Myostatin is a member of the transforming growth factor (TGF)-β superfamily, known for its ability to inhibit muscle growth. It can also regulate metabolism and glucose uptake in a number of tissues. To determine the mechanism of myostatin's effect on glucose uptake, we evaluated its actions using choriocarcinoma cell lines that are widely used as models for placental cells. Protein and mRNA were determined using immunoblotting and RT-PCR/PCR, respectively. Glucose uptake was assessed by uptake of radiolabeled deoxyglucose in vitro. All choriocarcinoma cell lines tested i.e., BeWo, JEG, and Jar, are used as models of placental cells, and all expressed myostatin protein and mRNA. Treatment of BeWo cells with myostatin resulted in inhibition of glucose uptake in a concentration-dependent manner (P < 0.01). At all concentrations tested, follistatin, a functional inhibitor of myostatin, completely blocked the inhibitory effect of myostatin (40 nM) on glucose uptake by BeWo cells (0.4 nM, P < 0.05). Follistatin treatment alone also increased glucose uptake (0.4 and 4 nM, P < 0.001; 40 nM, P < 0.05). Because BeWo cells proliferated and greater cell densities were achieved, glucose uptake declined irrespective of treatment. Myostatin treatment of BeWo cells did not alter the levels of myostatin receptor, ActRII A/B proteins. The levels of glucose transport proteins also remained unaltered in BeWo cells with myostatin treatment. This study has shown that myostatin specifically inhibits glucose uptake into BeWo cells, suggesting that locally produced myostatin may control glucose metabolism within the placenta.
- glucose uptake
- glucose transporter
- glucose transporter
- activin receptor
- activin type II receptor A
- activin type II receptor B
- cell line
myostatin is a member of the transforming growth factor (TGF)-β super family, which was first described in skeletal muscle and characterized as a negative regulator of skeletal muscle growth (28). It has since been found in other tissues in both mammalian and nonmammalian systems (13, 28, 34, 37, 38, 43, 44). Genetically engineered mice lacking myostatin show a significant increase in skeletal muscle mass because of muscle fiber hyperplasia and hypertrophy (28). Muscular cattle breeds like Belgian Blue and Piedmontese are found to have nonfunctional myostatin protein due to mutations in the highly conserved regions of the myostatin gene (15, 29). Similar loss of myostatin function has been characterized in a hypermuscular German boy who carried a mutation in his myostatin gene (41). Like other members of the TGF-β family, myostatin is synthesized as a precursor protein made of two protein domains, an NH2-terminal propeptide and a mature COOH-terminus domain (47). The myostatin propeptide exists as a homodimer and, following proteolytic processing, the propeptide and other proteins like follistatin bind to the COOH-terminus dimer to inhibit myostatin activity (19). Mature myostatin, following release from its inhibitory proteins, signals via binding to the activin type II receptors ActRII B and to a lesser extent with ActRII A (19).
In mouse models of obesity and type 2 diabetes, the absence of myostatin activity partially reduced hyperglycemia and adipogenesis indicative of a role for myostatin in glucose metabolism (30). Under a high-fat diet, transgenic mice overexpressing myostatin propeptide are able to maintain normal blood glucose levels and insulin sensitivity compared with their wild-type counterparts, indicating that myostatin can control plasma glucose (52). Inhibition of myostatin activity can improve insulin sensitivity and increase glucose uptake in tissues like muscle and fat by upregulation of glucose transporters such as GLUT1 and GLUT4 (21). Cumulatively, these studies indicate a role for myostatin in the regulation of metabolism and body composition. Myostatin is expressed in the human placenta, and expression changes with gestation (31), suggesting that myostatin may have a metabolic role in the human placenta.
A choriocarcinoma placental cell line was used in our studies here as opposed to primary trophoblasts because of its stability with passaging and easy maintenance necessary for extensive experimental studies (23). Additionally, working with placental trophoblasts has problems with lower yield and viability and batch-to-batch variability due to trophoblast isolation techniques; furthermore, their limited life span with in vitro culture rendered them unsuitable (10, 36). Hence, this study was undertaken with BeWo cells, a well-characterized model placental cell line, after initial comparison with JEG and JAr cells, to examine the mechanism of myostatin effect on glucose uptake. BeWo cells are one of the preferred model systems used for placental transport studies, in particular glucose transport across the placenta (23, 42). These cells also express glucose transporters, GLUT1, -3, and -4 (14, 42), and some of these glucose transporters have been shown to be affected by inhibition of myostatin activity (21). This study aims to determine the effects of myostatin on glucose uptake into BeWo cells and will also determine the expression and activity of the myostatin receptor ActRII A/B and glucose transporter proteins GLUT1, -3, and -4.
