following primers: 5′-TCAAGAACGAAAGTCGGAGG and 3′-GGACATCTAAGGGCATCACA and is expressed as the relative ratio between mBcl-2 and 18S.

Acridine orange/ethidium bromide staining

Freshly isolated wildtype and transgenic osteoblasts were plated into 24-well culture plates at a density of 1 × 104 cells/cm2 in DMEM containing 10% FBS, 100 U/ml of penicillin, and 100 μg/ml of streptomycin (Gibco BRL). At subconfluence, the cells were treated with 1, 10, 100, or 1000 nM of corticosterone for 72 h. Untreated cells were used as control. After treatment, medium was aspirated, and cells were labeled with the nucleic acid-binding dye mixture of 100 μg/ml acridine orange and 100 μg/ml ethidium bromide (Sigma) in PBS. The cells were examined by fluorescence light microscopy. Viable cells had yellow-green fluorescent nuclei with an organized structure. In early apoptotic cells, nuclei containing yellow chromatin were highly condensed or fragmented. Apoptotic cells also exhibited membrane blebbing. The late apoptotic cells had orange chromatin with nuclei that were highly condensed and fragmented. The necrotic cells had bright orange chromatin in round nuclei. Only cells with yellow, condensed, and fragmented nuclei were counted as apoptotic cells, in a blinded, nonbiased manner. For each sample, at least 500 cells per well and 4 wells per treatment group were counted, and the percentage of apoptotic cells was determined.

  • equation image

The apoptotic index was calculated as the ratio of percent of apoptotic cells in the treatment group versus control.

Calcium content

Primary wildtype and transgenic osteoblasts were cultured in α-MEM containing 10% FBS, 50 μg/ml of ascorbic acid, and 4 mM β-glycerophosphate after they became confluent. On days 7, 14, 21, and 28, cells were treated with 5% TCA overnight. Calcium content in the TCA cell extracts was measured colorimetrically with a calcium kit (Eagle Diagnostics, Desoto, TX, USA) according to manufacturer's instructions.

Von Kossa staining

Mineralization was assessed using a modified von Kossa silver nitrate staining method. Cells were fixed for 10 minutes in 2% paraformaldehyde in 0.1 M sodium cacodylate, pretreated for 20 minutes with saturated lithium carbonate solution, and washed in deionized water. The cells were incubated with 5% silver nitrate solution for 30 minutes under a bright light, followed by washing with water and air-drying.

Statistical analysis

Statistical analysis was performed by one-way ANOVA, followed by Student's t-test to determine significance between groups. In this text, significant differences are p ≤ 0.05.


  1. Top of page
  2. Abstract
  7. Acknowledgements

Establishment of the Col2.3Bcl-2 mouse

To explore the role of Bcl-2 in bone remodeling and development, we established and characterized a transgenic mouse that expresses hBcl-2 in osteoblasts (Col2.3Bcl-2). The Col2.3Bcl-2 mouse was created using a rat 2.3-kb Col1a1 promoter fragment that drives the transcriptional expression of the 1.8-kb region of human Bcl-2 cDNA (Fig. 1). Two transgenic heterozygote (tg/+) founder lines (1 and 2) derived from CD-1 wildtype (+/+) mice were established. Heterozygote Col2.3Bcl-2 transgenic mice were identified by PCR as described in the Materials and Methods section. The resulting mouse lines were derived from original founders that consist of a female for founder 1 and a male for founder 2. The transgenic mice and their corresponding littermates appeared healthy. No significant differences were found in litter size or the reproductive rate of the transgenic heterozygote mice compared with wildtype.

Tissue transgene expression

To confirm transgene protein expression, calvaria were isolated from 1-, 2-, and 6-month-old Col2.3Bcl-2 transgenic and wildtype mice in the two founder lines and were probed with anti-human specific Bcl-2 antibody by immunoassay. The linear range for hBcl-2 detection was determined by immunoblot assay using concentrations of 25–200 μg total calvaria protein. In these studies, 50 μg of total protein fell within the mid-portion of the linear detection range of staining intensity for actin and Bcl-2 (data not shown). Representative samples from the founder lines and wildtype controls were compared. The immunoblots were probed with anti-actin as a sample protein loading control and revealed a single band at 43 kDa. Staining with anti-hBcl-2 presented a single band corresponding to 26 kDa. Founder line 2 showed slightly greater staining intensity than founder 1 (Fig. 2A). Wildtype samples were consistently negative for hBcl-2 expression even with overloading of protein. There were also no significant differences in the expression of hBcl-2 between 1-month-old male and female transgenic mice from the same litter and from two litters of 2-month-old male and female transgenic mice (Fig. 2B). Densitometry scans of hBcl-2 expression of four 2-month-old male and female mice were performed, and after normalization to actin, the hBcl-2 ratio between male and female mice was 1.15, with similar results for 1-month-old mice, ratio 1.12. Therefore, according to the Western blots, no significant differences in hBcl-2 expression between male and female transgenic mouse calvaria could be found.

thumbnail image

Figure FIG. 2.. (A) A representative Western blot from three independent experiments from the calvaria of the progeny from the two heterozygote Col2.3Bcl-2 transgenic mouse lines shows the expression of human Bcl-2 at 2 months of age. Human Bcl-2 expression is absent in the 2-month-old CD-1+/+ control littermates. Actin was used as a control for standardization of equal amounts of protein per sample. (B) No significant differences were observed in hBcl-2 expression in males (M) and females (F) in the calvaria of one litter from 1-month-old Col2.3Bcl-2 mice and in the calvaria from two litters of 2-month-old Col2.3Bcl-2 mice. Wildtype littermates (+/+) were consistently negative for transgene expression. (C) Human Bcl-2 tissue-specific expression in the heterozygote Col2.3Bcl-2 mouse. A representative Western blot from three independent experiments with anti-human Bcl-2 antibody shows strong reactivity in collagen-containing tissues of founder 1 progeny.

