Fish scales are a form of calcified tissue similar to that found in human bone. In medaka scales, we detected both osteoblasts and osteoclasts and subsequently developed a new scale assay system. Using this system, we analyzed the osteoblastic and osteoclastic responses under 2-, 3-, and 4-gravity (G) loading by both centrifugation and vibration. After loading for 10 min, the scales from centrifugal and vibration loading were incubated for 6 and 24 hrs, respectively, after which the osteoblastic and osteoclastic activities were measured. Osteoblastic activity significantly increased under 2- to 4-G loading by both centrifugation and vibration. In contrast, we found that osteoclastic activity significantly decreased under 2- and 3-G loading in response to both centrifugation and vibration. Under 4-G loading, osteoclastic activity also decreased on centrifugation, but significantly increased under 4-G loading by vibration, concomitant with markedly increased osteoblastic activity. Expression of the receptor activator of the NF-κ;B ligand (RANKL), an activation factor of osteoclasts expressed in osteoblasts, increased significantly under 4-G loading by vibration but was unchanged by centrifugal loading. A protein sequence similar to osteoprotegerin (OPG), which is known as an osteoclastogenesis inhibitory factor, was found in medaka using our sequence analysis. The ratio of RANKL/OPG-like mRNAs in the vibration-loaded scales was significantly higher than that in the control scales, although there was no difference between centrifugal loaded scales and the control scales. Accordingly, medaka scales provide a useful model by which to analyze bone metabolism in response to physical strain.
INTRODUCTION
Various in vitro models have been developed elucidate the effect of physical strain, including the response of bone metabolism to different gravitational loads, (Tjandrawinata et al., 1997; Tanaka et al., 2003; Peng et al., 2011). Most studies using in vitro models noted osteoblastic responses to physical strain as the determinant in bone (Tjandrawinata et al., 1997; Tanaka et al., 2003). Bone consists of osteoblasts, osteoclasts, and the bone matrix, and both cell-to-cell and cell-to-matrix interactions are critical for cell response to physical stress (Harter et al., 1995; Owan et al., 1997; Hoffler et al., 2006). It was recently reported that mechanical stretch-induced calcium efflux from bone matrix and stimulated osteoblasts, suggesting that the bone matrix acts as a reservoir for mechanochemical transducers, which convert mechanical strain into a chemical signal for the onset of calcium efflux (Sun et al., 2012). A co-culture system of osteoblasts and osteoclasts with the matrix is essential to elucidate bone metabolism under gravity (G) loading conditions, indicating the need for the development of such a system.
The teleost scale, a calcified tissue, contains osteoblasts and osteoclasts (Bereiter-Hahn and Zylberberg, 1993, Suzuki et al., 2000; Yoshikubo et al., 2005) that are similar to those found in avian and mammalian membrane bone. Moreover, multinucleated osteoclasts, an active type of osteoclast, have been detected by tartrate-resistant acid phosphatase (TRAP) staining in the scales of goldfish (Suzuki et al., 2000; Azuma et al., 2007; Suzuki et al., 2011), carp (de Vrieze et al., 2010), and rainbow trout (Persson et al., 1999), together with the osteoblasts detected by alkaline phosphatase (ALP) staining (de Vrieze et al., 2010). With such typical components of bone matrix as type I collagen (Zylberberg et al., 1992), bone γ-carboxyglutamic acid protein (Nishimoto et al., 1992), osteonectine (Lehane et al., 1999), and hydroxyapatite (Onozato and Watabe, 1979), a teleost scale is a suitable model for mammalian bone.
Medaka (Oryzias latipes), a small teleost, is a particularly suitable model organism, as its entire genome sequence has been mapped, facilitating genetic analysis. Its relatively short life cycle and high productivity (Kasahara et al., 2007; Kawakami, 2007; Takeda, 2008) are also valuable features. Therefore, medaka appears to be a model organism that can be used to analyze various biological processes, including bone metabolism, at the molecular level (Inohaya et al., 2007; Watanabe-Asaka et al., 2010). Medaka were launched into space in 1994 as part of a microgravity experiment (Ijiri, 1995). This experiment included the first observation of the mating behavior, development, and hatching of a vertebrate in space. Using medaka, we can analyze the response of osteoblasts and osteoclasts, not only in a hypergravity environment on the ground but also under microgravity conditions in space.
