|Home | About | Journals | Submit | Contact Us | Français|
Odorant receptors (ORs) in the olfactory epithelium bind to volatile small molecules leading to the perception of smell. ORs are expressed in many tissues but their functions are largely unknown. We show multiple ORs display distinct mRNA expression patterns during myogenesis in vitro and muscle regeneration in vivo. Mouse OR23 (MOR23) expression is induced during muscle regeneration when muscle cells are extensively fusing and plays a key role in regulating migration and adhesion of muscle cells in vitro, two processes common during tissue repair. A soluble ligand for MOR23 is secreted by muscle cells in vitro and muscle tissue in vivo. MOR23 is necessary for proper skeletal muscle regeneration as loss of MOR23 leads to increased myofiber branching, commonly associated with muscular dystrophy. Together these data identify a functional role for an OR outside of the nose and suggest a larger role for ORs during tissue repair.
Odorant receptors (ORs) are G-protein-coupled receptors (GPCR) expressed within the olfactory epithelium of the nose where they function as chemosensors to detect small molecules we perceive as smell. ORs are the largest receptor family in mammals comprising 3-5% of all genes (Young and Trask, 2002), with approximately 913 potential OR genes in the mouse genome, and 390 in the human genome (Olender et al., 2008; Zhang and Firestein, 2002). Microarray analyses suggest ORs are expressed in many tissues (Feldmesser et al., 2006), but functions of these receptors are known for only three: chemosensing in olfactory epithelium, proliferation in prostate cancer and chemotaxis in sperm of humans and mice (Fukuda et al., 2004; Neuhaus et al., 2009; Spehr et al., 2003).
One of the ORs that regulates sperm chemotaxis is mouse OR23 (MOR23, Olfr16) (Fukuda et al., 2004). We unexpectedly observed an increase in MOR23 expression in a microarray analysis performed during myogenesis of primary cultured mouse muscle cells (unpublished data). This increase occurred when muscle cells were undergoing extensive cell migration, adhesion and fusion to form multinucleated cells. In adult skeletal muscle, myogenesis occurs to regenerate muscle after injury. This process is dependent on satellite cells, which are normally quiescent, but in response to injury proliferate, and their progeny myoblasts differentiate and fuse with each other or with existing myofibers to restore normal tissue architecture. Migration and cell-cell adhesion are necessary for muscle cell differentiation (Kang et al., 2004) and fusion in vitro (Jansen and Pavlath, 2006). Therefore, identifying the molecules that control muscle cell adhesion and migration may reveal potential molecular targets for improving muscle regeneration. As MOR23 regulates migration of sperm, we hypothesized that ORs may affect similar processes during myogenesis.
Here, we demonstrate that MOR23 is an important regulator of myogenesis in vitro and in vivo through affects on cell migration and adhesion, influencing downstream fusion events. In addition, our data suggest a MOR23 ligand is secreted by muscle cells in vitro and is also present in muscle tissue. Furthermore, a number of ORs were also expressed by regenerating muscle, which may suggest a larger role for ORs in tissue repair.
Multinucleated myofibers form through fusion of muscle cells, which requires cell migration and adhesion factors (Jansen and Pavlath, 2008). Myogenesis in vitro occurs in distinct phases (Fig. 1A): proliferating myoblasts terminally differentiate to become fusion-competent myocytes in differentiation media (DM). By 24 hrs in DM, myocytes fuse with one another to form small, nascent myotubes with few nuclei. Further rounds of fusion create mature myotubes with many nuclei, by 48 hrs in DM. By microarray analysis, MOR23 mRNA was upregulated at 24 hrs of myogenesis (unpublished data). To validate the expression of MOR23 and also to determine whether other ORs are expressed during myogenesis in vitro and in vivo, we performed Real Time RT-PCR. Given the large numbers of mouse ORs, we choose 18 ORs previously identified in microarray screens of skeletal muscle (Beggs et al., 2004; Melcon et al., 2006; Tseng et al., 2002; Zhao et al., 2002). Olfactory bulb RNA was used as a positive control for OR mRNA (Supp. Table 1). Thirteen ORs were expressed during myogenesis (Supp. Table 2 and 3, Fig. 1). The majority of ORs expressed in vitro were most highly expressed at 0 hrs, when the muscle cells are proliferating myoblasts (Supp. Table 2, Fig. 1B, C). A smaller number of ORs were increased with terminal differentiation, and most highly expressed at 24 hrs in DM, when myocytes are fusing. Interestingly, no ORs were upregulated or exclusively expressed at 48 hrs in DM, when most cells have already fused.