Download figure to PowerPoint

To validate the specificity of the collagen driven promoter and expression system, we examined hBcl-2 protein expression in 11 different tissues. In the immunoblots, human Bcl-2 protein expression was consistently observed in founder lines 1 and 2 in the collagen-rich tissues (skin, tail tendon, spine, calvaria, and femur) and no staining was observed in samples of other tissues (spleen, kidney, lung, brain, and heart) (founder 1; Fig. 2C) or tissues of wildtype littermates.

hBcl-2 expression was also examined by immunohistochemistry in femurs of 2-month-old mice. Immunohistochemical expression of hBcl-2 in wildtype femurs was consistently negative (Figs. 3A and 3B) as was the primary antibody isotype control (results not shown). These results as well as the Western blots (single band at 26 kDa) showed the specificity of the antibody for human Bcl-2. Osteoblasts stained positively in areas that appeared to be undergoing active bone formation along the endosteal (Fig. 3C) and periosteal (Fig. 3D) surface of cortical bone. Additionally, trabecular osteoblasts were positive for hBcl-2 (Fig. 3E). Together, these immunoassays show that mice derived from the original two founder lines express similar levels of hBcl-2 protein, the expression of hBcl-2 is limited to tissues rich in type I collagen, and mature osteoblasts are the exclusive expressers of hBcl-2 in femurs.

thumbnail image

Figure FIG. 3.. Immunohistochemistry shows that human Bcl-2 is expressed in the femurs of mature active osteoblasts in the 2-month-old Col2.3Bcl-2 male and female mice. (A and B) Nontransgenic (+/+) littermates were negative for hBcl-2 expression. Col2.3Bcl-2tg/+ mouse femurs showed osteoblast expression of hBcl-2 at regions of the (C) endosteal surface and areas of active bone formation on the (D) periosteal surface of cortical bone (arrows) and osteoblasts at the growth plate (E). Representative photomicrographs revealed no significant differences between males and females from 12 samples from three independent experiments. Bar = 20 μm.

Download figure to PowerPoint

Physical characterization

Wildtype mice were compared with Col2.3Bcl-2 littermates for weight, body length, and femur length at 1, 2, and 6 months of age (Fig. 4). Data from founder line 1 is shown with similar observations made for founder line 2. Overall, the transgenic mice were smaller than wildtype for all parameters except at 6 months of age, when the femur lengths were similar.

thumbnail image

Figure FIG. 4.. Analysis of body weight, body length, and femur length at 1, 2, and 6 months of age. Col2.3Bcl-2 heterozygote mice (tg/+) are significantly smaller than the corresponding nontransgenic littermates (+/+) at 1, 2, and 6 months of age. The number of animals evaluated were (1) 1 month old (28–34 days) +/+ = 15, tg/+ = 14, (2) 2 months old (58–72 days) +/+ = 64, tg/+ = 38, and (3) 6 months old (168–192 days) +/+ = 18, tg/+ = 19. Data represent means ± SD.

Download figure to PowerPoint

At 1 and 2 months of age, statistical differences were observed for body weight, body length, and femur length between wildtype and transgenic mice including differences between sexes. One-month-old mice showed the same phenotype as 2-month-old mice; therefore, only the 2-month data are discussed in detail (Fig. 4). Both 2-month wildtype males and females had a greater body weight than the corresponding transgenic littermates (males [33.5 ± 3.3 versus 29.4 ± 5.0 g; p≤ 0.005]; females [25.5 ± 1.7 versus 21.4 ± 4.2 g; p ≤ 0.005]). Body length was also greater in wildtype compared with transgenic male and female mice (male [180 ± 5.9 versus 168 ± 10.0 mm; p ≤ 0.005]; female [170 ± 7.5 versus 155 ± 13.4 mm; p ≤ 0.005]). Additionally, differences were observed for femur length between wildtype and the shorter transgenic mice (male [15.9 ± 0.5 versus 15.2 ± 0.9 mm; p ≤ 0.005]; female [15.4 ± 0.5 versus 14.1 ± 1.3 mm; p ≤ 0.005]).

By 6 months of age, although the wildtype mice were generally heavier (39.2 ± 10.0 versus 32.9 ± 7.8 g; p ≤ 0.05) and longer (188.7 ± 8.5 versus 180.8 ± 10.3 mm; p ≤ 0.05) than their corresponding transgenic littermates, no differences in femur length were observed (Fig. 4). Therefore, sex-related differences in body weight and length, which were observed at 1 and 2 months of age, were maintained at 6 months of age.

Static and dynamic histomorphometry

Static and dynamic histomorphometry was used to evaluate the skeletal phenotype of 2- and 6-month-old Col2.3Bcl-2 transgenic and wildtype mice. Osteoblast and osteoclast parameters were first analyzed to determine if the overexpression of Bcl-2 affected cell parameters. All histomorphometry was done using animals from founder line 1 with similar results from founder line 2. When we examined the effect of transgene by sex, male tg/+ had a 76.5% increase in the percent osteoblast surface (ObS/BS) compared with male +/+ (28.6 ± 3.9% versus 16.2 ± 1.5%; p ≤ 0.005). However, female tg/+ mice did not show a significant increase compared with female +/+. Examination of the relationship between male and female CD-1 +/+ mice showed that wildtype female mice had nearly 60% greater ObS/BS than wildtype males (27.3 ± 2.0% versus 16.2 ± 1.5%; p ≤ 0.001), but the Col2.3Bcl-2 mice did not have the same difference between male and female mice. The most remarkable difference was the increase in ObS/BS in male Col2.Bcl-2 mice. This increase abrogated the male/female differences normally observed in CD-1 wildtype mice (Fig. 5A).

thumbnail image

Figure FIG. 5.. (A) Static histomorphometry of osteoblast surface per bone surface (Ob.S/BS, %). The wildtype (+/+) female mice have a higher Ob.S/BS than male +/+ mice; however, the transgenic mice do not show this sex-based difference. Instead, transgenic (tg/+) male mice had a 76.5% increase in Ob.S/BS compared with +/+ male mice, therefore abrogating the sex-based difference observed in the wildtype mice. Number of mice evaluated: male +/+ = 10, female +/+ = 12, male tg/+ = 9, and female tg/+ = 9. Data represent means ± SE. (B) The number of osteoclasts per bone surface (N.Oc/BS) is not significantly different in the wildtype vs. the Col2.3Bcl-2 transgenic mouse femurs. However, N.Oc/BS is significantly higher in the female +/+ mice compared with male +/+ mice. Number of mice evaluated: male +/+ = 10, female +/+ = 12, male tg/+ = 9, and female tg/+ = 9. Data represent means ± SE. (C) The number of osteocytes per bone area (N.Ot/BA) is not significantly different in the male and female wildtype vs. the Col2.3Bcl-2 transgenic mouse femurs. Number of mice evaluated: male +/+ = 4, female +/+ = 6, male tg/+ = 6, and female tg/+ = 4. Data represent means ± SE.

Download figure to PowerPoint

We next examined osteoclast parameters in the 2-month-old Col2.3Bcl-2 and wildtype mice. Both male and female tg/+ mice were similar to male and female +/+ for NOc/BS (Fig. 5B). Wildtype mice displayed a sex difference in which the +/+ female mice had a greater NOc/BS than the corresponding male +/+ mice (6.8 ± 0.4/mm versus 4.8 ± 0.4/mm; p ≤ 0.05). Both ObS/BS (Fig. 5A) and NOc/BS (Fig. 5B) were higher in the female +/+ mice than in the male +/+ mice, showing that wildtype CD-1 female mice had a higher bone turnover rate than the corresponding CD-1 males. However, there was no significant effect of the transgene on osteoclast parameters. Neither were osteocyte numbers per bone area (N.Ot/BA) significantly different in male and female transgenic or +/+ mice, although a trend for increased osteocytes was found in the trabeculae of transgenic mice (Fig. 5C).