MATERIALS AND METHODS
Choriocarcinoma cell cultures.
BeWo, JEG, and JAr cells were cultured in Ham's F-12/DMEM (low glucose; GIBCO) with phenol red supplemented with 10% heat-inactivated FBS (GIBCO), 1% penicillin-streptomycin-glutamine liquid (GIBCO), and 1% glutamax liquid (GIBCO). Cells were seeded at a density of 2 × 105 cells/well in six-well plates for gene and protein expression studies. Cell cultures were maintained in a humidified 37°C incubator with 5% CO2.
Western blot analysis.
Cells were lysed in lysis buffer [60 mM Tris·HCl, pH 6.8, 20% glycerol, 2% SDS, and complete protease inhibitor tablet cocktail (Complete; Roche Diagnostics NZ, Auckland, New Zealand)]. Cell lysates were collected, briefly sonicated, and centrifuged at 20,817 g for 10 min. Ovine skeletal muscle used as positive control for myostatin protein and mouse muscle for all other proteins studied were processed as described previously (31). The protein concentration of the supernatant was determined using the DC assay (Bio-Rad), and samples were stored at −80°C until further use. Protein (20 μg) from each sample was mixed with reducing loading buffer (125 mM Tris·HCl, pH 6.8, 20% glycerol, 4% SDS, and 10% 2-mercaptoethanol) and was then either boiled for 5 min at 95°C or incubated for 30 min at room temperature (for GLUT protein analysis). Samples were then subjected to gel electrophoresis in a 10–12% SDS-polyacrylamide gel and transferred to nitrocellulose membranes (Bio-Rad). Membranes were first blocked for 1 h at room temperature and then incubated overnight at 4°C with appropriate primary antibody (Table 1). Following brief washes with PBS/TBS and Tween 20 the membranes were incubated for 2 h with appropriate horseradish peroxidase-conjugated secondary antibody (Table 1). The immunoreactive proteins were visualized using enhanced chemiluminescence (Super Signal; Pierce), and relative optical densities were measured using a densitometer (GS 800; Bio-Rad Laboratories, Auckland, New Zealand) and Quantity One software (Bio-Rad Laboratories). Membranes were later stripped with 0.2 M NaOH for 5 min and reprobed for β-actin as the loading control.
Total RNA extraction.
Following removal of culture media and brief washes with cold PBS, total RNA was extracted from cells using an RNA aqueous kit from Ambion following the manufacturer's suggested protocol (Ambion).
RT-PCR and PCR.
Total RNA was extracted from cells as described above and then spectrophotometrically assessed for concentration and quality at 260/280 nm. First-strand cDNA was synthesized using the Superscript II Preamplification kit (Invitrogen) in a 20-μl reaction volume with 1 μl (∼2 μg) of total RNA using oligo(dT) primers according to the manufacturer's protocol.
Myostatin cDNA was PCR amplified in a 25-μl reaction volume with 40 cycles using 1 μl of the RT product using the Taq PCR core kit (Qiagen), with mouse muscle cDNA as positive control. Amplification of myostatin transcripts was conducted using the primers (forward primer, 5′-GGTATTTGGCAGAGTATTGAT-3′; reverse primer, 5′-GTCTACTACCATGGCTGGAAT-3′; see Refs. 15 and 33) in a 3-step/cycling, with operating parameters as 94°C initial denaturation for 30 s, 55°C primer annealing for 1 min, by 72°C strand extension for 1 min, followed by a final extension at 72°C for 5 min. Q-solution (Qiagen) was also added to the PCR reaction for myostatin. Next, PCR products were separated on 1.5% agarose gel prestained with ethidium bromide.
