Hypoxia-Induced WSB1 Promotes the Metastatic Potential of Osteosarcoma Cells Free

Osteosarcoma is the most common type of bone cancer that usually occurs in juvenile groups (1,2). Despite remarkable advances in the combined used of chemotherapy, radiotherapy and surgical ablation of the primary tumor, survival rates have yet to improve over those of the 1970s (3). Approximately 40% to 50% of patients develop metastasis, often pulmonary metastasis, even after curative resection of the primary tumor, and the 5-year survival rate of osteosarcoma patients with metastases is even lower than 30% (4-6). Accordingly, the leading cause of fatal outcomes in relapsed osteosarcoma is tumor metastasis. Targeting tumor metastasis may thus represent a promising strategy to intervene into human osteosarcoma. However, little progress has been made in this area as the molecular mechanisms underlying osteosarcoma invasion and metastasis remain poorly understood.

Meanwhile, intratumoral hypoxia contributes to increased risk of invasion, metastasis, treatment failure, and patient mortality in most types of solid tumors, including osteosarcoma (7-10). Hypoxia-inducible factor-1 alpha (HIF1α), a transcription factor that is stabilized in hypoxia, regulates the expression of multiple target genes to help tumor cells adapt to and survive in hypoxic conditions, contributing to poor chemotherapy responses, as well as decreased overall survival and disease-free survival in osteosarcoma patients (11-13). Recently, exposure of osteosarcoma cells to hypoxic conditions was reported to increase the invasive potential of the tumor cells in vitro (14), suggesting that hypoxia might promote the metastasis of osteosarcoma cells. Considerable effort has been directed at understanding the potential mechanism of hypoxia-promoted cancer metastasis (7,15), but most of these studies were focused on angiogenesis, chemokines, and stromal–tumor cell interactions. Previous work also demonstrated that HIF1α adapts tumor cells to hypoxic conditions via context and often cell-type specific mechanisms. It is intriguing to ask whether and how hypoxia could promote the metastasis of human osteosarcoma.

WD repeat and SOCS box containing 1 (WSB1) were recently identified as a new member of the suppressor of cytokine signaling (SOCS) box protein family, that is upregulated in multiple types of human cancers (16). WSB1, in complex with elongin B/C-Cullin5-Rbx1, forms an E3 ubiquitin ligase for thyroid-hormone–activating type 2-iodothyronine deiodinase (D2; ref. 17) and homeodomain-interaction protein kinase 2 (HIPK2; ref. 18), thus negatively regulating tumor progression and chemoresistance. More recent evidence suggested that WSB1 might be a direct target of HIF1 in a human hepatocellular carcinoma model, with a proposed function of WSB1 in hypoxia-promoted chemoresistance. Given that WSB1 is a substrate-recognizing subunit of E3 ubiquitin ligase, it is intriguing to ask whether and how WSB1 might function in hypoxia-driven metastasis in human osteosarcoma, possibly through mechanisms involving substrates other than D2 and HPIK2.

In this study, we first examined the potential role of WSB1 in regulating osteosarcoma cell metastasis in vitro and in vivo, as well as its association with hypoxia. Quantitative proteomic and functional analyses revealed that Rho guanosine diphosphate dissociation inhibitor 2 (RhoGDI2) as a novel downstream protein involved in WSB1-mediated cell migration.

The U2OS and MG63 cell lines were purchased from the Cell Bank of Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China). The KHOS/NP cell line was kindly provided by Dr. Lingtao Wu (University of Southern California, CA). All the cell lines have been tested and authenticated utilizing short tandem repeat (STR) profiling every 6 months. All primary osteosarcoma blasts were from fresh tissue sections from biopsies of osteosarcoma patients as described previously (19). All cells were cultured in DMEM or RPMI-1640 medium supplemented with 10% FBS in a humidified atmosphere of 5% CO₂ at 37°C.