Expression of these ORs was also highly regulated during skeletal muscle regeneration (Supp. Table 3, Fig. 1D, E). Only one OR was most highly expressed in uninjured muscle tissue, whereas the majority of ORs were upregulated at day 5 of regeneration, when muscle cells are proliferating, differentiating and fusing into nascent myofibers. The ORs expressed at day 5 were those detected at both 0 and 24 hrs in vitro. Two ORs were most highly expressed at day 10 after regeneration, when most muscle cell fusion had already occurred. The mRNA for most ORs was expressed both in vitro and in vivo, except for two. Olfr70 was expressed at very low levels in vitro and was not apparent in vivo. This discrepancy may be due to problems with sensitivity of detection or time points chosen in vivo or may be related to cell culture conditions. Olfr49, expressed only in vivo, might be expressed by cells other than muscle, or in a sub-population of muscle cells that was not retained in vitro. Together, our results demonstrate multiple ORs display distinct expression patterns during myogenesis, suggesting that individual ORs may have non-redundant functions. ORs increased after differentiation and during fusion may have similar functions to ORs in sperm and olfactory neurons, regulating migration or adhesion. This subset of ORs was of specific interest as they might have conserved functions in muscle.
To investigate the role of a specific OR, MOR23 was selected, as its mRNA was upregulated during times of cell fusion and its expression was verified by use of a second primer pair (Supp. Table 2 and 3, Fig. 1B, D, ,2A,2A, SFig. 2); similarly immunoblot analysis revealed increased amounts of MOR23 protein at 24 hrs in DM (Fig. 2B). The MOR23 antibody likely cross-reacts with the closely related olfactory receptor, Olfr1403, which shares 85% homology to MOR23 and whose mRNA is most highly expressed in myoblasts (SFig. 1A, B). Based on the mRNA expression pattern for Olfr1403 most of the antibody reactive band at 24 hrs in DM likely reflects MOR23 and not Olfr1403.
In both olfactory neurons and sperm activation of ORs leads to similar downstream signaling events: activation of Gαolf, a G-protein specific to odorant receptors, initiates signal transduction through membrane adenylyl cyclase III (mACIII) (Spehr et al., 2004). We observed both isoforms of Gαolf, and mACIII in muscle cells at 24 hrs in DM by immunoblotting (Fig. 2B) suggesting that canonical OR signaling may occur within muscle cells.
Given that MOR23 is upregulated when myocytes are migrating and fusing, we hypothesized that MOR23 may have a conserved migratory function in myocytes. Thus, we examined the migration of myocytes in which MOR23 expression was knocked down by siRNA. First, to determine efficiency of knockdown, RNA was isolated from cells retrovirally infected with control or one of two distinct MOR23 siRNAs at 24 hr in DM and RT-PCR was performed; no MOR23 mRNA was observed with either MOR23 siRNA (SFig. 2). Since both MOR23 siRNAs resulted in similar knockdown, subsequent experiments were carried out using MOR23 siRNA 3, unless otherwise indicated. To determine whether expression of other ORs was affected by MOR23 siRNA, RNA was isolated from cells infected with control or MOR23 siRNA and Real Time RT-PCR was performed; no affect on mRNAs for other ORs was observed with MOR23 siRNA (Supp. Table 4). By immunoblotting, cells infected with MOR23 siRNA exhibited at least a 58% decrease in MOR23 protein although some of the residual antibody reactive band could have been Olfr1403, and hence a greater decrease in MOR23 protein actually occurred (Fig. 2C). To determine if MOR23 signaling was abrogated due to MOR23 siRNA, control and MOR23 siRNA cells were exposed to the synthetic MOR23 ligand lyral (Fukuda et al., 2004), and the levels of cAMP were detected by ELISA (Fig. 2D). Levels of cAMP were increased 60% in control siRNA cells and MOR23 siRNA abrogated this effect, indicating MOR23 protein levels are significantly decreased in MOR23 siRNA cells as they do not increase cAMP in response to lyral.
Subsequently, Boyden chamber experiments were performed using control and MOR23 siRNA cells with conditioned media collected from muscle cells at 24 hrs in DM as the chemoattractant (Fig. 2E). A 55% decrease in migration to CM was observed with each MOR23 siRNA. In order to determine if MOR23 siRNA cells had a general motility defect that could account for these results, time-lapse microscopy was performed at 24 hrs in DM in regular culture dishes (Fig. 2F). MOR23 siRNA did not significantly affect cell velocity when compared to control siRNA cells. To determine if removal of any OR was sufficient to alter migration, we focused on the following two ORs: Olfr1508 and Olfr1403; Olfr1508 was the only OR to share a similar expression pattern to MOR23, both in vivo and in vitro, (SFig. 3A, B), and Olfr1403 has the highest homology to MOR23. Knockdown of either Olfr1403 (SFig. 1D) or Olfr1508 (SFig. 3C) did not affect migration to CM (SFig. 1E, SFig. 3D). Hence, decreased migration to CM was specific to loss of MOR23 and not simply due to loss of any OR.