Cell seeding density of 5 × 104 cells/well in 24-well plates was determined after testing various densities for sensitivity to glucose uptake (data not shown). Glucose uptake by BeWo cells was carried out as described (31) with a few adaptations. At 24, 48, 72, and 96 h after seeding, the culture media were removed, and cells were washed briefly with tyrode buffer [in mM: 135 NaCl, 5 KCl, 1.8 CaCl2, 1 MgCl2 (6H2O), and 10 HEPES as described previously (2) with the exception of glucose; the solution was filter sterilized] to remove any residual FBS. Next, the cells were preincubated for 10 min at room temperature in 0.3 ml of warm Tyrode solution to acclimatize with the assay buffer and remove cellular glucose (14). The glucose uptake was initiated by the addition of assay buffer containing 1 μCi/ml 2-deoxy-d-[1-14C]glucose (GE Health Care) in tyrode buffer in the presence or absence of myostatin (0.2, 0.4, 4, and 40 nM; Research Diagnostics, division of Fitzgerald Industries International). Also, in a separate assay, the effect of myostatin function blocker, follistatin (0.4, 4, and 40 nM; R&D Systems) was examined on glucose uptake by the maximal concentration regime of myostatin of 40 nM. The ability of follistatin to block myostatin effects and produce any direct effects on glucose uptake was also examined with follistatin treatment alone (0.4, 4, and 40 nM). Glucose uptake studies were conducted for 20 min in a humidified 37°C incubator maintained at 5% CO2 and terminated by removal of the assay buffer followed by brief washes with cold tyrode buffer and chilling on wet ice. Radioactivity was precipitated with 0.4 ml ice-cold 1 M NaOH, and the protein levels were determined by bicinchoninic acid assay (Sigma Chemical). Radioactivity was determined by adding 3.5 ml of Starscint scintillation fluid (Perkin-Elmer Life and Analytical Sciences, Boston, MA) to the lysed samples and counted in Quantsmart 1.31 software in a Tri-Carb 2900TR Liquid Scintillation Analyzer (Perkin-Elmer Life and Analytical Sciences). Uptake was calculated as counts per minute per microgram of protein and expressed as means ± SE of four determinations.
Protein expression studies.
The molecular mechanism of myostatin action was examined on BeWo cells. Initially cells were cultured in the absence of FBS following identification of myostatin immunoreactive bands (100, 40, and 28 kDa) in the FBS supplement used for cell culture (data not shown); also, subsequent replacement with 0.1% bovine-γ globulins, documented elsewhere (7, 16), produced aberrant BeWo cell morphology seen with serum starvation (24). Subsequently, BeWo cells were seeded in six-well plates at densities of 2.5 × 105 cells/well in regular cell culture media; following 24 h after seeding, cells were briefly washed and replaced with 10% FBS-culture media with or without 4 nM myostatin daily for 3 days. Recombinant myostatin was not added to control wells. Proteins were extracted as described above for Western blotting following 24, 48, and 72 h of incubation periods, with treatments added in triplicates.
Statistical analyses and presentation of data.
Data were analyzed using ANOVA on ranks with post hoc Tukey's test and to make statistical comparisons. Individual t-tests were used for comparisons between groups. A P value <0.05 was considered significant, when compared with controls. Statistical analysis was done using the GraphPad Prism statistical software package.
Myostatin protein and mRNA expression in placental choriocarcinoma cell lines.
Myostatin expression in the BeWo, JEG, and JAr cell lines was determined using Western blotting and RT-PCR. A representative Western blot shows myostatin immunoreactive bands identified at ∼50, 30, and 16 kDa, which correspond to the precursor, dimer, and monomer, respectively, and are similar to sheep muscle, which was used as a positive control (Fig. 1). There was greater abundance of the COOH terminus/monomer (∼16 kDa) than the dimers in JAr, JEG, and BeWo placental cell lines when compared with sheep muscle. To further demonstrate that myostatin is present in these placental cell lines, RT-PCR was performed on the total RNA extracts. The expected amplified ∼500-bp product was identified from the cDNA pool derived from the placental cell lines BeWo, JEG, and JAr (Fig. 2) using the specifically designed PCR primers to amplify the processed portion of myostatin (15, 33), which is similar to mouse muscle, the positive control. The levels of myostatin PCR product in the placental cell lines were lower than in mouse muscle. We chose to use BeWo cells for further experiments since they are a commonly used in vitro placental transport model.
Myostatin inhibits glucose uptake in BeWo cells.
Myostatin inhibited glucose uptake in BeWo cells in a concentration-dependent manner following 20 min of treatment after 24 h of culture (Fig. 3A). This inhibition was significant at the higher concentrations of 4 and 40 nM of myostatin treatment (P < 0.001). The levels of glucose uptake decreased with time irrespective of treatment (P < 0.001) as cells proliferated and greater cell densities were achieved (Fig. 3A).
Follistatin blocks myostatin-mediated effects on BeWo cells.