To further understand how overexpression of Bcl-2 affected the activity of osteoblasts, dynamic measurements of bone formation were analyzed. Most interesting, female tg/+ mice had a 48% decrease in bone formation rate/bone surface (BFR) compared with +/+ females (0.49 ± 0.07 versus 0.95 ± 0.11 μm3/μm2/day; p ≤ 0.01), whereas male tg/+ were not different from male +/+. Between sexes, we observed that the +/+ males mice had a 61% lower BFR than that of the female +/+ mice (0.37 ± 0.04 versus 0.95 ± 0.11 μm3/μm2/day; p ≤ 0.001). Similarly, tg/+ mice also displayed a sex difference, albeit not as great as the +/+, with a 37% difference between male and female tg/+ mice (0.31 ± 0.04 versus 0.49 ± 0.07 μm3/μm2/day; p ≤ 0.05), with females having a greater BFR (Fig. 6A).

thumbnail image

Figure FIG. 6.. (A) Dynamic histomorphometry showing BFR. In wildtype mice, BFR is significantly elevated in the female trabecular bone vs. male trabecular bone; however, this sex difference is lessened in the transgenic mice, particularly in the females, in which the tg/+ female BFR is significantly less than the +/+ female BFR. Number of mice evaluated: male +/+ = 14, female +/+ = 10, male tg/+ = 11, and female tg/+ = 7. Data represent means ± SE. (B) Dynamic histomorphometry showing MAR. In wildtype mice, MAR is significantly elevated in the female femurs compared with the male femurs, and this difference persists in the transgenic mice. However, MAR is significantly lower in the tg/+ female mice compared with the +/+ female mice. Number of mice evaluated: male +/+ = 14, female +/+ = 10, male tg/+ = 11, and female tg/+ = 7. Data represent means ± SE.

Download figure to PowerPoint

Similar to BFR, the mineral apposition rate (MAR) between male +/+ and male tg/+ revealed no difference; however, the tg/+ females had a lower MAR compared with the +/+ females (2.07 ± 0.14 versus 2.59 ± 0.15 μm/day; p ≤ 0.05). The female +/+ mice had a higher MAR compared with the +/+ male mice (2.59 ± 0.15 versus 1.78 ± 0.15 μm/day; p ≤ 0.005). Similarly, the tg/+ mice displayed a sex difference with the females showing a larger MAR than the males (2.07 ± 0.14 versus 1.44 ± 0.12 μm/day; p ≤ 0.01), although the difference was not as great as was seen in the wildtype (Fig. 6B). Therefore, not only was the amount of bone surface actively involved in bone formation (BFR) decreased in female tg/+ mice but also the amount of mineral deposited at the active sites of bone formation was lower (MAR).

To understand if these differences in cellular parameters and BFRs affected bone mass, static histomorphometry was performed on femurs and calvaria at 2 and 6 months of age. Male tg/+ mice had a greater trabecular area than the female tg/+ mice (11.3 ± 1.1% versus 7.5 ± 1.3%; p ≤ 0.05), with no significant difference between sexes in the wildtype (Table 1). The cortical width was not statistically different for +/+ and tg/+ mice or between sexes. Tg/+ mice had thinner calvaria (CWi) than the corresponding +/+ littermates (108.7 ± 7.9 versus 130.9 ± 3.4 μm; p ≤ 0.05). No difference between male +/+ versus tg/+ calvaria was found, but calvaria width for female tg/+ mice was smaller than for the +/+ female mice (103.2 ± 14.1 versus 126.3 ± 3.6; p ≤ 0.05; Table 1). Interestingly, by 6 months, the differences in calvaria width for the female tg/+ compared with female +/+ were not significantly different (129.3 ± 8.3 versus 108.3 ± 7.2; N = 8–10 mice per group).

Table Table 1.. Static Histomorphometry of 2-month-old Mice
Thumbnail image of

In contrast, the differences between +/+ and Col2.3Bcl-2 mice at 6 months in static histomorphometric parameters of femurs were greater than at 2 months. Male and female tg/+ mice had greater bone volume (BV/TV) and trabecular number (TbN/mm), but decreased trabecular separation (TbSp) than sex-matched +/+ mice. Additionally, 6-month-old female tg/+ mice had a 3-fold greater percent osteoblast surface (%ObS) and higher osteocyte number (N.Ot/B.Ar) than the corresponding +/+ females (Table 2). A comparison between 2- and 6-month-old mouse histomorphometric data revealed a role for Bcl-2 expression in blunting the age-associated effects on bone, especially in female mice. Wildtype female mice at 2 versus 6 months showed a significant decrease in BV/TV (8.5 ± 0.6 versus 2.9 ± 0.6; p ≤ 0.00), %ObS (27.3 ± 1.7 versus 4.5 ± 1.2; p ≤ 0.00), and TbN/mm (3.1 ± 0.2 versus 0.7 ± 0.07; p ≤ 0.00), with increased TbSp (440.2 ± 28.6 versus 1502.8 ± 179.4; p ≤ 0.004) and TbTh (27.3 ± 0.97 versus 41.8 ± 5.1; p ≤ 0.05). Conversely, 2- versus 6-month-old tg/+ female mice show no significant differences for BV/TV (7.5 ± 1.3 versus 6.3 ± 0.5; p > 0.05), TbSp (555.5 ± 78.2 versus 719.6 ± 208.3; p > 0.05), TbTh (26.7 ± 2.4 versus 30.6 ± 2.9; p > 0.05), and TbN/mm (2.7 ± 0.3 versus 1.9 ± 0.3; p > 0.05). Although %ObS was significantly decreased in 6- versus 2-month tg/+ female mice (14.2 ± 3.05 versus 31.2 ± 3.4; p ≤ 0.002), the difference of 54% observed in the female tg/+ was not as great in the female +/+ mice (84%).

