Abstract
STC1, a mammalian homologue of stanniocalcin (STC) which plays a major role in calcium/phosphate homeostasis in fish, has been recently isolated. We have characterized the spatiotemporal distribution of STC1 mRNA and protein during mouse embryonic development generally and osteogenesis specifically. Northern blotting analysis of whole embryos showed that STC1 mRNA is highly and differentially expressed during embryogenesis. By in situ hybridization, STC1 mRNA was detected early in mesenchymal condensations and was then found to be highly expressed in perichondrial cells, periosteal cells, and then osteoblasts during endochondral bone formation. In bones forming by intramembranous ossification, STC1 mRNA was not detected until osteogenic cells appeared. The cellular distribution of STC1 protein closely corresponded to that of its mRNA, but the protein was also detected in hypertrophic chondrocytes. In the MC3T3-E1 osteogenic cell model, STC1 protein and mRNA were detectable throughout proliferation and differentiation stages but levels were relatively higher late during nodule formation/mineralization phases. For comparison, STC1 mRNA was also found in epithelial cells of both embryonic and adult intestine that had not previously been described among tissues responsive to calcium/phosphate transport. These results suggest that STC1 is expressed in a time- and cell-specific manner and may play an autocrine/paracrine role during osteoblast development and bone formation.
S
Similarly to the activity of fish STC, human STC1 recombinant protein (rhSTC1) is able to regulate calcium uptake in the gill (Zhang et al. 1998). However, in contrast to fish STC, which is expressed only in the specialized CS, human (Chang et al. 1995; Olsen et al. 1996) and mouse (Chang et al. 1996; Varghese et al. 1998) STC1 mRNA has been reported to be expressed in multiple tissues and rat STC1protein has been detected in a number of tissues by radioimmunoassay (De Niu et al. 2000). In addition, cell localization of STC1 mRNA and/or protein has been studied in detail in specific segments of the kidney in humans (De Niu et al. 1998), rats (Hadded et al. 1996; Wong et al. 1998; Worthington et al. 1999), and mice (Yoshiko and Maeda 1998). There is growing evidence that STC1 is able to regulate calcium/phosphate homeostasis in mammals (Wagner et al. 1997; Madsen et al. 1998). For example, rhSTC1 reduces renal phosphate excretion and increases the rate of Na-phosphate co-transport in renal cortical brush-border membrane vesicles of rats (Wagner et al. 1997). rhSTC1 also decreases calcium and increases phosphate absorption in voltage-clamped swine and rat duodenal tissues (Madsen et al. 1998). In addition, STC1 mRNA (Chang et al. 1995,1996) and protein (De Niu et al. 2000) were detected in rodent intestine, although their cellular distribution has not yet been established. Given these observations, STC1 has been hypothesized to play a significant autocrine/paracrine role(s) in mammalian calcium/phosphate homeostasis by reducing net calcium absorption and promoting calcium deposition into bone in the presence of increased plasma phosphate levels (Madsen et al. 1998). A role for STC1 protein in musculoskeletal development is also supported by our recent identification of STC1 mRNA in osteoblasts of developing mice (Yoshiko et al. 1999) and the presence of STC1 protein in developing mouse embryonic muscle and chondrocytes (Jiang et al. 2000).
In this study we sought to describe in detail the spatiotemporal distribution of STC1 mRNA and protein in embryonic mouse osteogenesis and in an in vitro mouse model of osteogenesis, the MC3T3-E1 cell line. Our data indicate that STC1 is already expressed at the stage of mesenchymal condensation in the developing embryo but becomes increasingly restricted to osteoblast lineage cells and mature osteoblasts as development advances. Taken together with the fact that STC1 is also distributed specifically in epithelial cells of intestine and kidney, our data support the view that STC1 may be an autocrine/paracrine factor in tissues responsive to calcium/phosphate transport.
Materials and Methods
Animals
Timed-pregnant or other ddY mice of appropriate ages for this study were purchased from Shizuoka Laboratory Animal Center (Hamamatsu, Japan). Mice were housed and handled according to protocols approved by Hiroshima University Research Facilities of Laboratory Animal Science.