Migration was also examined in cells expressing recombinant MOR23. First, to determine whether recombinant MOR23 was properly trafficked to the plasma membrane, cell surface biotinylation was used. The subsequent biotin pull-down was assayed for MOR23 by immunoblotting; a 3-fold increase in cell surface MOR23 was observed (Fig. 2G). We performed Boyden chamber experiments using diluted CM to determine if these MOR23 over-expressing (OE) cells were more sensitive to CM. As the CM was diluted, less migration occurred in both control and MOR23 OE cells; however, the MOR23 OE cells exhibited higher levels of migration at all dilutions (Fig. 2H). Thus, increasing MOR23 levels alone is sufficient to change the migratory behavior of myocytes. Together, these data indicate that MOR23 regulates myocyte migration and suggest that an endogenous MOR23 ligand is secreted by fusing muscle cultures.
To discern whether a MOR23 ligand is present in muscle tissue in vivo, Boyden chamber experiments were performed using crushed muscle extract. CME resulted in a 9-fold increase of myocyte migration, which was abrogated in MOR23 siRNA cells (Fig. 2I). In order to test the migration of MOR23 OE cells, a different optimal concentration of CME was used based on preliminary dose response curves. With MOR23 OE a 3-fold increase in myocyte migration occurred (Fig. 2J). These data suggest that a ligand(s) or its precursor is present within muscle tissue prior to induction of muscle regeneration.
The decreased migration of MOR23 siRNA cells to CM likely results from a defect in directed migration as no general motility defect was observed. To address this question, we used the Dunn chemotaxis chamber to determine if MOR23 siRNA cells respond to a gradient of lyral. Time-lapse microscopy was performed for 3 hrs, the paths of individual cells tracked and the final location of each cell in relation to its origin determined (Fig. 3A). Myocytes migrated toward lyral, yielding migration paths that ended on the lyral side of the chamber. To statistically analyze the pattern of migratory paths in response to lyral, directional data were summarized in circular histograms. Rayleigh's statistical test for clustering revealed that control cells (Fig. 3B) but not MOR23 siRNA cells (Fig. 3C) migrated toward lyral. A rescue experiment in which myocytes expressed both MOR23 siRNA and recombinant MOR23 (SFig. 4) demonstrated directed cell migration toward lyral (Fig. 3D), indicating that migration to lyral is specific to MOR23. To determine whether migration towards lyral requires mACIII, an inhibitor of membrane adenylyl cyclases, SQ22536, was used (SFig. 5). Addition of the mAC inhibitor abrogated migration towards lyral (Fig. 3E), indicating that canonical OR signaling is used in lyral-directed migration. As a control, to determine if lyral-dependent migration was specific to MOR23, myoblasts which express the highly homologous Olfr1403 and the lyral-responsive Olfr15 (Saito et al., 2009) but not MOR23, were tested in Dunn chambers. These cells did not undergo directed migration towards 10-7M lyral (Fig. 3F). In addition, knockdown of Olfr1403 or Olfr1508 in myocytes did not affect migration to lyral (SFig. 1F, SFig. 3E), revealing that migration of myocytes to lyral is specific to MOR23 and occurs in a directed manner.
Recent evidence suggests that ORs may regulate axon guidance of olfactory neurons via downstream changes in the expression of cell-cell adhesion molecules (Serizawa et al., 2006). To determine whether loss of MOR23 alters cell adhesion, a solution based cell adhesion assay was utilized in which control or MOR23 siRNA myocytes were suspended in media and aliquots taken over 60 min to count the number of adhered and un-adhered cells using phase-contrast microscopy. Myocytes infected with MOR23 siRNA did not efficiently adhere to one another (Fig. 4A). Although no difference in adhesion was observed at early time points, at later time points, a significant increase occurred in un-adhered MOR23 siRNA cells (Fig. 4B). At later times MOR23 siRNA cells remained in small clusters compared to control siRNA cells (Fig. 4C). Cell viability was ~99% at 60 min in all conditions (data not shown). Canonical OR signaling is also necessary for axonal migration and adhesion (Imai et al., 2006), and to determine whether it regulates cell-cell adhesion of myocytes, cells were treated with mAC inhibitor SQ22536, or vehicle (Fig. 4D). At early time points, no difference was observed, however at later time points a significant increase occurred in un-adhered cells with SQ22535 (Fig. 4E). Furthermore, SQ22536-treated cells formed smaller clusters than vehicle (Fig. 4F). Knockdown of Olfr1403 or Olfr1508 did not affect adhesion (SFig. 1G, SFig. 3F). Hence, adhesion affects were specific to loss of MOR23 and not simply due to loss of any OR. Together, these data suggest that MOR23 signaling affects adhesion of myocytes to one another likely through a downstream adhesion molecule.