Table Table 2.. Static Histomorphometry of 6-month-old Mice
Thumbnail image of

DXA analysis

Femoral and total (entire body) BMD (mg/cm2) and BMC (g) were examined for the Col2.3Bcl-2 transgenic mice and the corresponding wildtype littermates. DXA data were segregated by sex, and the results for BMD are displayed (Fig. 7). The BMD for tg/+ female mice was lower than the BMD for male tg/+ mice (73.9 ± 6.0 versus 80.2 ± 7.4 mg/cm2; p ≤ 0.002). This is distinct from +/+ mice, in which male and female mice have a similar BMD. Additionally, the transgenic female mice had a lower BMD and BMC than female +/+ mice (73.9 ± 6.0 versus 81.4 ± 7.2 mg/cm2; p ≤ 0.001 and 27.8 ± 3.9 versus 32.6 ± 6.0 g; p ≤ 0.004, respectively). The decreases in the BMD and BMC in female tg/+ mice compared with female +/+ mice correlate well with the histomorphometric results.

thumbnail image

Figure FIG. 7.. Femur BMD. Female Col2.3Bcl-2 mice had a decreased bone mass compared with BMD for female wildtype littermates and male Col2.3Bcl-2 mice. Number of femurs: n = 22 for +/+ males, n = 12 for +/+ females, n = 21 for tg/+ males, and n = 30 for tg/+ females. Data represent means ± SD.

Download figure to PowerPoint

Ex vivo cultures

Primary osteoblast cultures from +/+ and tg/+ mice were obtained and analyzed for their ability to proliferate and mineralize. In Fig. 8A, no differences in the growth curves were found in 9 days of culture. Growth had reached a plateau by day 6. No statistical difference was observed in the doubling time 42.6 h for the wildtype and 44.7 h for the Col2.3Bcl-2 cells (r2 = 0.99 and 0.99, respectively; p > 0.05). A Western blot of human Bcl-2 expression in culture showed that hBcl-2 was expressed throughout the culture period (Fig. 8B). Therefore, human Bcl-2 expression does not seem to affect cell proliferation in vitro.

thumbnail image

Figure FIG. 8.. (A) Growth curve for ex vivo cultures of wildtype and Col2.3Bcl-2 osteoblasts derived from neonatal calvaria. In 9 days of culture, no differences were found between +/+ and Col2.3Bcl-2 osteoblasts. (B) Western blot of human Bcl-2 expression in ex vivo cultures of Col2.3Bcl-2 osteoblasts compared with wildtype osteoblasts for 28 days. Human Bcl-2 is expressed only in the transgenic cultures for the entire time period. (C) Calcium content of ex vivo cultures of wildtype and Col2.3Bcl-2 osteoblasts derived from neonatal calvaria. Mineralization is significantly decreased (*p < 0.001) only at 28 days in the Col2.3Bcl-2 osteoblast cultures. (D) Von Kossa staining of the ex vivo cultures of wildtype and Col2.3Bcl-2 osteoblasts. More stained nodules are found at 28 days of culture of the osteoblasts from wildtype mice compared with those from Col2.3Bcl-2 mice. (E) Photomicrograph of TUNEL assay on nodules from 28-day wildtype and Col2.3Bcl-2 osteoblast cultures. Nodules from cell cultures from +/+ mice show a greater proportion of TUNEL+, brown nuclei than nodules from cultures from tg/+ mice; bar = 100 μm. Insets show the TUNEL+, condensed, and irregular nuclei (arrows) at higher magnification, whereas the inset from tg/+ osteoblast cultures have mostly normal nuclei; bar = 20 μm.

Download figure to PowerPoint

Differentiation of the Col2.3Bcl-2 and wildtype cell cultures was assessed by measuring calcium content at 7, 14, 21, and 28 days of culture. No significant differences between the Col2.3Bcl-2 and wildtype cultures were found in mineralization at 7, 14, and 21 days of culture. However by 28 days, the primary, wildtype osteoblast cultures increased their mineral content 2.0-fold more than the primary osteoblast cultures from the Col2.3Bcl-2 mice (Fig. 8C). Von Kossa staining for mineralized matrix confirmed the calcification data for all time-points. At 28 days, more Von Kossa-stained, mineralized nodules were found in the +/+ cultures compared with the tg/+ (Fig. 8D). This finding was consistent in three separate experiments with cells from different litters. These data supported the in vivo data and showed a decrease in bone formation from osteoblasts overexpressing Bcl-2. Wildtype osteoblast cultures had a significantly greater percentage of apoptotic cells shown by TUNEL+ staining (Fig. 8E) than the transgenic cultures at day 28 (23.4 ± 1.7% versus 15.5 ± 1.8%; p ≤ 0.02, respectively). Insets in Fig. 8E show TUNEL+, darkly stained, condensed, irregular shaped nuclei in the +/+ and normal, oval-shaped nuclei in the tg/+ cultures. No significant differences in the percentage of apoptotic cells were found at the earlier time points (data not shown), confirming the calcification data. Thus, mineralization is increased in osteoblast cultures in which apoptosis occurs, whereas prevention of apoptosis inhibits calcification in the later stages of bone formation in vitro.

Glucocorticoids effect on apoptosis

To assess the ability of the Col2.3Bcl-2 construct to inhibit apoptosis, primary cultures of Col2.3Bcl-2 and wildtype cells were treated for 72 h with varying concentrations of corticosterone. This protocol was previously shown to be optimal for a dose-dependent increase in apoptosis in mouse and rat-derived osteoblast populations.(19) Osteoblasts from +/+ cultures had a dose-dependent increase in apoptosis, whereas no increase in apoptosis was found in the Col2.3Bcl-2 tg/+ cells (Fig. 9A). In vivo analysis of apoptosis by TUNEL also showed that the Col2.3Bcl-2 construct was able to prevent glucocorticoid-induced apoptosis in 2-month-old mouse calvaria (Fig. 9B). Both apoptotic osteoblasts and osteocytes were decreased in the transgenic mice compared with wildtype. Light micrographs of TUNEL stained calvaria are shown in Fig. 9C, in addition to a negative control with no TdT and a positive control of DNase-treated bones. To determine if sex affected apoptotic rates in the wildtype mice, the metaphysis region of the distal femurs (Figs. 9D and 9E) were analyzed because the basal rate of apoptosis in 2-month-old mice was so low in calvaria that meaningful significant differences could not be obtained (Fig. 9B). Because of extensive bone formation and remodeling, the rate of apoptosis is greater near the growth plate compared with the rate in the calvaria. Thus, apoptosis seems to plays a role during active bone formation in the metaphysis rather than in less active regions such as calvaria at 2 months of age. In the metaphysis, although overall, tg/+ has significantly lower percentage apoptotic osteoblasts than +/+, no significant differences in the percentage of apoptotic ostseoblasts were found between male and female mice in wildtype and Col2.3Bcl-2 osteoblasts. The females seemed to have a greater apoptotic rate, probably because of increased bone remodeling compared with the males, but this was not statistically significant. A light micrograph of TUNEL-stained osteoblasts is shown in Fig. 9E. Thus, human Bcl-2 expression not only blocked glucocorticoid-induced apoptosis, but also decreased the basal rate of apoptosis.