The full-length coding sequence for WSB1 and RhoGDI2 was amplified from the U2OS cDNA library and subsequently subcloned into the PCMV6 plasmid (Origene). The WSB1-SOCS deletion was produced based on full-length WSB1 using a pair of primers (forward: 5′-GGAATTCATGGCCAGCTTTCCC-3′; reverse: 5′-GGAATTCTTAAACCTTATCGTCGTCATCCTT-3′). The indicated WSB1-SOCS mutation was produced by the Genescript Company.

The clinical samples of osteosarcoma patients were obtained from the Second Affiliated Hospital of Zhejiang University School of Medicine (Hangzhou, China) and Hangzhou First People's Hospital (Hangzhou, China). Written informed consents from patients and approval from the Institutional Research Ethics Committee of the hospital were obtained before the use of these clinical materials for research purposes.

The shRNA-expressing lentiviral vector pGFP-V-RS against the WSB1 gene was obtained from Origene. pCCL-WSB1 was constructed by the Genescript Company according to PCMV6-WSB1. The virus particles were harvested 48 hours after transfection of 293FT cells. For stable transfection, cells were grown in 6-well plates at 60% to 70% confluency, and 1 mL of viral supernatant was added with 1 μL polybrene.

Western blotting, real-time PCR, and immunoprecipitation were performed as described (20), with details in Supplementary information.

Chromatin immunoprecipitation (ChIP) was performed using the EZ ChIP kit (EMD Millipore). Briefly, KHOS/NP cells cultured under hypoxia conditions were collected and cross-linked with formaldehyde. Chromatin was sonicated on ice to an average length of 200 to 300 bp in an ultrasound bath. The chromatin was then incubated and precipitated with anti-IgG or anti-HIF1α antibodies. The immunoprecipitated DNA fragments were detected by PCR analysis using primers specific for PM1, PM2, and PM3.

Cells were cultured in DMEM and supplemented with either [U-12C6]-L-lysine (light) or [U-13C6]-L-lysine (heavy) for at least eight generations. The heavy labeling efficiency was measured by mass spectrometer analysis. Then, cells were continuously maintained in SILAC medium until they reached the desired confluence and were harvested by trypsinization, and combined in equal amounts. The enriched fractions were analyzed by mass spectrometry.

Tumor processing and immunofluorescence assays were performed as described previously (20), with details provided in the Supplementary information.

Cells were grown on 35-mm glass-bottom dishes and fixed with 4% paraformaldehyde for 15 minutes and then permeabilized with 0.1% Triton X-100 in PBS for 10 minutes. The cells were then blocked with 1% BSA for 20 minutes. Cells were incubated with Texas Red-X phalloidin (Life Technologies) diluted in 5 μL/200μL in PBS for 20 minutes and then stained with DAPI (100 μg/mL) for 3 minutes. The glasses were mounted with anti-fade reagent. The cells were observed under a confocal microscope.

Cells were seeded into 35-mm dishes coated with 0.5% gelatin. After 24 hours, dishes were placed onto an inverted microscope and imaged every 10 minutes up to 210 minutes. Temperature was maintained at 37°C.

The cell migration assay was performed in a 24-well Transwell plate with 8 μm polycarbonate sterile membrane (Corning Incorporated). Cells were plated in the upper chamber at 2×10⁴ cells per insert in 200 μL of serum-free medium. Inserts were placed in wells containing 600 μL of medium supplemented with 10% FBS. Twenty-four hours later, cells on the upper surface were detached with a cotton swab. Filters were fixed and cells in the lower filter were stained with 0.1% crystal violet for 15 minutes and counted. The quantified results were calculated by counting three random fields of migrated cells.