To test whether MOR23-dependent migration and/or adhesion is involved in muscle cell differentiation or fusion, we examined myogenesis in vitro in cells containing one of two MOR23 siRNAs. After 24 or 48 hr in DM, cells were immunostained using an antibody against embryonic myosin heavy chain (eMyHC). After 24 hrs in DM, both MOR23 siRNA cultures contained fewer and smaller nascent myotubes compared to control (Fig. 5A). This defect in myotube formation was not due to an affect on differentiation, as measured by the percentage of nuclei found in eMyHC+ cells (Fig. 5B), nor to a decrease in the total number of nuclei (data not shown). Rather, MOR23 siRNA myocytes exhibited a clear defect in cell fusion (Fig. 5C-E). The fusion index, as well as the number of myotubes, was transiently decreased 33% and 26% respectively in MOR23 siRNA cultures (Fig. 5C, D). A 15% decrease in myonuclear number was also noted at 24 hrs in DM in MOR23 siRNA cultures, which remained virtually unchanged at 48 hrs (Fig. 5E) and 72 hrs (data not shown). The transient fusion defects in MOR23 siRNA cultures may be due to compensation from other factors which regulate fusion.
To determine whether the impaired ability of MOR23 siRNA cultures to fuse is specific to the loss of MOR23; assays using MOR23 OE were also performed. MOR23 OE resulted in an increased number of nascent myotubes at 24 hrs in DM (Fig. 5F), with no change in differentiation (Fig. 5G). Fusion assays revealed a 30% increase in the fusion index and a 40% increase in the number of myotubes at 24 hrs in DM, with smaller increases in both parameters still noted at 48 hrs and 72 hrs (Fig. 5H, I and data not shown). Myonuclear number was also increased ~10% at both time-points in MOR23 OE cultures (Fig. 5J). Together these data suggest that MOR23-dependent migration and/or adhesion contributes to the fusion of myocytes to form nascent myotubes.
Changes in myonuclear number with MOR23 siRNA or OE at 48 hrs suggest that MOR23-dependent migration and/or adhesion may also regulate formation of mature myotubes from fusion of myocytes with nascent myotubes. Since the formation of nascent myotubes was affected in previous assays, we could not independently determine a role for MOR23 in this later stage of fusion. To analyze if MOR23 affects the fusion of myocytes with nascent myotubes (Fig. 1A), control and MOR23 siRNA nascent myotubes and myocytes were created separately, labeled with green or orange fluorescent dye, co-cultured for 24 hrs (Fig. 5K), and the number of dual labeled myotubes was quantified. With MOR23 siRNA in myocytes the percentage of myotubes with dual label was decreased 20% compared to control (Fig. 5L), whereas MOR23 siRNA in myotubes had no effect on dual labeling. To determine whether increased expression of MOR23 on myocytes could enhance fusion, control or MOR23 OE myocytes were co-cultured with control nascent myotubes (Fig. 5M). MOR23 OE in myocytes increased dual labeling 35% (Fig. 5N), indicating MOR23 expression on myocytes alone is sufficient to increase fusion of myocytes with nascent myotubes. Together, these data indicate that MOR23 is functionally required on myocytes. Therefore, MOR23-dependent processes can affect myonuclear addition at later stages of myogenesis.
To determine the detailed expression pattern of MOR23 during adult regenerative myogenesis, RNA was isolated from mouse gastrocnemius muscles at different times after injury. Using Real Time RT-PCR, MOR23 mRNA was induced 4-fold during days 3-7 post-injury (Fig. 6A), a time of extensive cell fusion. By 14 days post-injury, when regeneration is nearly complete, MOR23 mRNA levels had returned to uninjured control levels. MOR23 protein levels were increased at day 5 post-injury, a time when fusion is ongoing and nascent myofibers are present, as determined by immunoblot (Fig. 6B, G). Olfr1403 mRNA was not increased during muscle regeneration at these times (SFig. 1C). Sections from gastrocnemius muscles at day 5 post-injury were immunostained with an antibody against MOR23 (Fig. 6C). Regenerating myofibers, identified by their centrally located nuclei and small size, contained MOR23. To determine whether mononucleated muscle cells express MOR23 during regeneration, cells were isolated from regenerating muscle and myogenic cells (α7-integrin+CD31-CD45- (Kafadar et al., 2009)) were separated from immune and endothelial cells (α7-integrin-CD31+CD45+) by FACS (SFig. 6). Real Time RT-PCR revealed expression of MOR23 mRNA in α7-integrin+CD31-CD45- cells, which also expressed myogenin mRNA, confirming their myogenicity (Fig. 6D). MOR23 mRNA was not apparent in the α7-integrin-CD31+CD45+ fraction; this fraction did express Mac1 mRNA, suggesting many of these cells are immune cells. Furthermore, immunostaining for MOR23 in α7-integrin+CD31-CD45- cells indicated ~86% of muscle cells express MOR23 (Fig. 6E). The myogenic purity of these cells was >99% as determined by the expression of Myf5. Thus, MOR23 expression is induced in myogenic cells during muscle regeneration when myofibers are forming. Overall these data suggest that the induction of MOR23 expression may be tied to migration and/or adhesion events that occur during cell fusion in muscle regeneration.