thumbnail image

Figure FIG. 9.. (A) Apoptosis in ex vivo osteoblast cultures from Col2.3Bcl-2 and wildtype mice treated with or without (control) varying concentrations of corticosterone (1.0–1000 nM) for 72 h. Human Bcl-2 expression was able to block glucocorticoid-induced apoptosis at all concentrations of corticosterone. (B) Graph of the percent of apoptotic cells determined by TUNEL in 2-month-old calvaria treated for 7 days with or without 1 mg/kg body weight of dexamethasone. A significant decrease in the percent of apoptotic osteoblasts and osteocytes was found in the Col2.3Bcl-2 tg/+ calvaria than the +/+ (*p < 0.05). (C) Light micrographs of the vehicle and dexamethasone-treated calvaria are shown with a DNase positive control and a no TdT negative control. Bar = 30 μm. (D) Graph of the percent apoptotic cells determined by TUNEL in 2-month-old male and female femurs of wildtype and Col2.3Bcl-2 mice. A significant decrease of apoptosis (*p < 0.05) in the osteoblasts from male and female transgenic femurs is shown compared with the wildtype control male and female littermates. (E) Light micrographs of femurs from female wildtype and Col2.3Bcl-2 trabeculae at the growth plate show the differences in basal rates of apoptosis. Dark arrows are TUNEL+ osteoblasts and empty arrows are TUNEL osteoblasts. Bar = 20 μm.

Download figure to PowerPoint

Estrogen regulation of Bcl-2

Because the Col2.3Bcl-2 transgenic mice did not show by histomorphometry the sex-related bone phenotype as strongly as the wildtype mice, we studied the effect of estrogen on endogenous mouse Bcl-2 in ex vivo osteoblast-like cultures by RT-PCR and qualitative real-time PCR. After 24 h of administration of 10−8, 10−7, and 10−6 M 17β-estradiol, a dose-dependent effect on mouse Bcl-2 levels (mBcl-2) was found. All three doses stimulated mBcl-2 message levels compared with untreated and 10−8 M 17α-estradiol-treated groups, with the 10−8 M 17β-estradiol producing the maximal stimulation (Fig. 10A). The qualitative real-time PCR confirmed the above results, and after normalization to 18S, 10−8 M 17β-estradiol produced a maximal increase of 17-fold in mBcl-2 levels (Fig. 10B). There was no significant difference in mBcl-2 levels between untreated (c) and 10−8 M 17α-estradiol groups. Thus, estrogen seems to be one of the regulators of endogenous Bcl-2 levels in mouse osteoblasts.

thumbnail image

Figure FIG. 10.. (A) RT-PCR of estrogen stimulation of mouse Bcl-2 (mBcl-2) in ex vivo cultures of wildtype osteoblasts. A dose-dependent stimulation of mBcl-2 message levels was found with 10−8, 10−7, and 10−6 M 17β-estradiol compared with untreated (c) and 10−8 M 17α-estradiol-treated groups. (B) Quantitative real-time PCR of mBcl-2 normalized to 18S and expressed as a relative ratio of mBcl-2/18S.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  7. Acknowledgements

To begin to understand the role of Bcl-2 and apoptosis in bone development and remodeling, we established and characterized a transgenic mouse model that expresses human Bcl-2 protein in osteoblasts. Our model was created in a background of CD-1 outbred mice because previous studies have shown the variability in the bone phenotypes (e.g., density and length) between inbred strains of mice.(29) Using the CD-1 strain more closely mimics the natural population, and therefore, any significant bone phenotype changes in such outbred animals would more likely reflect changes caused by the transgene expression and not reflect mouse strain specific responses. Heterozygote Col2.3Bcl-2 founder lines were established and they appeared healthy; however, they were significantly smaller than the corresponding nontransgenic littermates (+/+) at 1, 2, and 6 months of age. Measurements of body length and femur length revealed an increase in femur and body length from 1 to 2 months, but no significant difference between 2- and 6-month-old wildtype mice. Therefore, the 2-month-old time-point was chosen to study the role of Bcl-2 on bone formation and remodeling, when bone formation is much more active than at 6 months. Histomorphometric analysis of the wildtype mice and transgenic mice at 2 months revealed sex-related differences in bone parameters in the wildtype, which were minimized in the Col2.3Bcl-2 mice. Estrogen was shown to increase mouse Bcl-2 message levels in osteoblast cultures from the wildtype mice. Therefore, the genetic manipulation of Bcl-2 levels seemed to have affected the sex-related differences in the skeleton. To understand further what role Bcl-2 has in affecting osteoblast activity, we studied the effect of apoptosis and Bcl-2 on proliferation, differentiation, and mineralization in ex vivo cultures. These studies showed an effect of Bcl-2 expression on mineralization but not proliferation. The expression of human Bcl-2 in osteoblasts decreased mineralization at 28 days of culture.

Human Bcl-2 expression was found mainly in active osteoblasts on the bone surface because of the DNA construct that uses the 2.3-kb region of the type I collagen promoter (Col2.3). Fragments of the rat Col1a1 promoter have been shown to be preferentially expressed in different type I collagen-producing tissues.(30) In transgenic mice, Col3.6 directed transgene expression in bone, teeth, and nonosseous tissues, such as tendon, skin, and lung. Col2.3 was more restricted in its expression with strong activity in bones, very low activity in tendon, and almost undetectable activity in other tissues.(31) The Col2.3 and Col3.6 promoters have been shown to reflect different stages of differentiation in the osteoblast lineage. Characteristics of pre-osteoblasts are the expression of alkaline phosphatase (ALP) and type I collagen mRNA.(32) In vitro, activation of the Col3.6 promoter has been shown to coincide with pre-osteoblasts.(24) The next stage in the progression of osteoblast differentiation is characterized by bone sialoprotein expression, cuboidal cell morphology, and was consistent with the activation of the 2.3-kb collagen promoter restricted to cells within differentiated nodules in bone forming cultures in vitro.(24, 25, 33, 34) We chose to use the Col2.3 promoter because it is active in functional osteoblasts expressing collagen. Col2.3Bcl-2 osteoblasts stained for human Bcl-2 by immunohistochemistry at the growth plate, endosteal surface, and areas of bone formation on the periosteal surface of cortical bone. Western blots also showed that human specific Bcl-2 was present only in type I collagen-containing tissues: skin, tendon, and bone.