Various concentrations of recombinant human follistatin, an inhibitor of myostatin action (1, 19), were tested to determine the effect of follistatin on the inhibitory effect of 40 nM myostatin on glucose uptake. Follistatin reversed myostatin inhibition of glucose uptake to control levels, and a significant reversal was found with a lower follistatin treatment of 0.4 nM after 24 h of culture (P < 0.05; Fig. 3B).
Follistatin treatment alone increased glucose uptake (0.4 and 4 nM, P < 0.001, 40 nM P < 0.05) after 24 h of culture (Fig. 3C).
Exogenous myostatin has no effect on ActRII A/B proteins level.
To examine whether myostatin treatment could alter the protein level of ActRII A and ActRII B, BeWo cells were treated with recombinant myostatin for 24, 48, and 72 h in culture. In BeWo cells, ActRII A/ActRII B protein expression levels were unaltered with 4 nM myostatin treatment for 24, 48, and 72 h in culture (Fig. 4B). Immunoreactive ActRII A/B protein bands in BeWo cell line were identified at 47, 50, and 55 kDa, which was similar to that found in SKOV-3, an ovarian cancer cell line used as positive control (Fig. 4A). The strongest immunoreactive band was the 47 kDa in both the BeWo samples and the positive control.
Exogenous myostatin has no effect on GLUT transporter protein levels.
Expression of glucose transporter proteins GLUT1, 3, and -4 was measured to examine whether myostatin directly affects glucose transporters when affecting glucose uptake in BeWo cells. GLUT1, GLUT3, and GLUT4 immunoreactive bands were detected at ∼50 kDa, with GLUT1 and -3 as a doublet (Figs. 5A, 6A, and 7A). There was no significant effect on the protein expression levels of GLUT1, -3, and -4 following incubation with 4 nM myostatin vs. controls with no myostatin for 24, 48, and 72 h in culture. However, the abundance of GLUT1 and -4 decreased, whereas that of GLUT3 increased with time in culture (Figs. 5B, 6B, and 7B).
The results from the present study demonstrate that myostatin protein and mRNA are expressed in the placental choriocarcinoma cell lines BeWo, JEG, and Jar; however, the levels were lower than those found in muscle tissue, the positive control. There was a differential expression level of myostatin protein, with the relative levels of the monomer being higher in the cell lines than that found in muscle; this may be because of glycosylation and/or posttranslational modifications associated with cell lineage and target tissue (11). Previous work from this laboratory has shown that human placenta expresses myostatin protein and mRNA (31). Although myostatin was identified in other placental cell lines, BeWo cells were used for studying myostatin functions since they closely resemble placental trophoblast cells in vitro (23) and have also been studied extensively as an in vitro model for placental transport studies (6, 8, 9, 23, 25, 32, 42, 49, 50).
In this study, it was shown that short-term glucose uptake was inhibited by myostatin in BeWo cells in a concentration-dependent manner. This is in agreement with myostatin's role in other tissues, since systemic glucose levels are elevated in wild-type nude mice overexpressing myostatin from CHO-myostatin tumors (21). Also, knockout of myostatin gene in obese mice models partially suppressed adipogenesis and abnormal glucose metabolism (30). In addition, mice overexpressing myostatin propeptide, a functional inhibitor of myostatin put on a high-fat diet, had increased muscle growth, reduced hyperglycemia, and maintained normal blood glucose levels and insulin sensitivity compared with wild-type mice (52). Furthermore, inhibition of myostatin activity has been suggested to improve insulin sensitivity and increase glucose uptake by upregulation of glucose transporters like GLUT1 and GLUT4 in muscle and fat tissues (21). These findings in a placental cell line are not, however, reflected in the only study to date in fresh placental tissue (31) in which we reported that myostatin increases human placental glucose uptake, using an explant system. The disparity in these research findings may be attributed to the use of a fresh placental explant, which involves a complex interaction of various cell types within the placenta and may regulate trophoblast cell behavior (5, 10). Furthermore, differences between primary tissue cultures and cell lines derived from cancer tissue can also account for the different results. The significant responsiveness to myostatin seen early in culture might be associated with cell proliferation and/or differentiation state of BeWo cells in culture at that time. Endogenous and exogenous myostatin are known to elicit a differential response in cultured myoblast cell lines; the present study possibly mimics in vivo paracrine or endocrine effects of myostatin (39). It is possible in this study that the exogenous myostatin either alone or in combination with endogenous myostatin may have contributed to the inhibition of glucose uptake. As in the case of muscle regeneration following myotrauma, dramatic changes in the local levels of myostatin occur during the wound healing process (18), although systemically the levels of myostatin are kept latent complexed with inhibitory proteins such as follistatin (19). Alternatively, the seeding density has altered the transport capacity of these cells, since BeWo cell seeding density has been shown to influence their permeability properties (23). The overall decrease in glucose uptake may be because of changes in transporter density on the BeWo cell membrane surface that are involved in glucose uptake and associated morphological changes as cell numbers increase. Myostatin has been shown to directly regulate metabolic rates of proliferating myoblast and muscle fiber number via inhibition of cell proliferation in a tissue-specific manner (3, 46, 47).
