Tartrate-resistant acid phosphatase (TRAP) is a secreted product of osteoclasts and a lysosomal hydrolase of some tissue macrophages. To determine whether TRAP expression is rate-limiting in bone resorption, we overexpressed TRAP in transgenic mice by introducing additional copies of the TRAP gene that contained the SV40 enhancer. In multiple independent mouse lines, the transgene gave a copy number–dependent increase in TRAP mRNA levels and TRAP activity in osteoclasts, macrophages, serum, and other sites of normal low-level expression (notably, liver parenchymal cells, kidney mesangial cells, and pancreatic secretory acinar cells). Transgenic mice had decreased trabecular bone consistent with mild osteoporosis. Measurements of the bone formation rate suggest that the animals compensate for the increased resorption by increasing bone synthesis, which partly ameliorates the phenotype. These mice provide evidence that inclusion of an irrelevant enhancer does not necessarily override a tissue-specific promoter.
The presence of tartrate-resistant acid phosphatase (TRAP; E.C. 220.127.116.11) activity has been used extensively as a marker for terminally differentiated bone-resorbing osteoclasts. TRAP activity has been localized to the lysosomes and resorptive hemivacuole of osteoclasts and to the lysosomes of a subset of macrophages.(1) TRAP also is localized to the lysosome in a range of other cell types outside the mononuclear phagocyte system and has been implicated as the major enzyme responsible for cleavage of the mannose-6-phosphate lysosomal targeting signal in these cells.(2) The involvement of TRAP in osteoclastic bone resorption was suggested by the reduced resorptive activity of cultured primary osteoclasts observed when the cells were treated with inhibitors of TRAP activity, including specific antibodies.(3) More recently, targeted disruption of the TRAP gene has provided clear evidence that the enzyme is essential for normal bone homeostasis in vivo.(4)
The gene for the porcine TRAP homolog, uteroferrin,(5) as well as the mouse TRAP gene, including 1.8 kb of 5′ flanking DNA(6) and the human TRAP gene sequences,(6,7) have been described. Some limited analysis of the mouse TRAP promoter was performed when it was first sequenced.(6,8) Expression of TRAP mRNA and secretion of TRAP protein is regulated in terminally differentiated osteoclasts by agents that alter the intracellular Ca2+ concentration. These agents include phorbol esters, calcium ionophores, and elevated extracellular Ca2+ concentration.(9,10) In murine macrophages, TRAP mRNA is reportedly inducible by the cytokine interleukin 4.(11) Subsequently, Alcantara et al.(12) observed that the mouse TRAP promoter was inducible in the rabbit endometrial cell line HRE H9 by iron-saturated transferrin. This observation was correlated with the finding that the TRAP induction that occurs when monocytes are cultured in vitro could be prevented by desferral, an iron chelator.(12) By contrast, protoporphyrin IX and hemin both inhibited the activity of the TRAP promoter, acting through a novel, tandemly repeated GAGGC motif.(13)
Clearly, all of the available model systems are poor substitutes for analysis of the regulation and function of the TRAP gene in osteoclasts because these cells are its major site of expression. Our two aims in the present study were to determine whether the currently known mouse genomic DNA sequences are sufficient to direct tissue-specific expression, and to create a mouse transgenic model in which to study the role of TRAP via its overexpression from additional copies of the TRAP gene inserted into the mouse germ line. In an attempt to overcome position and copy-number dependence of the transgene, we included the SV40 early enhancer within intron 3 of the gene. The resulting transgenic mice displayed highly tissue-restricted overexpression of TRAP, confirming that all the elements required for tissue-specific expression were present. Preliminary studies indicate that overexpression of the gene in bone causes an increase in resorption rate that is partly compensated by an increase in rate of bone formation, yielding a net increase in bone turnover.
MATERIALS AND METHODS
Cell lines and culture
The RAW 264.7 macrophage cell line was obtained from the ATCC and cultured in RPMI-1640 with 10% fetal calf serum.