Tumors were established via by intravenous injection of lentivirus-transfected KHOS/NP cells (1×10⁶ cells/animal) into the tails of 3- to 4-week-old female BALB/c (nu/nu) mice (National Rodent Laboratory Animal Resource, Shanghai, China). At 14 days after injection, the mice were sent to the Second Affiliated Hospital of Zhejiang University School of Medicine and subjected to PET scanning. For the orthotopic transplantation model, 4-week-old BALB/c nude mice were anesthetized with a mixture of xylazine and ketamine. A 26-gauge needle was used to perforate the femur head, after which a second needle, coupled to a 1-mL syringe loaded with 1×10⁵ cells, was introduced through the puncture hole into the medullary cavity of the femur. After the mice were sacrificed, all lungs were dissected or further imaged by fluorescence stereomicroscopy and then fixed with formalin. Tissue sections (3 μm) were stained with hematoxylin/eosin. The Animal Research Committee at Zhejiang University approved all animal studies and animal care was provided in accordance with institutional guidelines.

The relationship between HIF1α and WSB1 expression was analyzed using the Spearman rank correlation test. Survival curves were plotted via the Kaplan–Meier method. The values for all samples in the different experimental conditions were averaged, and the standard error or SD of the mean was calculated. Differences between means were determined using the unpaired Student t tests and were considered significant at P < 0.05.

We first explored WSB1 expression in 40 primary osteosarcoma tissues by IHC staining. Interestingly, high expression levels of both the hypoxia marker protein HIF1α and the WSB1 protein were detected in most tumor specimens (Fig. 1A and B). Moreover, a statistically significant correlation was also found between these two proteins (R = 0.825, P < 0.01; Fig. 1B), suggesting a connection between high WSB1 levels and hypoxic microenvironments. To better understand the effect of hypoxia on WSB1 expression, three osteosarcoma cell lines and two primary osteosarcoma blasts (19) were cultured under hypoxic conditions (1% O₂) for 24 hours. The protein expression of WSB1, as well as HIF1α and its downstream target BNIP3, was significantly induced by hypoxia in all five osteosarcoma cells (Fig. 1C). Because WSB1 expression is significantly increased at the mRNA level under hypoxia (Fig. 1D), it is reasonable to question whether HIF1α mediates hypoxia-induced WSB1 transcription. By targeting HIF1α using two specific siRNAs, we found that HIF1α-depletion significantly downregulated its target, BNIP3, and also significantly blocked the hypoxia-induced WSB1 expression (Fig. 1E). In addition, both the mRNA and protein levels of WSB1 were upregulated, along with the expression of HIF1α and BNIP3, in cells that overexpressed HIF1α (Fig. 1D and F). Moreover, based on WSB1−promoter = luciferase reporter assays and site-directed mutagenesis, we confirmed that HIF1α was able to transactivate the WSB1 promoter via direct binding to its -339 bp region (Fig. 1G and H). ChIP assays also validated that HIF1α bound to the -339 bp genomic region of WSB1 under hypoxia (Fig. 1I). These results not only suggest that WSB1 is a direct target of HIF1α, but also imply that WSB1 might be involved in hypoxia-triggered tumorigenesis.

Given that hypoxia has been reported to play a crucial role in promoting tumor progression and metastasis (7,8), we further hypothesized that hypoxia-induced WSB1 expression might be pathologically relevant to the progression of osteosarcoma. Importantly, WSB1 protein levels were not only higher in tumors with metastasis than without it (Supplementary Fig. S1A), but were also prognostic for metastasis-free survival (Fig. 1J). Furthermore, hypoxia could promote the migration of primary osteosarcoma blasts (Fig. 1K and L), where WSB1 expression levels were significantly upregulated under hypoxic conditions (Fig. 1C). Thus, hypoxia-induced WSB1 is clearly associated with the metastatic potential of human osteosarcoma.