Cell Culture
MC3T3-E1, a clonal mouse normal osteoblastic cell line, was obtained from the Riken Cell Bank (Tsukuba, Japan). Cells were grown at 37C in Dulbecco's modified Eagles medium containing 10% fetal calf serum (Upstate Biotechnology; Lake Placid, NY) and antibiotics in a humidified 5% CO2 atmosphere. For the developmental sequence study, the medium was supplemented with 50 μg/ml ascorbic acid and changed every second or third day until mineralized nodules had developed. In some cases, 10 mM β-glycerophosphate was added to cultures for 2 days before culture termination to induce matrix mineralization. To detect mineralization, cultures were fixed in neutral buffered formalin for 15 min and incubated in 2.5% silver nitrate solution (von Kossa staining).
Western Blotting Analysis
Cells were rinsed with PBS on ice, suspended in 0.5% IGE-PAL CA-630 (Sigma; St Louis, MO), 150 mM NaCl, 5 mM EDTA, Tris-HCl, pH 7.3, including 1 mM phenylmethylsulfonyl fluoride, and incubated for 2 hr at 4C on a rotating platform. After centrifugation, an aliquot of the supernatant (50 μg total protein) was mixed with 2 × SDS reducing buffer and boiled for 5 min. Standard SDS-PAGE procedures were used with a Mini-PROTEAN II cell system (Bio-Rad; Hercules, CA) and 15% acrylamide gels. Thirty ng of bacterial recombinant human STC1 (rhSTC1) was loaded as a positive control. Protein blotting from the gels to nitrocellulose membranes (Hybond ECL; Amersham Pharmacia Biotech, Poole, UK) was performed using a TRANS-BLOT SD (Bio-Rad). The membranes were blocked at room temperature (RT) for 2 hr with 0.1% Tween-20, 0.1 M NaCl, 0.1 M Tris-HCl, pH 7.5 (TTBS), including 0.2% casein. Membranes were incubated overnight at 4C with mouse anti-rhSTC1 monoclonal antibody (anti-rhSTC1, 0.5 μg/ml), followed by a Vectastain ABC kit (Vector Lab; Burlin-game, CA). Briefly, the membranes were incubated with biotinylated secondary antibody (horse anti-mouse IgG, 1:500) at 4C overnight and then avidin-biotinylated horseradish peroxidase (HRP) macromolacular complex (ABC reagents, 1:200) at RT for 30 min. Each incubation step was followed by three washes (15 min) with TTBS. Chemiluminescence detection was carried out with luminol visualization solution (0.4 mg/ml luminol, 0.1 mg/ml p-iodophenol, 0.015% H2O2 in 50 mM Tris-HCl, pH 7.5). The uniformity of protein loading was assessed by rabbit anti-actin antibody (1:1000; Miles, Elkhart, IN) and HRP-conjugated goat anti-rabbit IgG antibody for secondary antibody (1:3000, Bio-Rad). The specificity of anti-rhSTC1 was demonstrated by preabsorption of anti-rhSTC1 with rhSTC1 (100 μg/ml) and by normal mouse IgG (Santa Cruz Biotech; Santa Cruz, CA) instead of the anti-rhSTC1.
Northern Blotting Analysis
Digoxygenin (DIG)-11-UTP-labeled single-stranded RNA probes were prepared using a DIG labeling kit (Roche Diagnostics; Mannheim, Germany) according to the manufacturer's instructions. Full-length (1.2-kb) human STC1, a 0.6-kb fragment of mouse alkaline phosphatase (ALP), a 0.47-kb fragment of mouse osteocalcin (OCN), and a 0.5-kb fragment of mouse Type I (α) collagen (CollαI) cDNAs were used to generate antisense and sense probes.