Both male and female transgenic mice had shorter femurs and body lengths than those of the wildtype mice, suggesting that endochondral bone lengthening was reduced. Chondrocytes did not show any significant staining for hBcl-2. Therefore, it is unlikely that cartilage growth was affected. No major differences in the width of the growth plate were apparent. At 2 months of age, calvaria showed more significant differences in bone mass compared with femurs, with a decrease in the width of the calvaria in transgenic mice. This finding in calvaria also correlates well with the transgene expression, which was greatest in calvaria compared with femurs. We hypothesized that a possible explanation for the smaller transgenic mice may come from Bcl-2's ability to affect either cell proliferation or differentiation. Bcl-2 has been shown to diminish proliferation of fibroblasts transfected by retrovirus with murine Bcl-2.(35) Alternatively, in transgenic mice in which human Bcl-2 overexpression driven by the myosin heavy chain promoter, Bcl-2 stimulated proliferation of myocytes.(36) Bcl-2 was also shown to prolong the lifespan of epithelial cells and allow them to differentiate and mature.(29) In the immune system, Bcl-2 was able to repress cell death and promote lymphocyte differentiation.(37) Our ex vivo data showed that Bcl-2 had no effect on proliferation but decreased mineralization in Col2.3Bcl-2 osteoblast cultures compared with results with osteoblasts from wildtype littermates. In vivo growth data showed that the transgenic mice continued to grow after 2 months, whereas there was little to no growth in the wildtype mice, suggesting that there is a delay of growth with hBcl-2 expression. This delay may be explained by our data showing that apoptosis is required for osteoblast bone formation in vitro and in vivo. Bone-like tissue formation in rat osteoblast cultures were shown to have increased TUNEL+ cells, increased Bax, and decreased Bcl-2 levels, showing an increase in apoptosis, particularly in the mineralizing nodules,(38) which we have confirmed in our ex vivo mouse cultures. At bone remodeling sites, 50–70% of the osteoblasts die by apoptosis after they have completed their bone forming function.(39–41) Therefore, in a growing animal, apoptosis may be needed for the skeleton to reach its normal size. However, histomorphometric evaluation of 6-month-old mice, when the transgenic mouse femur is comparable in length with the wildtype femur, revealed that the age-induced loss of trabecular bone was not as great in tg/+, resulting in more percent trabecular bone area in the tg/+ than the +/+. Thus, cell renewal of osteoblasts on the bone surface seems to affect skeletal growth and the age-induced decrease in trabecular bone.

The prevention of apoptosis by Bcl-2 overexpression during skeletal growth may slow the turnover of osteoblasts on the bone surface resulting in a slower rate of bone formation, which was confirmed by our BFR and MAR data. In addition, the expression of human Bcl-2 in mature osteoblasts increased the percentage of osteoblasts on the bone surface without a similar effect on osteoclast parameters. This uncoupling of bone remodeling may also contribute to the decrease in bone formation and bone size found in the transgenic mice. The decrease in bone formation in the transgenic mice (through Bcl-2) may also be caused by the decreased activity of osteoblasts through mechanisms that regulate the renewal of active (collagen producing) cells. Ex vivo data of osteoblast cultures correlated with the in vivo data and showed that mineralization was decreased in the cultures from tg/+ mice compared with the +/+ cultures. Interestingly, previous work in our laboratory with calvaria organ culture showed that mineralization was increased in those cultures where cell death was prevalent.(42) Lynch et al.(38) have also shown that apoptosis was increased during osteoblast differentiation and bone formation in vitro.

In wildtype mice, the sex-related differences in bone parameters were also affected by Bcl-2 levels. Female wildtype mice had increased osteoblast and osteoclast activity, which resulted in a higher bone formation rate than wildtype males; however, the expression of Col2.3Bcl-2 resulted in loss of the sex differences between male and female tg/+ mice. Our finding that estrogen stimulates the transcript levels of Bcl-2 suggests that overexpression of Bcl-2 abrogates the sex hormone-related differences in bone turnover. In contrast to male +/+ mice, which have fewer Ob/S compared with female +/+ mice, overexpression of Bcl-2 caused the male and female transgenic mice to have a similar percent of bone surface covered by active osteoblasts. A comparison of sex differences in femurs of transgenic mice confirmed the osteoblast data because tg/+ males had significantly greater percentage of trabecular bone area than tg/+ females. Males maintained their bone in the calvaria, whereas female tg/+ mice had a significant decrease in CWi compared with +/+ mice. This resulted in a significant decrease in CWi for Col2.3Bcl.2 mice compared with wildtype mice. When we looked at differences based on sex, the male +/+ and male tg/+ mice had no difference in BFR. We hypothesize that this was caused by the increase in percentage ObS/BS in the tg/+ males without a concomitant decrease in osteoclasts. This increase in %ObS/BS was able to offset the negative effect of hBcl-2, which prolongs the life of the osteoblasts on the bone surface and does not allow normal cell turnover. In the females, there was a significant decrease in BFR in the tg/+ compared with +/+ female mice. The female tg/+ mice did not show an increase in %ObS/BS, so they were not able to offset the decrease in turnover of the osteoblasts resulting in a slower rate of bone formation. We speculate that estrogen levels may be responsible for the increased percentage of osteoblast surface and activity in the wildtype female mice, because estrogen has been shown to regulate Bcl-2 mRNA and protein levels in hippocampal neurons,(9) cerebral ischemia,(43) splenic B cells,(44) and human MCF-7 breast cancer cells.(11) In addition, transcriptional activation of Bcl-2 by 17β-estradiol has been shown in breast cancer cells.(11, 12) In one case, two estrogen responsive elements were identified within the Bcl-2 coding region,(11) and in another case, 17β-estradiol was shown to interact with the Bcl-2 gene promoter through multiple enhancer elements.(12) We have previously shown that estrogen is able to block glucocorticoid and lipopolysaccharide-induced apoptosis by preventing the decrease in Bcl-2.(19) In addition, we showed that estrogen regulates Bcl-2 message levels in ex vivo cultures of wildtype osteoblast cultures. Therefore, estrogen may be able to maintain a greater number of osteoblasts on the bone surface in the wildtype females because of its stimulation of Bcl-2, but when overexpressed with the introduction of human Bcl-2, the lack of osteoblast turnover on the bone surface of the transgenic females leads to a decrease in bone formation. Therefore, for normal bone formation to occur, precise control of apoptosis is necessary, with Bcl-2 playing an important role.