In the present study, we investigated bone metabolism under G-loading using medaka scales as a bone model. To demonstrate the coexistence of bone cells in medaka scales, we first analyzed the morphological features of both osteoblasts and osteoclasts. Second, we developed a new in vitro assay system using medaka scales. In this system, ALP and TRAP were used as respective markers of osteoblasts and osteoclasts. Third, we examined static (centrifugation) and dynamic (vibration) G-loading using the developed assay system with medaka scales. We demonstrated that osteoblasts and osteoclasts in medaka scales responded with certain degrees of sensitivity to G-loading by both centrifugation and vibration. Using our original system with medaka scales, we measured the difference between static and dynamic G-loading.
MATERIALS AND METHODS
Animals
Medaka were purchased from a commercial source (Higashikawa Fish Farm, Yamatokoriyama, Japan) and used for the in vitro scale assay. All experimental procedures were conducted in accordance with the Guidelines for the Care and Use of Laboratory Animals of Kanazawa University, Japan.
Morphological study of osteoblasts and osteoclasts in medaka scales
Scales were collected from medaka anesthetized with ethyl 3aminobenzoate, methanesulfonic acid salt (Sigma-Aldrich, Inc., St. Louis, MO, USA) and fixed using 4% paraformaldehyde solution neutralized with phosphate buffer solution (pH 7.2; Wako, Co., Ltd., Osaka, Japan). Subsequently, osteoblasts were detected by ALP staining using NBT/BCIP Stock Solution (Roche Applied Science, Mannheim, Germany). The scales were TRAP stained using the methods described by Cole and Walters (1987). After staining, the osteoblasts and osteoclasts were observed under a microscope.
Development of an in vitro assay system using medaka scales
Medaka were anesthetized with ethyl 3-aminobenzoate, methanesulfonic acid salt (Sigma-Aldrich, Inc., St. Louis, MO, USA), after which all scales collected from the left side of the animal were placed into a 1.5-mL microtube and all scales collected from the right side of medaka were transferred into a different 1.5-mL microtube. One hundred microliters of distilled water were added to each microtube. After sonication, the tube was centrifuged and the supernatant was used to detect both ALP and TRAP activities. The methods for measuring ALP and TRAP activities were reported by Suzuki et al. (2007). The ALP and TRAP data obtained for the scales from the left and right sides of the medaka were compared.
Effect of osteoblastic and osteoclastic activities under 2-, 3-, and 4-G loading by centrifugation and vibration
Scales were collected from medaka anesthetized with ethyl 3-aminobenzoate, methanesulfonic acid salt. All the left-side scales (loaded experimental scales) and right-side scales (unloaded control scales) from each individual were put into respective microtubes, followed by the addition of a 500-µL aliquot of Leibovitz's L15 medium (Invitrogen, Grand Island, NY, USA) containing a 1% penicillin-streptomycin mixture (ICN Biomedicals, Inc., Aurora, OH, USA). To fix the scales, a cotton ball (1 cm in diameter) was placed into each microtube. The microtube containing the scales was loaded to 2-, 3-, and 4-G by centrifugation (LIX-130; Tomy Digital Biology Co., Ltd., Tokyo, Japan) or by vibration with the original apparatus (Suzuki et al., 2007) for 10 min at room temperature. The loaded scales were compared with unloaded (1-G control) scales. The loading times were determined following our previous study using goldfish scales (Suzuki et al., 2007). After loading, the centrifuged and vibrated scales were incubated for 6 and 24 hrs, respectively, at 15°C. We previously reported that calcemic hormones such as calcitonin and estrogen were effective at these incubation times (Suzuki et al., 2000; Yoshikubo et al., 2005); therefore, we used these times in the present study. After incubation, the ALP and TRAP activities in the medaka scales were measured as described above.