Whole embryos at known developmental ages (E10.5, E12.5, E14.5, and E16.5; embryonic days post coitum), several adult tissues, and MC3T3-E1 cells were subjected to total RNA extraction by acid-guanidinium thiocyanate-phenol-chloroform extraction (Chomczynski and Sacchi 1987). Total RNA (20 μg for the tissues and 10 μg for the cultured cells) was electrophoresed on 1% agarose formaldehyde gels and transferred to positively charged nylon membranes (Roche Diagnostics). After immobilization by baking at 120C for 30 min, the membranes were prehybridized and hybridized with 100 ng/ml of probes of interest. Hybridization and washing conditions were as recommended by the manufacturer. After washing, ALP-conjugated sheep anti-DIG antibody (Roche Diagnostics) was applied and chemiluminescence detection was carried out using CSPD (Roche Diagnostics). The uniformity of RNA loading was assessed by rehybridization with a β-actin probe (Roche Diagnostics).
Tissue Preparation
For immunohistochemistry, 7-mm frozen sections from embryonic mouse tissues (littermates served for total RNA preparation) were stored at −70C until use. For ISH, the tissues were immersion-fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) overnight at 4C. The specimens were then dehydrated in a graded series of ethanol, defatted in chloroform, and embedded in paraffin. Sections of 6 μm were placed on 3-aminopropyltriethoxysilane-treated slides and stored at 4C.
Immunohistochemistry
Frozen sections were air-dried, fixed in PLP solution (10 mM NaIO4, 75 mM lysine, 37.5 mM phosphate buffer, pH 7.4, containing 2% paraformaldehyde) at RT for 10 min. To block nonspecific staining, the sections were incubated with Dako Protein Block (Carpinteria, CA) at RT for 2 hr. The sections were incubated with anti-rhSTC1 (5 μg/ml) at 4C overnight and then with biotinylated secondary antibody (1:1000) at RT for 2 hr, followed by incubation with 0.3% H2O2 in methanol at RT for 30 min. To make the avidin-biotin complex, ABC reagents (1:100; Vector) as described above were applied for 30 min at RT. Each incubation step was followed by two washes (15 min) with PBS. Staining was developed with a Vector VIP substrate kit (Vector). The sections were counterstained with methyl green and observed under a light microscope. As negative control, normal mouse IgG (Santa Cruz Biotech) was used in place of anti-rhSTC1.
In Situ Hybridization
ISH analysis was carried out as previously described (Yoshiko et al. 1999), with a slight modification. Briefly, sections were deparaffinized in xylene, rehydrated, and treated with 0.2 N HCl for 20 min, 0.2% Triton X-100 for 10 min, and proteinase K (10 μg/ml) for 10 min. After post-fixation with 4% paraformaldehyde in phosphate buffer for 5 min, aldehyde was inactivated twice in 2 mg/ml glycine in phosphate buffer for 10 min. RNase A (100 μg/ml) treatment at 37C for 30 min before hybridization with antisense probes was performed as a negative control. After dehydration and air-drying, the sections were incubated in hybridization buffer (50% formamide, 200 μg/ml yeast tRNA, 1 × Denhardt's solution, 10% dextran sulfate, 600 mM NaCl, 0.25% sodium dodecyl sulfate, and 1 mM EDTA in 10 mM Tris-HCl, pH 7.6) at 37C for 16 hr with 1 μg/ml probes. Post-hybridization treatment consisted of washing twice in 2 × SSC plus 0.075% Brij 35 (Sigma) at 37C for 20 min, followed by treatment with 1 μg/ml RNase A for 15 min, washing in 2 × SSC plus 0.075% Brij 35 for 20 min, twice in 0.1 × SSC plus 0.075% Brij 35 at 3C for 20 min, and once for 10 min at RT. The sections were then incubated in blocking reagent (Roche Diagnostics), followed by ALP-conjugated goat anti-DIG antibody. After color detection using NBT/BCIP (Roche Diagnostics) for 20–60 min, the sections were counterstained with methyl green and examined under a light microscope.