To analyze MOR23 function during muscle regeneration, gastrocnemius muscles were induced to regenerate with BaCl2 two days prior to electroporation of either MOR23 siRNA or OE plasmids. Induction of regeneration allowed myoblasts to take up plasmid, as determined in preliminary experiments. Gastrocnemius muscles were isolated 5, 10 or 20 days post-electroporation (Fig. 6F). Knockdown of MOR23 protein was determined by immunoblot at day 5, and an 85% reduction of MOR23 was observed (Fig. 6G), suggesting that the majority of the immunoreactive band is MOR23 and not Olfr1403. Although little to no difference existed in proliferation or differentiation in muscles that received MOR23 siRNA compared to control (SFig. 7), abnormalities in muscle regeneration were noted in MOR23 siRNA muscles at all time points (Fig. 6H). The mean cross-sectional area (XSA) of regenerating myofibers in MOR23 siRNA muscles was decreased 26-38% (Fig. 6I) with a corresponding 30% increase in the number of myofibers smaller than 1500μm2 at day 20 (Fig. 6J). In addition, the number of regenerating myofibers per field increased ~50% with MOR23 siRNA (Fig. 6K). In contrast, MOR23 OE in vivo resulted in a 5-fold increase in MOR23 mRNA (SFig. 8), and increased myofiber XSA ~50% (Fig. 6L), with a corresponding 39% decrease in the number of myofibers smaller than 1500μm2 at day 20 (Fig. 6M). In addition, the number of regenerating myofibers per field decreased 21-36% with MOR23 OE (Fig. 6N). These data suggest that optimal muscle regeneration is dependent on MOR23.
Small clusters of myofibers in MOR23 siRNA muscles were observed, which suggested that myofiber branching may have occurred with loss of MOR23. Myofiber branching is characterized by a single myofiber having a plasma membrane contiguous with several smaller myofibers (Fig. 7A). Myofiber branching is an important phenomenon as it is increased with multiple injuries, aging, and muscular dystrophy (Ontell et al., 1982). The molecules which regulate this process are unknown, although cell-cell adhesion molecules are hypothesized to play a role fusing branched myofibers to form mature myofibers. We hypothesized that the increase in small myofibers observed in MOR23 siRNA muscles may be due to an increase in unresolved myofiber branches, yielding the appearance of many small myofibers rather than large myofibers apparent in control siRNA muscles. To determine whether MOR23 may regulate myofiber branching, cross-sections 170μm apart were isolated from gastrocnemius muscles 5 days after electroporation. Myofiber branching was quantified by determining the number of regenerating myofibers that were a single myofiber in one section and multiple fibers occupying the same area in a second section (Fig. 7B). Muscles with MOR23 siRNA exhibited a 40% increase in the percentage of branching myofibers (Fig. 7C), and more branches per myofiber (Fig. 7D). MOR23 OE did not result in changes in myofiber branching at day 5 (Fig. 7E). In order to visualize the entire myofiber, single myofibers were isolated to determine the extent of branching. We were unable to isolate intact regenerating myofibers at day 5; however myofiber branches could be visualized at day 10 and 20 using phase-contrast microscopy and DAPI-staining was used to identify regenerating myofibers by the presence of centrally located nuclei (Fig. 7F). Myofibers with multiple tracks of central nuclei, but unbranched cytoplasm, are mature regenerated myofibers (Fig. 7F, arrowheads). Muscles electroporated with MOR23 siRNA exhibited a 60% increase in regenerating myofibers with branches (Fig. 7G) and 90% more branches per myofiber (Fig. 7H). In contrast, a 50% decrease in the number of branched myofibers (Fig. 7I) occurred in MOR23 OE muscles at day 20 with 70% fewer branches per myofiber (Fig. 7J). These data suggest that myofiber branching is regulated by MOR23, possibly through its effects on migration and adhesion of muscle cells.