Cultivation of osteoclast-like cells
Primary osteoblasts were extracted from the calvaria of neonatal BALB/C mice and cultured in minimum essential medium, alpha modification (α-MEM; Life Technologies, Melbourne, Australia), that contained 10% fetal calf serum. Bone-marrow cells were prepared from BALB/C mice more than 4 weeks old. Osteoclast-like cells were differentiated in vitro by the coculture of calvarial osteoblasts with splenocytes or bone marrow cells. Osteoclast cells were differentiated from murine femoral bone marrow in culture. The bone marrow cells were flushed out of the femur and plated on calvarial osteoblast monolayers grown on gelatin-coated plates.(14) Cells were grown for 7 days in 6-well tissue culture plates in Dulbecco's modified Eagle medium (DMEM; Life Technologies) with 15% heat-inactivated fetal bovine serum, 2 mM l-glutamine, 20 U penicillin/ml, and 20 μg/ml streptomycin supplemented with 10 nM 1α,25(OH)2 vitamin D3at 37°C in humidified air that contained 5% CO2. The medium was replaced on days 3 and 6, and adherent osteoclasts were harvested on day 7 by treatment with 0.1% collagenase/0.2% dispase in phosphate-buffered saline (PBS) to remove the gelatin layer that contained the osteoblasts. The osteoclast population was further enriched by incubation with 0.25% trypsin in PBS for 5 minutes to remove residual osteoblasts.
RNA extraction and Northern blot analysis
Total cell RNA was isolated by using a modified version of Chomczynski and Sacchi's(15) method of acid guanidinium isothiocyanate phenol-chloroform extraction. Ten micrograms of total RNA was subjected to electrophoresis on a 1.2% agarose gel under denaturing conditions. The RNA was transferred to Hybond N membrane (Amersham, Sydney, Australia). The filter was probed with a32P-labeled 371–base pair (bp) fragment of the murine TRAP cDNA coding region; washed with 0.5× standard saline citrate (SSC) and 0.1% sodium dodecyl sulfate (SDS) at 65°C; and exposed to Fujifilm-RX with two intensifying screens at− 70°C. The 18S ribosomal RNA was detected by hybridization with a 5′-32P-labeled oligonucleotide complementary to this RNA.
Transient transfection and luciferase assays
Cells for transient transfection were replated in fresh medium the day before transfection, which was performed under conditions described previously.(16) After transfection, the cells were incubated in RPMI-1640 that contained 10% heat-inactivated fetal calf serum for 24 h. The cells were thoroughly washed with PBS and lysed in 1× Lysis solution (Promega, Annandale, Australia). Cell extracts were incubated for 15 minutes at room temperature, then centrifuged at 15,000gfor 5 minutes to remove cellular debris. Aliquots of cell lysate were assayed for luciferase activity in 100 μl of luciferase assay reagent (Promega) by using a Bertoldt Luminometer. The protein concentration of the extracts was assayed by using the micro-Bradford technique (Biorad, Regents Park, Australia).
Reporter plasmid construction
The TRAP promoter from the 5′ end to the splice acceptor site of intron 1 was isolated and, with the TRAP ATG, was used to create luciferase reporter plasmids controlled by the TRAP promoter. The region immediately 3′ to the ATG was mutated by polymerase chain reaction (PCR) to generate a HindIII site (5′-A+1TGGATTCAT to 5′-A+1TGGAAGCTT), and then the full-length promoter to this new site was subcloned into pGL2–Basic (Promega) to yield pGL2–MAP5.
Generation of transgenic mice containing the TRAP-SV40E construct
To generate the transgene, a 6585-bp TRAP gene fragment was isolated(6) and inserted into the pBluescript IISK to yield pTRAP. The SV40 enhancer (SV40E) was isolated as a HpaI/BamI fragment from pGL2-Control (Promega), and after blunting of the ends with T4 DNA polymerase, the fragment was ligated into the EcoRV site in intron 3 of the TRAP gene in pTRAP to yield pTRAP-SV40E that contained two copies of the SV40E. The SalI/SacI-linearized pTRAP-SV40E DNA was injected into the pronuclei of fertilized mouse embryos at a concentration of 2 ng/ml, and embryos were transferred to pseudo-pregnant mice.(17) The resulting transgenic animals were screened by probing a Southern blot of PstI-digested mouse genomic DNA with the unique SV40E sequence.
Brain, spleen, kidney, liver, heart, pancreas, and lung tissues were extracted and immediately frozen in O.C.T. embedding medium (Miles Inc., Elkhart, IN, U.S.A.). Cryostat sections (8 μm thick) were cut for use in enzyme histochemistry. The sections were fixed and stained for TRAP activity (Sigma kit 386-A, Sigma Chemical Co., St. Louis, MO, U.S.A.). Sections were counterstained with methyl green and mounted in gelatin/glycerol medium.
Preparation of osteoclast lysates
Adherent osteoclasts were lysed in 100 μl of MilliQ H2O for 5 minutes, and the osteoclast lysate was harvested from the wells and centrifuged at 15,000g for 5 minutes to remove cellular debris. Osteoclast lysate supernatant was stored at− 70°C until assayed.