Analysis of the interaction between osteoblasts and osteoclasts under 4-G loading by centrifugation and vibration
We analyzed the mRNA expression of both the receptor activator of the NF-κ;B ligand (RANKL), an activating factor of osteoclasts expressed in osteoblasts, and osteoprotegerin (OPG), an osteoclastogenesis inhibitory factor in osteoblasts. OPG, a decoy receptor of RANKL, inhibits osteoclastogenesis by binding to RANKL (see the review by Lacey et al., 2012). To analyze the interaction between osteoblasts and osteoclasts, the ratio of the expression of the RANKL/OPG-like mRNA was examined.
After 4-G loading, the scales were incubated for 24 hrs at 15°C and then frozen at -80°C for mRNA analysis. The loaded scales were compared with unloaded (1-G control) scales. Total RNAs were prepared from medaka scales using a total RNA isolation kit for fibrous tissue (Qiagen GmbH, Hilden, Germany). Complementary DNA was synthesized using the PrimeScript™ RT reagent kit (Takara Bio Inc., Otsu, Japan). The primer sequences—sense: AGGCAAAACGGCAAAGAAAT; anti-sense: CCCAGCTTTATGGCTCCAA—were designed from medaka RANKL (JN119285) (To et al., 2012). The OPG-like sequence in medaka was determined by sequence analysis, as follows. All amino acid sequences of medaka were retrieved from the Genome to Protein Structure and Function database. We analyzed the amino acid sequence of OPG in fugu (ENSTRUP00000023772) as a query to all amino acid sequences of medaka, and selected the best-hit sequence as the candidate for the OPG amino acid sequence of medaka. Thereafter, we used CLUSTAL X2 (Larkin et al., 2007) for multiple sequence alignment of homologous sequences. The primers for the OPG-like sequence in medaka were as follows: sense: 5′-GGATCCGTCCACTGGTAAAA-3′; antisense: 5′-GAGCACTCGATTTCCACCTC-3′.
Fig. 1.
Microscopic views of medaka scales stained for alkaline phosphatase (ALP). Arrows indicate ALP positive cells. Bar: 100 µm.
β-actin (ENSORLT00000021168) was amplified using the following primers: sense: 5′-TGTGCTACGTGGCTCTTGAC-3′; antisense: 5′-GCCAATGAAAGAAGGTTGGA-3′. The PCR amplification was performed using the real-time Mx3000p PCR apparatus (Agilent Technologies. Santa Clara. CA. USA) (Suzuki et al.. 2011). The annealing temperature of RANKL, OPG-like, and βactin fragments was 60°C. The initial reaction condition was 10 s at 95°C, followed by 40 cycles of denaturation at 95°C for 5 s, and annealing/extension at 60°C for 40 s. The RANKL and OPG-like mRNA levels were normalized to the β-actin mRNA level.
RESULTS
Osteoblasts and osteoclasts in medaka scales
In the medaka scales, ALP-stained osteoblasts (Fig. 1) and TRAP-stained mono- and multi-nucleated osteoclastic cells were detected (Fig. 2).
Fig. 2.
Microscopic views of medaka scales stained for tartrateresistant acid phosphatase. Arrowheads and arrows indicate monoand multi-nucleated osteoclastic cells, respectively. Bar: 100 µm.
Comparison of ALP and TRAP activities between scales from the left and right sides of medaka
ALP activity in scales from the left side of medaka was similar to that in scales from the right side (Fig. 3A) with no significant difference. The TRAP activity in the scales from the left side was nearly equal to that in scales from the right side (Fig. 3B).
Effect of osteoblastic and osteoclastic activities under 2-, 3-, and 4-G loading by centrifugation
The ALP activity significantly increased under 2-G loading by centrifugation after 6 hrs of incubation (Fig. 4A). The ALP activity increased significantly under 3- and 4-G loading by centrifugation after 24 hrs of incubation (Fig. 4B). In contrast, under 2- to 4-G loading by centrifugation, the TRAP activity significantly decreased after both 6 and 24 hrs of incubation (Fig. 5A, B).