In many systems, tissue repair requires cell migration and adhesion. Our results suggest these processes are also necessary for proper regeneration of muscle tissue. We show multiple ORs are expressed by muscle cells during myogenesis in vitro and muscle regeneration in vivo. In depth studies on one of these ORs, MOR23, revealed a role in regulating migration and adhesion of muscle cells, which are critical for cell fusion. Importantly, these studies identify an unexpected functional role for MOR23 in skeletal muscle regeneration, demonstrating a novel function for an OR in tissue repair.
MOR23 was expressed by terminally differentiated myocytes in vitro, functioning to regulate their migration. Interestingly, MOR23 also regulates migration of mouse sperm in vitro (Fukuda et al., 2004) and ORs are hypothesized to play a role in axonal growth cone migration (Feinstein et al., 2004), suggesting that MOR23 has a conserved function in migration of multiple cell types. Although synthetic ligands are known for many odorant receptors, endogenous ligands have not been identified. Both crushed muscle extract and conditioned media contained a ligand that stimulated myocyte migration in a MOR23-dependent manner. CME was prepared from uninjured muscle tissue, which indicates that the MOR23 ligand is present prior to the onset of muscle injury. The ligand may be present in an inactive form in normal muscle tissue and become activated by the “crushing” used in preparing CME. In vivo, HGF is expressed in an inactive form, which is expressed by muscle cells and then bound by the extracellular matrix until matrix metalloproteases cleave and release the active form (Allen et al., 2003). Alternatively, the ligand may be stored within myofibers and released upon myofiber degeneration. CM collected from differentiating muscle cultures in vitro also contained the MOR23 ligand, demonstrating that muscle cells themselves can be the source of the ligand. These results do not rule out that other cell types may also express the MOR23 ligand in vivo. Preliminary data indicate the MOR23 ligand is likely not sensitive to heat or Proteinase K digestion (data not shown). Identification of sources for the MOR23 ligand will allow future purification of an endogenous OR ligand from a readily available source, as opposed to follicular fluid, which is the only other hypothesized source of an endogenous OR ligand (Spehr and Hatt, 2004). If the ligand for MOR23 can be isolated, it may give insight into the nature of endogenous ligands for other odorant receptors, including the human odorant receptor found in sperm.
MOR23 also regulated cell-cell adhesion, indicating another conserved function for ORs between muscle and olfactory neurons. In olfactory neurons, OR signaling is required for expression of adhesion molecules, which correlates with altered axon guidance into specific glomeruli (Imai et al., 2006). However, changes in the adhesiveness of the olfactory neurons were not assayed. In contrast, we performed adhesion assays which demonstrated that loss of MOR23 or OR signaling decreased cell-cell adhesion, suggesting that MOR23 regulates adhesion indirectly, possibly through regulating expression of specific downstream adhesion molecules. Further studies are needed to determine which adhesion molecules may be affected by MOR23 signaling.
MOR23 expression was induced with muscle regeneration and manipulating the levels of MOR23 affected several aspects of muscle regeneration, including myofiber cross-sectional area, number and branching. Although branched myofibers have been noted in the literature for 100 years (Schmalbruch, 1976; Volkmann, 1893), no molecule is known to regulate their formation. We show that loss of MOR23 increased myofiber branching. Expression of MOR23 by mononucleated muscle cells and regenerating myofibers in vivo suggests both migration and adhesion could contribute to these effects. If myocyte migration is required for fusion with existing regenerating myofibers, then defects in migration would increase the formation of de novo myofibers. Alternatively, formation of small myofibers may normally occur, which adhere to one another, and then fuse to form larger myofibers, and when adhesion is disrupted, branched myofibers occur. Therefore, loss of MOR23 may disrupt myofiber formation on several levels. Since muscles containing unbranched myofibers are stronger than those with extensive branching (Chan et al., 2007) and MOR23 OE can decrease the occurrence of branching, MOR23 may be an important target for therapies directed at increasing strength of dystrophic muscle.
As MOR23 functions in the repair of skeletal muscle, and multiple ORs are expressed during myogenesis, other ORs may regulate similar or distinct processes. Recent evidence suggests that activation of an OR inhibits proliferation of prostate cancer cells (Neuhaus et al., 2009), therefore, ORs that were upregulated in myoblasts may regulate proliferation. Also, as muscle cells must interact directly with one another for terminal differentiation to occur (Krauss et al., 2005), ORs in addition to MOR23 may play a role in the migration or adhesion of these cells. As Olfr49 was expressed only in muscle tissue, other cell types present in muscle, such as immune cells, endothelial cells, fibroblasts, or neurons may also express ORs. Finally, ORs present in uninjured muscle or at later time points in regeneration may be necessary for quiescent satellite cells, either to keep these cells in their niche or to affect their proliferation state. Our data suggest that multiple ORs may regulate myogenesis, and therefore, may be necessary for proper tissue repair. Additional studies will be necessary to determine the specific effects of each of these ORs both in vitro and in vivo.