Serum isolation and TRAP activity assays
Mouse serum was obtained from transgenic and wild-type mice by extraction of 1 ml of whole blood through heart puncture. The blood was clotted at room temperature for 30 minutes, then centrifuged at 500g for 10 minutes. The serum was removed and frozen at −70°C until assayed. TRAP assays were performed on serum and osteoclast lysates in 96-well microtitre plates and were recorded as the change in absorbance at 415 nm due to the cleavage of p-nitrophenol phosphate by the enzyme.(11,18)
Bone histology and osteoclast number analysis
Tibiae from both wild-type and TRAP-SV40E mice were fixed for 24 h in 10% Formalin and decalcified in 10% ethylenediaminetetraacetic acid (EDTA) for 7 days. The bones were dehydrated, cleared using petroleum spirit, and embedded in paraffin wax. Midsagittal sections (6 μm) were incubated in 0.1 M Tris-HCl (pH 9.0) for 18 h and transferred to 0.1 M sodium citrate (pH 5.2) for 3 h before staining for TRAP activity.(19) Activated sections were reacted for TRAP by using the hexasotized para-rosaniline method(20,21) with naphthol AS-TR phosphate (disodium salt; Sigma) as the substrate in the presence of 1 mg/ml sodium tartrate. The sections were counterstained with 0.2% methyl green and mounted in DPX (BDH Ltd., Poole, U.K.). To determine representative osteoclast cell numbers in each section, photographs were taken across the proximal tibiae with ×25 magnification, and image analysis was completed by using Adobe Photoshop (version 5.0). Ten fields (160 μm × 160 μm each) were systematically chosen across the region of the secondary spongiosa and analyzed for the presence of osteoclasts, which were defined as TRAP-positive cells that were apposed to trabecular surfaces. The multinuclearity of these cells was not taken into consideration because of the thickness of the sections.
Bone turnover studies
Mice were injected (10 mg/kg body weight, ip) with labeled calcein(22) (Sigma) on days 1 and 4 and were culled on day 5. Tibia samples were removed and fixed in 10% buffered Formalin for 16 h, then embedded in methylmethacrylate.(23) Sagittal sections of 80 μm were cut on a diamond saw (O-well Ahlberg Technical Equipment Co., U.S.A.) through the proximal tibia. Labeled calcein was visualized by using a fluorescence microscope (Olympus BX-60). Labeled bone was analyzed with the KSS stereology package (KSS, Salt Lake City, Utah, U.S.A.). Histomorphometric measurements of the bone surface included trabecular area, expressed as a percentage of the total bone area (Tb.Ar, %); double-labeled or mineralizing surface (dLS/BS, %), measured as the length of labeled surface divided by the total bone surface; mineral appositional rate (MAR, μm/day), obtained by dividing the distance between labels by time; and bone formation rate (BFR/B.Ar, μm · μm−2 · year−1), the product of MAR and dLS/BS. Bone formation rate per unit area is equivalent to bone turnover rate.(24)
Effect of the SV40 enhancer on the TRAP promoter
Our aim in studying the TRAP gene is to understand the processes that control osteoclast differentiation and activity. Our approach has been to generate transgenic mice in which TRAP itself is overexpressed. Previously, other researchers have used the 2.1-kb TRAP promoter to drive expression of an SV40 large T antigen,(25,26) or to reconstitute the expression of the c-src proto-oncogene in c-src knockout mice.(27) However, these studies did not examine whether the expression was appropriately restricted to specific cells that normally produce TRAP mRNA. One difficulty in such studies is the absence of a reliable, readily detected histochemical reporter gene for macrophages and osteoclasts. As a result, we decided to use the TRAP enzyme itself as a marker. This approach offered the advantage that successful production of transgenic animals would provide a model in which the TRAP gene was overexpressed in a tissue-restricted manner, allowing assessment of its biological function.
The production of transgenes to identify key cis-acting elements of a promoter can be complicated by several requirements. These requirements include the need for large amounts of flanking genomic sequences to isolate the gene from position effects of neighboring genes, and the commonly observed nonlinear relationship between transgene copy number and the level of reporter gene expression. In an attempt to obviate these effects and to amplify transgene expression, we tested the effect of the SV40E on TRAP promoter activity initially using luciferase reporter plasmids. As shown in Fig. 1A, the addition of the SV40E sequence to the native TRAP promoter caused a 3-fold increase in luciferase expression when transfected into the RAW 264.7 macrophage cell line. To determine the effect of SV40E on the TRAP promoter in the context of the whole gene, we have taken a construct containing the entire 7-kb murine TRAP gene (pTRAP) and introduced two copies of SV40E into intron 3 of the TRAP gene to generate the construct, pTRAP-SV40E (Fig. 1B). Previously, we showed that the SV40E is active in cells of the macrophage and B lymphocyte lineages.(28) We examined the enzyme activity of TRAP expressed from the pTRAP and pTRAP-SV40E constructs by cytochemical staining of transiently transfected RAW 264.7 cells. TRAP activity staining was not visibly higher in RAW 264.7 cells transiently transfected with the pTRAP construct (Fig. 1B, center) than the cells transfected with the pBluescript plasmid negative control (Fig. 1B, upper panel). RAW 264.7 cells transfected with pTRAP-SV40E expressed higher levels of histochemically detectable TRAP enzyme activity, as indicated by the stained cells (Fig. 1B, lower panel), than cells transfected with either pTRAP or pBluescript.