Effect of osteoclastic and osteoblastic activity under 2-, 3-, 4-G loading by vibration
The ALP activity increased significantly under 3- and 4-G loading by vibration after 6 hrs of incubation (Fig. 6A). Under 2- to 4-G loading by vibration, the ALP activity significantly increased after 24 hrs of incubation (Fig. 6B). Under 3-G loading by vibration, the TRAP activity significantly decreased after 6 hrs of incubation (Fig. 7A). After 24 hrs of incubation, the TRAP activity significantly decreased under 2-G loading but significantly increased under 4-G loading by vibration (Fig. 7B).
Determination of the OPG-like sequence in medaka
Using our original system, the OPG-like sequence was determined from the medaka database. After analysis using CLUSTAL X2, we found that 18 cysteines are strictly conserved in fish, birds, and mammals (Fig. 8). The amino acid identity of the medaka OPGlike sequence was 69.9% compared to the fugu OPG, 38.3% compared to the chicken OPG, 39.0% compared to the mouse OPG, and 41.2% compared to the human OPG. Furthermore, similarity analysis using the BLOSUM62 matrix confirmed that the medaka OPG-like sequence shows a high similarity to fugu (80.1%) and a relatively high similarity to other vertebrate OPG (chicken, 60.0%; mouse, 58.0%; human, 58.3%).
Analysis of the interaction between osteoblasts and osteoclasts under 4-G loading by centrifugation and vibration
RANKL mRNA expression was increased significantly by vibration loading, but not by centrifugal loading (Fig. 9A). The expression of the OPG-like mRNA was not significantly different between centrifugal and vibration loading (4-G) (Fig. 9B). The ratio of the expression of RANKL/OPG-like mRNA in the vibrationloaded scales was significantly higher than that in the control scales (Fig. 10).
Fig. 5.
Effect of osteoclastic activity after incubation for 6 (A) and 24 hrs (B) under 2-gravity (G), 3-G, 4-G loading by centrifugation. * and ** indicate statistically significant differences at P < 0.05 and P < 0.01, respectively, from the values in the control scales.
Fig. 6.
Effect of osteoblastic activity after incubation for 6 (A) and 24 hrs (B) under 2-gravity (G), 3-G, 4-G loading by vibration. * and ** indicate statistically significant differences at P < 0.05 and P < 0.01, respectively, from the values in the control scales.
Fig. 7.
Effect of osteoclastic activity after incubation for 6 (A) and 24 hrs (B) under 2-gravity (G), 3-G, 4-G loading by vibration. * and ** indicate statistically significant differences at P < 0.05 and P < 0.01, respectively, from the values in the control scales.
Fig. 8.
Multiple protein sequence alignment of osteoprotegerin (OPG) with other animals: fugu OPG (ENSTRUP00000023772), chicken OPG (ENSGALP00000025915), mouse OPG (PRO_0000034588), and human OPG (PRO_0000034587), and concatenated cysteines represent disulfide bonds in OPG sequences. This multiple protein sequence alignment showed that disulfide bonds are highly conserved in the OPG sequences of vertebrates.
DISCUSSION
The present study is the first to demonstrate that medaka scales respond to G-loading by using ALP and TRAP as marker enzymes of osteoblasts and osteoclasts stained by the respective substrates (Figs. 1, 2), supporting the notion that this new assay system can be a useful tool in analyzing the response of these cells to gravitational stress.
Osteoblasts in medaka scales were activated by loading with increasing gravity (2-G, 3-G, and 4-G) for 10 min by centrifugation. Medaka scales responded to G-force < 5, which is much lower than the 5 to 50 G needed for mammalian osteoblasts to respond (Gebken et al., 1999; Saito et al., 2003; Searby et al., 2005), suggesting that medaka scales are very sensitive to G-loading. In addition, the osteoclastic activity in medaka scales decreased under 2-G, 3-G, and 4G loading by centrifugation, suggesting a very good response of osteoclasts to low G loading, although the mechanisms have yet to be determined. Moreover, osteoblastic and osteoclastic activities in medaka scales were sensitive not only to dynamic G-loading by vibration with a G-load apparatus, but also to static G-loading by centrifugation. Fish scales contain osteoblasts, osteoclasts, and a matrix similar to that in bone (Bereiter-Hahn and Zylberberg, 1993; Suzuki et al., 2000; Yoshikubo et al., 2005; Azuma et al., 2007). Bone matrix plays an important role in the response to physical stress (Owan et al., 1997; Hoffer et al., 2006); in future studies, we will seek to identify the difference in matrix components, if any, to elucidate the mechanisms underlying the extreme sensitively of medaka scales to G-loading.