Results
STC1 mRNA Transcripts Are Differentially Regulated During Embryogenesis
Because high levels of STC1 mRNA have been detected in whole developing mouse embryos (Chang et al. 1996; Varghese et al. 1998), we first assessed whether the mRNA expression levels are regulated during embryogenesis and compared this with levels in adult tissues. We also characterized the size and pattern of the mRNA transcripts, because multiple transcripts of STC1 mRNA have been reported in other tissues (Chang et al. 1995,1996; Varghese et al. 1998; Wong et al. 1998; Yoshiko et al. 1999) and cell lines (Chang et al. 1995; Yoshiko et al. 1999). At all embryonic stages assessed (E10.5, E12.5, E14.5, and E16.5), multiple STC1 transcripts were detected (Figure 1). During embryogenesis, three STC1 transcripts (4.0, 2.6, and 1.4 kb) were upregulated, but the 1.2-kb transcript was downregulated by E16.5. Again, multiple STC1 transcripts were detectable in various adult tissues, but with differences in relative transcript abundance in different tissues. No signal was detected in any sample hybridized with the sense probe (data not shown).
Northern blotting analysis of STC1 in RNA extracts from whole mouse embryos at different developmental stages and adult mouse tissues. The position of size standards in kb are shown. RNA from embryonic mice contains high levels of STC1 mRNA even compared to those from various adult tissues. Several transcripts of the STC1 mRNA are differentially expressed during embryogenesis.
Distribution of STC1 mRNA during embryonic mouse osteogenesis. (
STC1 Is Expressed in Specific Cell Types During Embryonic Osteogenesis
To elucidate the spatiotemporal profile of STC1 mRNA expression during embryonic mouse osteogenesis, ISH was performed; sections of representative examples of bones forming by endochondral or intramembranous ossification are shown in Figure 2. For bones forming by endochondral ossification, STC1 mRNA was expressed in the mesenchymal condensations (e.g., developing vertebral column at E10.5; Figure 2A) but, as cartilage development progressed, STC1 mRNA was highly expressed in perichondrial cell layers surrounding cartilage primordia (e.g., developing ribs at E12.5; Figure 2B). Perichondrial/periosteal cells of growing but still cartilaginous bones also expressed high levels of STC1 mRNA (e.g., occipital
Distribution of STC1 mRNA in embryonic and adult mouse intestine. In embryonic intestine at E16.5, STC1 mRNA is observed throughout the epithelium (
Immunohistochemistry showed that the STC1 protein largely co-localized with STC1 mRNA (Figure 4). For example, diffuse staining of STC1 protein was seen in the perichondrium of the developing endochondral sphenoid bone at E12.5 (Figure 4A). Osteoblasts and associated periosteal cells of the intramembranous frontal (Figure 4B) and parietal (Figure 4C) bones stained intensely for STC1 protein at E16.5. In E16.5 metatarsals (endochondral bones), hypertrophic chondrocytes but not other chondrocytes and osteoblasts labeled for STC1 protein (Figure 4D). When anti-rhSTC1 antibodies were replaced with normal mouse IgGs, no labeling over background was seen (e.g., metatarsals; Figure 4G).
STC1 Is Upregulated During MC3T3-E1 Cell Development
We previously demonstrated that STC1 mRNA is expressed in several osteoblastic cell lines, including the mouse calvaria-derived line, MC3T3-E1 (Yoshiko et al. 1999). MC3T3-E1 cells are a useful in vitro model in which to study osteoblast development because they display a time-dependent and sequential expression of osteoblast phenotypic markers similar to that seen in bone formation in vivo, including production of extracellular mineralized matrix (Franceschi and Iyer 1992; Quarles et al. 1992). We therefore next examined the expression patterns of STC1 protein and mRNA at several different time points encompassing proliferation and differentiation stages in MC3T3-E1 cell cultures, i.e., at 80% confluence (Day 4 for RNA) and Days 7, 13, 18, 23, and 28 of culture. In this study, bone nodules (unmineralized) were first observed at day 12 and continued to develop throughout the duration of culture, with mineralization first detected at Day 20 (data not shown). Anti-rhSTC1 antibodies detected low levels of a single protein product. The levels increased up to Day 18 and remained high thereafter (Figure 5). The molecular mass seen (approximately 25 kD) was almost same as that of rhSTC1 recovered from bacteria and without glycosylation. The specificity of anti-rhSTC1 was confirmed by preabsorption of the antibody with rhSTC1 or by normal mouse IgG instead of anti-rhSTC1.