To further address the effect of WSB1 on osteosarcoma cell metastasis, we next examined the effects of manipulating WSB1 levels. Lentiviral transduction enabled the stable overexpression of WSB1 compared with vector-transduced cells, at levels comparable with those observed under hypoxia (Fig. 2A and Supplementary Fig. S1B). The migration assays showed that, compared with the vector-transduced cells, WSB1-overexpressing KHOS/NP and U2OS cells traversed more efficiently into the lower chamber of the Transwell under normoxia (Fig. 2B). Consistent with this finding, we observed that WSB1 overexpression enhanced the wound-closure capability of KHOS/NP cells (Supplementary Fig. S2). Therefore, WSB1 may modulate the migration capacity of osteosarcoma cells in vitro.

To further quantify metastatic potential in vivo, we performed tail-vein xenografts in BALB/c (nu/nu) mice and examined the rates of lung colonization by micro-PET scan. As shown in Fig. 2C and D, increased signal intensities in lungs were observed on day 14 in mice transplanted with WSB1-overexpressing KHOS/NP cells compared with control mice, indicating the formation of more metastatic nodes in mice injected with WSB1-overexpressing cells. The lentiviral vector pCCL expresses a GFP tag (Supplementary Fig. S1C), allowing us to visualize metastatic nodes in lungs via fluorescence imaging. As shown in Fig. 2E, WSB1-overexpressing cells formed more metastatic nodules in the lung than control cells. Histologic examination confirmed that the number of micrometastatic lesions was markedly increased in the lungs of mice transplanted with WSB1-overexpressing cells (Fig. 2F and Supplementary Fig. S3). Moreover, we further examined the metastasis phenotype of those two cell lines using an orthotopic transplantation model. Consistent with our intravenous injection model, WSB1-overexpressing cells induced more lung metastasis nodes than control cells (Fig. 2G and H).

We next verified the effects of WSB1 depletion on the migration and metastasis of KHOS/NP cells using shRNA. Two specific sequences against WSB1 significantly inhibited the hypoxia-dependent expression of WSB1 (Fig. 2I). Meanwhile, in contrast with overexpression (Fig. 2B), WSB1 depletion markedly inhibited the hypoxia-driven migration of KHOS/NP cells (Fig. 2J). In agreement with the in vitro results, WSB1 depletion significantly decreased the signal intensity in the lungs of tail-vein-injected nude mice in micro-PET scans (Fig. 2K and L). Importantly, histologic examination of lung tissues revealed that in some cases depleting WSB1 totally blocked the metastatic nodules in vivo (1/4 in shRNA-#1 group and 2/4 in shRNA-#2 group), whereas mice injected with control cells formed 8 to 18 metastatic nodules per lung in all four mice (Fig. 2M and N). Our data indicate that WSB1 is necessary for the aggressive, highly metastatic phenotype of osteosarcoma cells.

Collectively, these overexpression and knockdown results generated from in vitro and in vivo studies, together with the results from osteosarcoma biopsies and primary cell lines, demonstrate a new phenotype of WSB1 wherein the hypoxia-dependent induction of WSB1 is necessary and sufficient to promote the metastatic potential of osteosarcoma cells.

Previous reports have revealed that WSB1 is an E3 ubiquitin ligase (17, 18); therefore, it is interesting to know whether E3 ubiquitin ligase activity is critical for WSB1 to promote cell migration. WSB1 is composed of seven WD40 repeats and a C-terminal SOCS box (Fig. 3A), and the SOCS box is known to be involved in protein degradation via the ECS ubiquitin ligase complex (21). We therefore constructed a mutant WSB1 that lacks the SOCS box (Fig. 3B, ΔSOCS) to determine whether E3 ubiquitin ligase activity is required for WSB1-promoted cell migration. As illustrated in Fig. 3C and D, infection with wild-type WSB1 lentivirus significantly promoted migration compared with the vector control. In contrast, ΔSOCS completely lost its migration-promoting activity, as did a WSB1 point-mutant (339-421, AAA) lacking E3 ligase activity (Fig. E–H). Taken together, these data further suggest that E3 ubiquitin ligase activity is required for the WSB1-driven migration of osteosarcoma cells.