Increased expression of TRAP in murine tissues
Initially, 15 independent lines of transgenic mice with detectable integrated copies of the TRAP-SV40E transgene were established. Insertion of the SV40E provided a simple probe to identify the transgenic mice and compare relative copy number of the TRAP-SV40E transgene. Preliminary screening for transgenic animals (by assaying the serum TRAP enzyme activity) showed a clear positive correlation between transgene copy number and enzyme activity (Fig. 2). Five lines, which encompassed low (range 1–10), middle (range 10–30), and high (range 30–50) copy numbers of the transgene, were arbitrarily chosen for detailed analysis.
To assess whether overexpression of the TRAP gene was observed in osteoclasts, we produced osteoclasts by the in vitro coculture system. The production of large multinucleated giant cells in these cultures was not adversely affected by the transgene, indicating that TRAP overexpression does not inhibit growth and differentiation of osteoclasts. Direct assays of enzyme activity in extracts of cultured transgenic osteoclasts showed that TRAP expression was related to the copy number of the inserted transgenes (Fig. 2). Histochemical staining of bone marrow in vivo provided visible evidence of overexpression of the TRAP gene in the osteoclasts but not in residual mononuclear cells that do not normally express the gene (data not shown).
The retention of tissue-specific expression by the transgene outside of the skeletal system was assessed by a combination of Northern analysis and histochemical staining. RNA from a range of murine tissues was probed with labeled TRAP cDNA that had been amplified by PCR, and its identity was confirmed by DNA sequence analysis. A 1.5-kb mRNA species was detected in large amounts in liver, kidney, spleen, and to a lesser degree in heart, brain, and lung (Fig. 3). The level of detectable TRAP mRNA was clearly elevated in all of the tissues of the middle and high copy number transgenics and was in proportion to the transgene copy number. This relativity of TRAP expression between wild-type and transgenic mice was retained in all of the examined tissues (Fig. 3). Cryostat sections of these tissues were stained histochemically to assess the specific distribution of the enhanced TRAP activity. As seen in Fig. 4, expression in the heart, spleen, lung, brain, and kidney (Fig. 4A–4E, respectively) was restricted to macrophage-like cells. In the kidney, TRAP was expressed exclusively in the glomerulus, most probably in mesangial cells (Fig. 4E). Expression in the liver (Fig. 4F) was detectable in the Kupffer cells lining the vascular sinuses, but the major site appeared to be the hepatic parenchymal cells. This site was confirmed by in situ hybridization and histochemical staining of 11.5- and 12.5-day post coitum murine embryos (data not shown). In the pancreas, TRAP was expressed specifically at the highest levels in the secretory acinar cells (Fig. 4G). In each of these sites, the tightly focused increase in activity in the TRAP-SV40E animals supports the view that the enzyme detected histochemically is indeed TRAP and not another acid phosphatase.
Bone turnover in TRAP-SV40E transgenic mice
The effect of tissue-restricted overexpression of TRAP in skeletal metabolism was assessed by histomorphometric analysis. Fluorescent calcein was administered to mature high-copy-number transgenic mice and to wild-type control mice of the same genetic background. Histomorphometric analyses of calcein incorporation into bone were performed on the tibiae. Animals that expressed increased levels of TRAP had significantly decreased levels of trabecular bone relative to their nontransgenic littermates (Table 1). Conversely, the bone formation rate was increased more than 2-fold in the transgenic animals, indicating that increased resorptive activity can be compensated in large measure by an increase in the rate of bone formation. This observation indicates that in TRAP-SV40E transgenic mice, the rate of bone turnover is higher than in their wild-type counterparts. Sections of decalcified bone taken from the tibiae of wild-type and high-copy-number TRAP-SV40E mice revealed no obvious difference in osteoclast numbers; however, the osteoclasts from TRAP-SV40E mice were stained more intensely (Fig. 5). Measurement of osteoclast numbers in secondary spongiosa of tibial sections revealed no significant difference between osteoclast numbers in the wild-type and transgenic samples (Table 2). We infer that the increased bone resorption can be attributed to increased activity of individual osteoclast cells when TRAP activity is increased.