Expression of STC1 mRNA in differentiating MC3T3-E1 cell cultures was compared to that of several well-established osteoblast markers, including Collα1, ALP, and OCN (Figure 6). As expected, Collα1 was detectable throughout the culture time but levels were highest at earlier times, whereas ALP mRNA was detected and upregulated from Day 13 and OCN mRNA from Day 18 as bone nodules formed and mineralized. STC1 mRNA was detectable from the earliest time analyzed (Day 4) but increased and remained high during late differentiation and mineralization phases (from Day 18).
Discussion
We previously demonstrated that STC1 mRNA is expressed in osteoblasts of developing mouse long bone, calvaria, and several osteoblastic cell lines (Yoshiko et al. 1999). As a step towards determining the physiological functions of STC1 in bone, we have now extended our recent observations by assessing the spatiotemporal distribution of STC1 protein and mRNA during mouse embryonic osteogenesis and the osteoblast differentiation sequence in the MC3T3E-1 cell model. Our results suggest that STC1 protein and mRNA are co-localized as early as the mesenchymal cell condensation stages of bone development and that expression becomes restricted to cells of the osteoblast lineage as bones develop. Expression is highest in mature osteoblasts in bones forming by both endochondral and intramembranous routes. STC1 protein (but not its mRNA) was also selectively identified in hypertrophic chondrocytes (see below), suggesting that these cells also play a role in STC1 production and action.
Distribution of STC1 protein during embryonic mouse osteogenesis. (
Northern blotting analysis revealed that STC1 mRNA is highly expressed in extracts from whole mouse embryos and in various adult tissues. In contrast to an earlier analysis in which steady-state levels of a 4-kb transcript were detected in whole embryos (Varghese et al. 1998) and in some embryonic tissues (Stasko and Wagner 2001), we found at least four transcripts that were differentially regulated during embryogenesis. Our data are consistent with the reported presence of two transcripts in ovary (Varghese et al. 1998) and multiple transcripts in several other tissues (Chang et al. 1995,1996; Yoshiko and Maeda 1998; Yoshiko et al. 1999), including E16 mouse embryos (Chang et al. 1995) and in several cell lines (Chang et al. 1995; Yoshiko et al. 1999). These multiple transcripts are probably due to alternative usage of polyadenylation signals in the long 3′UTR, because alternative splicing of coding exons among the various STC cDNAs has not been detected (Varghese et al. 1998). A correlation between the tissue- and development-specific expression patterns of the multiple transcripts is still under investigation. However, taken together with previous observations that no change in STC1 mRNA levels is observed in the major embryonic tissues expressing it, i.e., kidney, heart and lung, except for thymus in which mRNA levels decrease, between E14.5 and 18.5 of mouse embryo development (Stasko and Wagner 2001), the present results suggest that the increase in the transcriptional levels of STC1 mRNA in whole mouse embryos appears to correspond to formation of the skeleton, most notably to development of osteoblasts and mineralized bone.
Western blotting analysis of STC1 protein during osteoblast development in the MC3T3-E1 cell model. The positions of size standards, in kD, are shown. Anti-hrSTC1 recognizes an approximately 25-kD band with rhSTC1 and in cell extracts. Higher levels of the protein are present from Day 18 to culture termination.
It has been observed that both STC1 protein and mRNA expression are ubiquitous in mouse embryonic tissues but, as shown in Figures 2 and 4, both STC1 protein and mRNA are higher in osteoblastic cells than in surrounding musculoskeletal tissue cell types as well as in developing brain, skeletal muscle, and skin. The high and specific localization of STC1 in osteoblast lineage cells during embryonic osteogenesis suggests a significant role for STC1 in bone development and is in partial agreement with an earlier study in which the protein was found to be abundant in the musculoskeletal system in fetal mice (Jiang et al. 2000). Notably, however, the antibody used in the earlier study found STC1 protein to be restricted to a zone of cells in which chondrocytes undergo transition from a proliferating to a hypertrophic phenotype (Jiang et al. 2000). It is not entirely clear why our results and the previous studies detect different cohorts of STC1-producing cells, although a plausible explanation for the discrepancy may be differences of epitope retention/accessibility in the different procedures for the preparation of tissue sections and/or by the preparation of antibodies. In any case, the almost exact (with the exception of hypertrophic chondrocytes; see below) co-localization of cells expressing STC1 protein and mRNA in our study suggests that osteoblasts and associated periosteal cells are the main sources of STC1 in the skeleton. The latter is in agreement with our analyses of STC1 mRNA and protein expression in the MC3T3-E1 model, a non-transformed clonal cell line that undergoes a proliferation-differentiation sequence resembling that in primary osteoblast cultures (Franceschi and Iyer 1992; Quarles et al. 1992).