In mature bone cells, estrogen causes osteoclast apoptosis suppressing bone resorption, whereas estrogen promotes bone formation by increasing osteoblast survival.(45, 46) In males, testosterone-mediated bone resorption is inhibited mostly by aromatization of testosterone to estrogen, which in turn acts through the estrogen receptor (ER) and is an important pathway for normal growth in male rats.(47) As such, estrogen and testosterone have been shown to affect the lifespan of both osteoblasts and osteoclasts by regulating apoptosis. However, recent data on the effects of estrogen and androgens on apoptosis in vitro did not show sex hormone specificity. Osteocytes, embryonic fibroblasts, and HeLa cells were protected from etoposide-induced apoptosis through the action of sex steroids on the Src/Shc/ERK signaling pathway.(45) Nuclear targeting of ERα, ERβ, and AR were not required for this anti-apoptotic effect. In addition, the anti-apoptotic effects of estrogen could be abrogated by the androgen receptor antagonist flutamide, and conversely, the estrogen receptor antagonist ICI 182,780 was able to block the effects of 5α-dihydrotestosterone. Thus, estrogens and androgens may equally prevent etoposide-induced apoptosis through a “nongenotropic” pathway; however, engagement of the classical sex hormone receptors have sex-specific effects on bone remodeling in vivo. Because the skeleton of males and females is different, our study with the Col2.3Bcl-2 mice suggests that sex hormones may affect differences in bone formation through Bcl-2. The phenotype of the Col2.3Bcl-2 suggests that altering the survival potential of osteoblasts affected the differences in bone remodeling found between sexes, and decreased the bone formation rate and volume in the females.

According to the present histomorphometric data from Col2.3Bcl-2 mice, hBcl-2 increased the number of a specific population of osteoblasts on the bone surface, probably by increasing their survival potential. In ex vivo cultures from the Col2.3Bcl-2 and wildtype mice, expression of human Bcl-2 blocked glucocorticoid-induced apoptosis. However it had no effect on cell proliferation but decreased mineralization. Thus, both ex vivo and in vivo data suggest that apoptosis is a requirement for normal bone formation. In addition, our study is the first to show that differences in bone phenotype related to the sex of the animal are affected by Bcl-2, which involves estrogen's ability to regulate Bcl-2 levels. These data further substantiate a central role of Bcl-2 in osteoblast apoptosis, which was discovered in our previous work,(19) and suggest that the regulation of apoptosis is not only important in bone formation but also in maintaining bone during the aging process.