Table Table 1. Histomorphometric Measurements of Bone Turnover Using Calcein Double Labeling of Tibiae
Table Table 2. Osteoclast Numbers Within the Proximal Tibia
TRAP activity has been used for many years as the principal cytochemical marker for osteoclasts, but its role in osteoclast biology is contentious. The enzyme is secreted by osteoclasts into the resorptive space, and enzyme activity reaches its highest level in multinucleated cells actively resorbing bone. TRAP activity is localized to both the lysosomes and the resorptive region of osteoclasts. Recently, it has been shown that TRAP can dephosphorylate mannose-6-phosphate residues in lysosomal proteins(2) after their delivery to the endosome/lysosome compartment. Because TRAP itself contains the mannose-6-phosphate motif, it may regulate its own transport, because premature dephosphorylation may prevent correct targeting. The localization of TRAP to the resorptive space implies that it plays an active role in the process of bone resorption.(25) This proposal is supported by the demonstration that the TRAP inhibitor Mo6+ and anti-TRAP antibodies markedly reduce bone resorption in vitro.(3,10) The strongest support for a role of TRAP in bone resorption comes from TRAP-deficient mice generated by gene knockout.(4) The bones of these mice are decreased in length, with increased mineralization density and a defect in endochondral ossification. We have observed very high levels of TRAP associated with this process in developing mouse embryos (Kathleen Murphy, Department of Microbiology and Parasitology, University of Queensland, Queensland, Australia; unpublished observations). Osteoclasts from TRAP-knockout mice were found to differentiate and form resorption pits in vivo, but their ability to resorb bone was markedly reduced.(4)
In the present study, we took a complementary approach. We investigated whether an increase in TRAP expression would also change osteoclast function. TRAP-SV40E mice exhibited changes in bone structure consistent with changes in both osteoclast and osteoblast function. Additionally, the detectably raised levels of TRAP documented in the transgenic mice provided the first reported system for studying the regulation of an osteoclast-/macrophage-specific gene in vivo. As such, the production of transgenic mice using the entire TRAP gene confirmed that the gene fragment used contains all the information required to generate tissue-restricted expression.
Using these transgenics, we have confirmed the presence of TRAP activity outside of the mononuclear phagocyte system: in the secretory cells of normal liver and pancreatic tissue, and in the mesangial cells of the renal glomerulus. Although TRAP activity has been reported previously in a range of organs, including the liver, cellular expression has not been localized.(29) The TRAP-SV40E transgenic mice, in combination with the previously characterized knockout mice, provide systems in which to study the functions of TRAP outside of macrophages and osteoclasts.
Inclusion of the SV40E did not override tissue specificity of the TRAP promoter, and expression of TRAP was associated with transgene copy number. For our purposes, we generated animals in which the biological role of TRAP could be assessed and a transgene backbone could be systematically mutated to identify the cis-acting elements required for TRAP gene expression.
With respect to the role of TRAP in bone resorption, we showed (Table 1) that TRAP overexpression can cause changes in bone homeostasis without causing a significant change in osteoclast numbers (Table 2). We infer that individual osteoclasts have an increased resorption rate that is compensated in part by a 2-fold increase in the rate bone formation, so there is relatively little net bone loss. There is no overt pathology in young animals, but it remains to be seen how the transgenic lines respond to physiological stresses that alter calcium homeostasis, such as ovariectomy and aging.
The mechanism whereby TRAP overexpression increases bone turnover remains unclear because of the many activities that TRAP has been shown to possess in vitro. These activities include the hydrolysis of pyrophosphate,(30) the dephosphorylation of mannose-6-phosphate residues in lysosomal proteins,(2) the dephosphorylation of bone phosphoproteins such as osteopontin and bone sialoprotein,(31) the regulation of endochondral ossification,(4) and the generation of hydroxyl radicals that may be active in collagen degradation.(32,33) Each of these activities independently is likely to regulate the efficiency with which osteoclasts resorb bone. The effect of TRAP overexpression implies that TRAP is a target for development of anti-osteoporotic drugs. Our recent solution of the crystal structure of mammalian TRAP enzyme from pig(34) should expedite the development of such agents.
This research was supported by National Health and Medical Research Council grants 961167 and 961245. The Centre for Molecular and Cellular Biology is a Special Research Centre of the Australian Research Council.