Northern blotting analysis of STC1 and osteoblastic markers during osteoblast development in the MC3T3-E1 cell model. The positions of size standards, in kb, are shown. STC1 mRNA is relatively higher from Day 18 to culture termination. Stages of osteoblast development are also marked by high ALP and osteocalcin mRNA expression.
STC1 protein was also located in hypertrophic chondrocytes but the mRNA was not, a discrepancy shown previously in kidney, in which STC1 protein is detected exclusively in the segments that do not express the mRNA (Wong et al. 1998). Discrepancies between protein and mRNA expression have been previously noted for several molecules and are generally believed to result from differences in, e.g., the stability of the mRNA/protein reflecting post-transcriptional/post-translational modulation, differences between synthesizing cells vs target cells, and/or differences in detection limits of probes and antibodies used. The localization of STC1 protein in cartilaginous tissue appears to correspond to the differentiation stage, during which chondrocytes start to mineralize their extracellular matrix (Alini et al. 1994; Sommer et al. 1996). The cell surface receptor for gibbon ape leukemia virus (Glvr-1/PiT1), the Type III sodium-dependent phosphate transporter/retrovirus receptor gene, has also been reported to be localized to the same zone where it may regulate inorganic phosphate (Pi) uptake in osteogenic cells and contribute to bone mineralization (Palmer et al. 1999). Interestingly, the expression pattern of Glvr-1 mRNA in kidney (Tenenhouse et al. 1998) also appears to match that of STC1 mRNA in kidney (Yoshiko and Maeda 1998). These observations suggest that it would be interesting to examine whether STC1 regulates the sodium/phosphate co-transport system in bone and cartilage, because rhSTC1 is able to stimulate sodium/phosphate co-transport in kidney (Wagner et al. 1997). We have recently tested this hypothesis and indeed have found a close correlation among STC1, Pi transport, and mineralization in the primary rat osteoblast model (our unpublished observations).
In contrast to a recent immunohistochemical study of E15.5 embryos in which STC1 protein was not detected in intestine (Jiang et al. 2000), we found STC1 mRNA in epithelial cells in both embryonic and adult intestine. Whereas in embryonic intestine STC1 mRNA was detected throughout the epithelium, STC1-expressing cells were restricted to the base of the villi and mainly to the crypt region in adult intestine, a region in which stem cells are located. In adult rodent intestine, both STC1 protein and mRNA are also detectable by radioimmunoassay (De Niu et al. 2000) and Northern blotting (Chang et al. 1995,1996; and our unpublished data), respectively. Taken together, the data suggest that STC1 may be involved in enterocyte development as well as intestinal calcium/phosphate transport.
In summary, we conclude that STC1 is expressed in specific cell types in tissues responsive to calcium/phosphate transport including intestinal epithelial cells in embryonic and adult tissues. Notably, we have demonstrated that STC1 expression is present as early as the mesenchymal condensations and subsequently is high in cells committed to the osteoblast lineage, especially the mature osteoblasts, during embryonic mouse osteogenesis. Our results suggest that osteoblasts may be a rich source of and a target for action of STC1 during both endochondral and intramembranous bone formation.
Footnotes
Acknowledgements
Supported in part by grants-in-aid from the Ministry of Education, Science, Sports and Culture of Japan and by CIHR (MT-12390 to JEA) of Canada.
We are grateful to Dr Roger R. Reddel of Children's Medical Research Institute (NSW, Australia) for providing us with the recombinant plasmid pAC 143 including human STC cDNA, and to Akira Igarashi and Dr Shoichi Takano of BML Inc. (Saitama, Japan) for providing us with rhSTC1 and anti-rhSTC1. We also thank Aoi Son and Usha Bhargava for technical assistance.