  1. Top of page
  2. Abstract
  7. Acknowledgements
  • 1
    Adams JM, Cory S 2001 Life-or-death decisions by the Bcl-2 protein family. Trends Biochem Sci 26: 6166.
  • 2
    Antonsson B, Martinou JC 2000 The Bcl-2 protein family. Exp Cell Res 256: 5057.
  • 3
    Desagher S, Martinou JC 2000 Mitochondria as the central control point of apoptosis. Trends Cell Biol 10: 369377.
  • 4
    Gross A 2001 BCL-2 proteins: Regulators of the mitochondrial apoptotic program. IUBMB Life 52: 231236.
  • 5
    Lagasse E, Weissman IL 1997 Enforced expression of Bcl-2 in monocytes rescues macrophages and partially reverses osteopetrosis in op/op mice. Cell 89: 10211031.
  • 6
    Martinou JC, Dubois-Dauphin M, Staple JK, Rodriguez I, Frankowski H, Missotten M, Albertini P, Talabot D, Catsicas S, Pietra C, Joaquin H 1994 Overexpression of BCL-2 in transgenic mice protects neurons from naturally occurring cell death and experimental ischemia. Neuron 13: 10171030.
  • 7
    Farlie PG, Dringen R, Rees SM, Kannourakis G, Bernard O 1995 bcl-2 transgene expression can protect neurons against developmental and induced cell death. Proc Natl Acad Sci USA 92: 43974401.
  • 8
    Lacronique V, Mignon A, Fabre M, Viollet B, Rouquet N, Molina T, Porteu A, Henrion A, Bouscary D, Varlet P, Joulin V, Kahn A 1996 Bcl-2 protects from lethal hepatic apoptosis induced by an anti-Fas antibody in mice. Nat Med 2: 8086.
  • 9
    Nilsen J, Diaz Brinton R 2003 Mechanism of estrogen-mediated neuroprotection: Regulation of mitochondrial calcium and Bcl-2 expression. Proc Natl Acad Sci USA 100: 28422847.
  • 10
    Bhavnani BR 2003 Estrogens and menopause: Pharmacology of conjugated equine estrogens and their potential role in the prevention of neurodegenerative diseases such as Alzheimer's. J Steroid Biochem Mol Biol 85: 473482.
  • 11
    Perillo B, Sasso A, Abbondanza C, Palumbo G 2000 17beta-estradiol inhibits apoptosis in MCF-7 cells, inducing bcl-2 expression via two estrogen-responsive elements present in the coding sequence. Mol Cell Biol 20: 28902901.
  • 12
    Dong L, Wang W, Wang F, Stoner M, Reed JC, Harigai M, Samudio I, Kladde MP, Vyhlidal C, Safe S 1999 Mechanisms of transcriptional activation of bcl-2 gene expression by 17beta-estradiol in breast cancer cells. J Biol Chem 274: 3209932107.
  • 13
    Schorr K, Li M, Krajewski S, Reed JC, Furth PA 1999 Bcl-2 gene family and related proteins in mammary gland involution and breast cancer. J Mammary Gland Biol Neoplasia 4: 153164.
  • 14
    Razandi M, Pedram A, Levin ER 2000 Plasma membrane estrogen receptors signal to antiapoptosis in breast cancer. Mol Endocrinol 14: 14341447.
  • 15
    Veis DJ, Sorenson CM, Shutter JR, Korsmeyer SJ 1993 Bcl-2-deficient mice demonstrate fulminant lymphoid apoptosis, polycystic kidneys, and hypopigmented hair. Cell 75: 229240.
  • 16
    Nakayama K, Negishi I, Kuida K, Sawa H, Loh DY 1994 Targeted disruption of Bcl-2 alpha beta in mice: Occurrence of gray hair, polycystic kidney disease, and lymphocytopenia. Proc Natl Acad Sci USA 91: 37003704.
  • 17
    Michaelidis TM, Sendtner M, Cooper JD, Airaksinen MS, Holtmann B, Meyer M, Thoenen H 1996 Inactivation of bcl-2 results in progressive degeneration of motoneurons, sympathetic and sensory neurons during early postnatal development. Neuron 17: 7589.
  • 18
    Boot-Handford RP, Michaelidis TM, Hillarby MC, Zambelli A, Denton J, Hoyland JA, Freemont AJ, Grant ME, Wallis GA 1998 The bcl-2 knockout mouse exhibits marked changes in osteoblast phenotype and collagen deposition in bone as well as a mild growth plate phenotype. Int J Exp Pathol 79: 329335.
  • 19
    Gohel A, McCarthy MB, Gronowicz G 1999 Estrogen prevents glucocorticoid-induced apoptosis in osteoblasts in vivo and in vitro. Endocrinology 140: 53395347.
  • 20
    Cleary ML, Smith SD, Sklar J 1986 Cloning and structural analysis of cDNAs for bcl-2 and a hybrid bcl-2/immunoglobulin transcript resulting from the t(14;18) translocation. Cell 47: 1928.
  • 21
    Lichtler A, Stover ML, Angilly J, Kream B, Rowe DW 1989 Isolation and characterization of the rat alpha 1(I) collagen promoter. Regulation by 1,25-dihydroxyvitamin D. J Biol Chem 264: 30723077.
  • 22
    Lichtler A, Stover ML, Angilly J, Kream B, Rowe DW 1994 Isolation and characterization of the rat alpha 1(I) collagen promoter. Regulation by vitamin D. J Biol Chem 269: 16518.
  • 23
    Krebsbach PH, Harrison JR, Lichtler AC, Woody CO, Rowe DW, Kream BE 1993 Transgenic expression of COL1A1-chloramphenicol acetyltransferase fusion genes in bone: Differential utilization of promoter elements in vivo and in cultured cells. Mol Cell Biol 13: 51685174.
  • 24
    Kalajzic Z, Liu P, Kalajzic I, Du Z, Braut A, Mina M, Canalis E, Rowe DW 2002 Directing the expression of a green fluorescent protein transgene in differentiated osteoblasts: Comparison between rat type I collagen and rat osteocalcin promoters. Bone 31: 654660.
  • 25
    Visnjic D, Kalajzic I, Gronowicz G, Aguila HL, Clark SH, Lichtler AC, Rowe DW 2001 Conditional ablation of the osteoblast lineage in Col2.3deltatk transgenic mice. J Bone Miner Res 16: 22222231.
  • 26
    Singh I 1969 A modification of the Masson-trichrome stain. Indian J Pathol Bacteriol 12: 172173.
  • 27
    Parfitt AM, Drezner MK, Glorieux FH, Kanis JA, Malluche H, Meunier PJ, Ott SM, Recker RR 1987 Bone histomorphometry: Standardization of nomenclature, symbols, and units. Report of the ASBMR Histomorphometry Nomenclature Committee. J Bone Miner Res 2: 595610.
  • 28
    Wong GL, Cohn DV 1975 Target cells in bone for parathormone and calcitonin are different: Enrichment for each cell type by sequential digestion of mouse calvaria and selective adhesion to polymeric surfaces. Proc Natl Acad Sci USA 72: 31673171.
  • 29
    Beamer WG, Donahue LR, Rosen CJ, Baylink DJ 1996 Genetic variability in adult bone density among inbred strains of mice. Bone 18: 397403.
  • 30
    Pavlin D, Bedalov A, Kronenberg MS, Kream BE, Rowe DW, Smith CL, Pike JW, Lichtler AC 1994 Analysis of regulatory regions in the COL1A1 gene responsible for 1,25-dihydroxyvitamin D3-mediated transcriptional repression in osteoblastic cells. J Cell Biochem 56: 490501.
  • 31
    Bogdanovic Z, Bedalov A, Krebsbach PH, Pavlin D, Woody CO, Clark SH, Thomas HF, Rowe DW, Kream BE, Lichtler AC 1994 Upstream regulatory elements necessary for expression of the rat COL1A1 promoter in transgenic mice. J Bone Miner Res 9: 285292.
  • 32
    Cowles EA, DeRome ME, Pastizzo G, Brailey LL, Gronowicz GA 1998 Mineralization and the expression of matrix proteins during in vivo bone development. Calcif Tissue Int 62: 7482.
  • 33
    Stover ML, Wang CK, McKinstry MB, Kalajzic I, Gronowicz G, Clark SH, Rowe DW, Lichtler AC 2001 Bone-directed expression of Col1a1 promoter-driven self-inactivating retroviral vector in bone marrow cells and transgenic mice. Mol Ther 3: 543550.
  • 34
    Kalajzic I, Kalajzic Z, Kaliterna M, Gronowicz G, Clark SH, Lichtler AC, Rowe D 2002 Use of type I collagen green fluorescent protein transgenes to identify subpopulations of cells at different stages of the osteoblast lineage. J Bone Miner Res 17: 1525.
  • 35
    Borner C 1996 Diminished cell proliferation associated with the death-protective activity of Bcl-2. J Biol Chem 271: 1269512698.
  • 36
    Limana F, Urbanek K, Chimenti S, Quaini F, Leri A, Kajstura J, Nadal-Ginard B, Izumo S, Anversa P 2002 bcl-2 overexpression promotes myocyte proliferation. Proc Natl Acad Sci USA 99: 62576262.
  • 37
    Linette GP, Korsmeyer SJ 1994 Differentiation and cell death: Lessons from the immune system. Curr Opin Cell Biol 6: 809815.
  • 38
    Lynch MP, Capparelli C, Stein JL, Stein GS, Lian JB 1998 Apoptosis during bone-like tissue development in vitro. J Cell Biochem 68: 3149.
  • 39
    Jilka RL, Weinstein RS, Bellido T, Parfitt AM, Manolagas SC 1998 Osteoblast programmed cell death (apoptosis): Modulation by growth factors and cytokines. J Bone Miner Res 13: 793802.
  • 40
    Weinstein RS, Manolagas SC 2000 Apoptosis and osteoporosis. Am J Med 108: 153164.
  • 41
    Manolagas SC 2000 Birth and death of bone cells: Basic regulatory mechanisms and implications for the pathogenesis and treatment of osteoporosis. Endocr Rev 21: 115137.
  • 42
    Gronowicz G, Woodiel FN, McCarthy MB, Raisz LG 1989 In vitro mineralization of fetal rat parietal bones in defined serum-free medium: Effect of beta-glycerol phosphate. J Bone Miner Res 4: 313324.
  • 43
    Dubal DB, Shughrue PJ, Wilson ME, Merchenthaler I, Wise PM 1999 Estradiol modulates bcl-2 in cerebral ischemia: A potential role for estrogen receptors. J Neurosci 19: 63856393.
  • 44
    Bynoe MS, Grimaldi CM, Diamond B 2000 Estrogen up-regulates Bcl-2 and blocks tolerance induction of naive B cells. Proc Natl Acad Sci USA 97: 27032708.
  • 45
    Kousteni S, Bellido T, Plotkin LI, O'Brien CA, Bodenner DL, Han L, Han K, DiGregorio GB, Katzenellenbogen JA, Katzenellenbogen BS, Roberson PK, Weinstein RS, Jilka RL, Manolagas SC 2001 Nongenotropic, sex-nonspecific signaling through the estrogen or androgen receptors: Dissociation from transcriptional activity. Cell 104: 719730.
  • 46
    Manolagas SC, Kousteni S, Jilka RL 2002 Sex steroids and bone. Recent Prog Horm Res 57: 385409.
  • 47
    Vanderschueren D, van Herck E, Nijs J, Ederveen AG, De Coster R, Bouillon R 1997 Aromatase inhibition impairs skeletal modeling and decreases bone mineral density in growing male rats. Endocrinology 138: 23012307.