Interestingly, in the olfactory neuron field, only one OR is expressed in each cell (Touhara et al., 1999). As multiple ORs were expressed by muscle cultures, sub-populations of muscle cells may each express a single OR. Potentially the “one cell, one OR” hypothesis may not hold true for muscle cells, and several ORs are expressed by each muscle cell. A redundant system may exist, so that even if one OR signal is somehow blocked, other ORs may substitute. Alternatively, cells expressing multiple ORs may utilize specific ORs in a spatial-temporal manner. Additional studies may elucidate whether multiple populations of muscle cells each express a single OR, or whether each muscle cell expresses several ORs.
In summary, MOR23 is a key regulator of myogenesis through its actions on cell migration and adhesion, affecting downstream fusion in vitro and muscle regeneration in vivo. Cell migration within muscle is a major issue in transplantation of cells for treatment of muscular dystrophy (Skuk and Tremblay, 2003; Smythe et al., 2001). Study of the receptor-ligand pairs that regulate migration and/or adhesion of muscle cells may allow for more efficient therapeutic strategies. In addition, further studies on ORs may reveal additional unexpected functions in various tissues.
Adult mice between 8–24 weeks of age were used and handled in accordance with the IACUC guidelines of Emory University. To induce regeneration, gastrocnemius muscles of male C57BL/6 mice were injected with BaCl2 (O'Connor et al., 2007) and collected as described (Abbott et al., 1998).
Primary myoblasts were derived from the hindlimb muscles of Balb/C mice and cultures were >99% myogenic as assessed by MyoD immunostaining (Jansen and Pavlath, 2006). To induce differentiation, cells were switched to differentiation media (DM), immunostained with an eMyHC antibody (F1.652; Developmental Studies Hybridoma Bank) and analyzed as described (Jansen and Pavlath, 2006). At least 500 nuclei per condition were analyzed for each assay. Cell mixing experiments were performed as described (Jansen and Pavlath, 2006). The presence of dual label was analyzed in 50-100 myotubes with ≥3 nuclei in each experiment. For all experiments 3-6 independent isolates were analyzed.
Oligonucleotides (see Supp. Methods) encoding siRNA to MOR23 were cloned into the retroviral plasmid pSUPER.retro.puro according to OligoEngine protocol. The control plasmid contained a scrambled siRNA sequence. Retroviral production and infection were performed as described (Abbott et al., 1998) and cells were used 48 hrs later in experiments. The efficiency of retroviral-mediated gene transfer was >95%, based on cell survival in the presence of puromycin following two rounds of retroviral infection.
To generate a MOR23 retroviral expression vector, a cDNA fragment encoding a FLAG-Rho-MOR23 was subcloned from the pME18S construct (Katada et al., 2003) into the retroviral vector pTJ84 using the AvrII and SalI/XhoI sites (Abbott et al., 2000). pTJ84 was used as the control vector.
Real-time RT-PCR was performed on DNase treated total RNA, reverse transcribed to cDNA, amplified using primers from SABiosciences, and analyzed using the iCycler iQ Real-Time Detection System and software (Bio-Rad) as described (Bondesen et al., 2004). For conditions and controls, see Supp. Methods. MOR23 mRNA was quantified in reference to MOR23-pME18S plasmid standard and normalized to 18S rRNA levels.
Cells were harvested in RIPA-2 (O'Connor et al., 2007) with protease inhibitors (Mini-Complete; Roche) and immunoblots were performed as described (Friday and Pavlath, 2001) with primary antibodies (α-MOR23, GeneTex; α-Golf, α-mACIII and α-tubulin, Abcam, Inc.) and detected with appropriate secondary HRP-conjugated antibodies (Jackson ImmunoResearch Lab., Inc.).
Cell surface biotinylation was performed as described (Salazar and Gonzalez, 2002). Sulfo-NHS-SS-Biotin was used at 500μM (Pierce) and protein was isolated in RIPA-2. Immunoprecipitation was performed with NeutrAvidin Agarose Resins (Pierce) with 300μg of protein at 4°C overnight.
cAMP was detected using the CatchPoint kit from Molecular Probes. Myocytes were generated as for Boyden chambers and plated in 96-well plates at 1×105 cells/well. Cells were treated with either buffer or 10-4M lyral (International Flavors and Fragrances) for 15 min in triplicate.