Treatment with follistatin overcame the inhibitory effect of myostatin on glucose uptake in BeWo cells. Follistatin is known to block the effects of myostatin in muscle (1, 12, 19). However, there is also a significant increase in glucose uptake with follistatin treatment alone, which might indicate that follistatin inhibits the actions of locally produced myostatin, that follistatin is inhibiting other members of the TGF-β family (4, 22, 35, 51), or that follistatin has a direct effect on glucose uptake. In addition, mice overexpressing follistatin and the ActRII A/B receptor, have dramatic increases in muscle mass, which is much greater than that seen in myostatin null mice (19). Also, mice deficient in follistatin exhibit multiple craniofacial defects and imminent death after birth; these effects are contrary to that observed with lack of TGF-β members like activins and myostatin that can interact with follistatin (26–28). The lack of dose dependency when treating with follistatin may suggest that higher dose of 40 nM of follistatin is a saturating dose and unable to stimulate glucose uptake when compared with the lower dose of 0.4 nM. It is known that a follistatin dose of 100 ng/ml, which is equivalent of 4 nM, in cytotrophoblast cell cultures blocks both exogenous and endogenous activins, another member of the TGF-β family (4). Alternatively, higher treatment doses of follistatin may have activated endogenous feedback mechanisms to regulate follistatin levels, independent of myostatin/other TGF-β member's effects on glucose uptake. Follistatin is known to exist as several isoforms, as is the recombinant follistatin used in the present study, with different kinetics and binding affinities for various TGF-β members (40, 45). Perhaps these follistatin-TGF-β interactions are occurring over the lower follistatin dose. Based on cross-linking studies, myostatin signaling is proposed to be mediated by binding to ActRII A and B (19, 20, 48). There was no effect of myostatin treatment on the expression levels of ActRII A/B proteins, which might suggest ActRII A/B are not solely regulated by myostatin levels and other ligands are involved (20). Inhibition of myostatin activity using the soluble ActRII B/Fc enhanced muscle mass in wild-type mice and to a lesser extent in myostatin null mice (20). Furthermore, muscle mass was increased significantly in wild-type mice administered the inhibitor, ActRII B/Fc; however, in mice lacking ActRII or ActRII B, the increase in muscle mass was only reduced and not eliminated (20).
There was also no significant effect of myostatin treatment on the glucose transporter proteins (GLUT1, -3, and -4), which suggests that myostatin-mediated effects on glucose uptake are independent of GLUT transporters. Irrespective of treatment, the protein levels of GLUT1 and -4 decreased, whereas GLUT3 increased with culture incubation time. The decline in the glucose transporter GLUT1 and -4 levels explains at least in part the decrease in glucose uptake associated with culture time. The rise in the levels of glucose transporter GLUT3 may also be an adaptive mechanism to meet energy requirements with increasing cell density (17). In the myostatin-treated groups, GLUT1 and -4 levels seem to be below control groups; the lower levels suggest that glucose transport in BeWo cells may be regulated by myostatin, which is in agreement with findings from Liang et al. (21), who reported that myostatin downregulates glucose transporters GLUT1 and GLUT4 in tissues such as muscle and fat.
This study has clearly shown that myostatin inhibits glucose uptake in BeWo cells. Given that myostatin regulates nutrient uptake in other tissues, these interesting observations suggest that local placental myostatin may control placental glucose metabolism within the fetoplacental-maternal unit.
This work was supported by the National Research Centre for Growth and Development, New Zealand.
We thank Dr. Andrew Shelling and Anna Ramachandran for ActRII A/B antibodies and Claire Osepchook, Prudence Grandison, Nagarajan Kannan, Peter Shepherd, and Denis Evseenko for expert technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 by American Physiological Society