The interaction between osteoblasts and osteoclasts has been reported in mammals, and cytokine from osteoblasts, in particular, is required to produce differentiated osteoclasts (Suda et al., 1999). RANK in osteoclasts binds RANKL, the ligand, resulting in osteoclast activation whereby multinucleated osteoclasts (an active type of osteoclast) are formed (Teitelbaum et al., 2000). OPG, a decoy receptor of RANKL, inhibits osteoclastogenesis by binding to RANKL (see the review by Lacey et al., 2012). We demonstrated that RANKL mRNA expression increases significantly in response to vibration loading, but not centrifugal loading. In addition, the ratio of the expression of RANKL/ OPG-like mRNA in the scales loaded by vibration was significantly higher than that in control scales, while there was no difference on centrifugation. Because the RANKL/OPG ratio is an indicator of bone resorption (Lacey et al., 2012), medaka scales provide a good model by which to investigate bone metabolism.
In a study of a widely used hind-limb-elevation (tail suspension) model, the results of bone resorption by osteoclasts were inconsistent (Carmeliet et al., 2001); in this field, most subsequent research has thus been focused on osteoblastic response in bone formation. Recently, in isolated osteoclasts, it was reported that osteoclastic activity increased in those osteoclasts that were cultured in space (Tamma et al., 2009); however, there has been no data relating the interaction between osteoblasts and osteoclasts in space. Although bone mass is reportedly increased by mechanical strain, it has also been reported that mechanical strain on isolated osteoclasts upregulated their bone-resorbing activity (Kurata et al., 2001); therefore, the results obtained from the isolated osteoclast system may differ from that obtained from the in vivo system. Moreover, we demonstrated that melatonin, a major hormone secreted from the pineal gland, activated the growth of isolated osteoblasts in culture, although this hormone also suppressed the functions of osteoblasts using an in vitro assay system with scales (Suzuki and Hattori, 2002). These results obtained using an in vitro assay system resembled those of an in vivo study in rats (Ladizesky et al., 2003). Based on these results, we propose that our study using a scale-organ-culture system provides more accurate reproduction of the in vivo study.
Medaka has a number of beneficial features as a model organism, and transgenic systems have been developed using this species. For example, fluorescent protein markers of osteoblasts can be observed in vivo (Inohaya et al., 2007). Furthermore, the launch of the aquatic animal habitation module (Aquatic Habitat), which is planned for the near future, will enable breeding medaka in the International Space Station (Sakimura et al., 2003; Watanabe-Asaka et al., 2010). We will be able to perform space experiments using the scales of space-bred medaka, both in vitro and in vivo, in the near future. Together with the present data, medaka scales provide a promising model system by which to study bone metabolism, and will also be useful in evaluating the physical strain associated with gravity and microgravity in space flight.
ACKNOWLEDGMENTS
This study was supported in part by grants to N.S. (Kurita Water and Environment Foundation; Grant-in-Aid for Space Utilization by the Japan Aerospace Exploration Agency; Grant-in-Aid for Scientific Research [C] Nos. 21500404 and 24620004 by JSPS), to A.H. (Grant-in-Aid for Scientific Research [C] Nos. 21570062 and 24570068 by JSPS), to K.K. (Grant-in-Aid for Scientific Research [C] Nos. 21500681 and 24500848 by JSPS), to T.S. (Grant-in-Aid for Young Scientists [B] Nos. 22770069 and 40378568 by JSPS), to H.M. (Grant-in-Aid for Scientific Research [C] No. 23592727 by JSPS), and to K.H. (the Environment Research and Technology Development Fund [B-0905] sponsored by the Ministry of the Environment, Japan; Health, Labour Sciences Research Grants of the Ministry of Health, Labour and Welfare, Japan; Grant-in-Aid for Scientific Research [B] No. 21390034 and for Exploratory Research No. 24651044 by JSPS).