Migration of muscle cells was quantified using time-lapse microscopy as described (Jansen and Pavlath, 2006). Images were recorded (QImaging Camera and OpenLab 3.1.4 software) every 5 min for 3 hours. Cell velocities were calculated in μm/hr using ImageJ software by tracking the paths of mononucleated cells. Approximately 20 mononucleated cells were tracked for each experiment. Boyden chambers were performed as described (Mylona et al., 2006). Primary myoblasts were seeded on 150-mm plates at low density (9 × 105 cells/plate) and switched to DM for 24 hrs to generate myocytes in the absence of myotube formation. Cells (7.5 × 104 cells in 200μL DM) were loaded in the upper wells of the Boyden chamber and incubated at 37°C for 5 hrs. Migrated cells were fixed, stained and counted. To prepare CM, myoblasts were incubated in DM for 24 hrs; the media, which had been “conditioned” with secreted factors, was then collected, filtered (0.45 μm), flash frozen, and stored at -80°C until use. Crushed muscle extract was created as described (Chen and Quinn, 1992) using gastrocnemius muscles from C57BL/6 mice (N=10) and stored at -80°C.
Dunn chambers were performed as described (Jansen and Pavlath, 2006). The outer well of the Dunn chamber was filled with DM or 10-7M lyral in DM. SQ22536 (Sigma) was used at 2.5mM. Directional analyses were performed using at least 15 cells per assay.
Cell-cell adhesion in suspension was analyzed by incubating myocytes at 2×105 cells in 2mL of DM after lifting using Cell Dissociation Buffer Enzyme-Free PBS-Based (Invitrogen). SQ22536 was used at 2.5mM. Duplicate 50μL aliquots of cells were taken at regular intervals for 60 min. This cell concentration was used to permit counting of individual cells within clusters using phase-contrast microscopy. Trypan blue staining was used at 60 min to determine cell viability.
FACS was performed as described (Kafadar et al., 2009) using cells isolated from day 5 regenerating gastrocnemius muscles. Cells were sorted using a BD FACSVantage SE and placed into Trizol Reagent (Invitrogen) for RNA isolation or plated for immunostaining. For detailed methods, see supplement.
MOR23 immunostaining was performed on muscle sections or muscle cells isolated by flow cytometry from gastrocnemius muscles 5 days post-injury. After blocking, sections and cells were incubated overnight at 4°C with MOR23 antiserum or purified rabbit IgG (Genetex Inc.) diluted 1:100 in blocking buffer. Immunostaining was visualized using the TSA Rhodamine Tyramide Signal Amplification kit (PerkinElmer). Nuclei were counterstained with 25 μM 4,6-diamidino-2-phenylindole in 0.1% BSA. For detailed methods, see supplement.
Gastrocnemius muscles of C57BL/6 mice were injected with BaCl2 to induce muscle regeneration. Two days later the muscles were injected with the indicated plasmid (25μg DNA; 1μg/μL in PBS) and electroporated as described (O'Connor et al., 2007). Post-electroporation, muscles were collected and homogenized in RIPA-2 for protein isolation or frozen and sectioned as described (Abbott et al., 1998). Single myofibers were isolated from previously injured and electroporated gastrocnemius muscles as described (Kafadar et al., 2009). Serial 14μm cross-sections were stained with hematoxylin and eosin and analyzed by quantifying myofiber number and cross-sectional area (XSA) for three 307,200μm2 fields in the core of each regenerating muscle (O'Connor et al., 2007). In preliminary experiments, electroporation of a lacZ plasmid was utilized to determine efficiency of gene delivery to MyoD+ cells in the gastrocnemius muscle; 70% of adherent cells were lacZ+, 80% were MyoD+. All analysis were performed blinded with N=4-8 mice per condition and 90-100 myofibers per mouse.
To determine significance between two groups, comparisons were made using Student's t-tests or the Mann Whitney test (in vivo). Analyses of multiple groups were performed using a one-way or two-way analysis of variance with Bonferroni's posttest as appropriate. Statistical analyses were performed using GraphPad Prism 4.0 (GraphPad). For all statistical tests, a confidence interval of p < 0.05 was accepted for statistical significance.
We thank Drs. K. Touhara, S. Traynelis, V. Faundez and F. Rossi for reagents and technical help. GKP was supported by grants AR-047314, AR-051372, and AR-052730 from the National Institutes of Health and the Muscular Dystrophy Association. CAG was supported by National Institutes of Health training grant T32-GM08367. KAK was supported by an MDA Development grant.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.