To further understand the downstream pathways induced by WSB1 in the metastatic process of osteosarcoma, a large-scale proteomic analysis was subsequently performed. The proteomes of ectopic WSB1-overexpressing and control vector-infected KHOS/NP cells were quantified using SILAC followed by mass spectrometry analysis (Fig. 4A). A total of 4,163 proteins were identified from both cell lysates, and 4,094 proteins were further quantified, of which 1,078 proteins were significantly perturbed by WSB1 overexpression, with 560 proteins being upregulated and 518 proteins downregulated (>3.0-fold, Fig. 4B and Supplementary Table S1). Bioinformatics analysis further revealed that these altered proteins are mainly involved in protein binding (50%) and catalytic activity (29%) and also participated in both cellular (23%) and metabolic processes (18%; Fig. 4C). The expression of metastasis-associated proteins was further analyzed. Twenty-four proteins involved in cell metastasis were identified, and 7 of these 24 proteins changed by more than 3.0-fold in response to WSB1 overexpression (Supplementary Fig. S4). These quantitative proteomic data provide additional confirmation that WSB1 promotes osteosarcoma cell metastasis.

Interestingly, RhoGDI2, a critical and potent inhibitor of Rho proteins (22), was the top candidate from the SILAC analysis, with the most significant decrease in protein abundance (Fig. 4D). Because our data suggest that E3 ubiquitin ligase activity is required for the WSB1-driven migration of osteosarcoma cells, we hypothesized that the downregulation of RhoGDI2 may serve as the downstream target of WSB1. To confirm this, the RhoGDI2 protein expression level in WSB1-overexpressing KHOS/NP cells was determined. Forced WSB1 overexpression significantly inhibited RhoGDI2 protein expression, whereas RhoGDI1, another member of the RhoGDI family, did not change (Fig. 4E), indicating the selective inhibition of RhoGDI2 by WSB1. Because WSB1 is specifically induced by hypoxia, we examined whether the downregulation of RhoGDI2 could also be achieved under hypoxic conditions. As expected, an obvious decrease in RhoGDI2 was observed to accompany the induction of WSB1 under hypoxia (Fig. 4F). In contrast, WSB1 knockdown effectively reversed the hypoxia-dependent decrease in RhoGDI2 levels (Fig. 4G). Moreover, a negative correlation between WSB1 and RhoGDI2 was also found in 40 primary osteosarcoma tissues by IHC staining (R = −0.65, P < 0.01; Fig. 4H and I). In addition, RhoGDI2 protein levels were inversely correlated with metastasis of osteosarcoma (Supplementary Fig. S5A). These data clearly support the idea that RhoGDI2 is the downstream target of WSB1 signaling.

Given that our data further demonstrated that RhoGDI2 mRNA levels were not significantly decreased in WSB1-overexpressing cells and under hypoxia (Supplementary Fig. S5B), as well as the fact that ΔSOCS failed to decrease the amount of RhoGDI2 (Fig. 5A), we are encouraged to further address whether WSB1 mediates the degradation of RhoGDI2. To this end, cycloheximide (CHX) was used to prevent de novo protein synthesis; thus, RhoGDI2 levels would primarily reflect the protein degradation process. We exposed WSB1-overexpressing cells and control cells to CHX and estimated the RhoGDI2 expression. As shown in Fig. 5B, RhoGDI2 degradation rates were remarkably increased in WSB1-overexpressing cells. Furthermore, MG132 (a specific proteasome inhibitor) treatment totally reverses the WSB1 overexpression-triggered decrease in RhoGDI2 protein levels (Fig. 5C). In keeping with these results, MG132 also inhibited RhoGDI2 protein degradation under hypoxic condition (Supplementary Fig. S5C).

We next examined whether WSB1 regulates RhoGDI2 protein degradation through a direct interaction. Immunofluorescence results demonstrated that exogenous WSB1 colocalizes with exogenous RhoGDI2 in 293T cells transfected with FLAG-WSB1 and RhoGDI2-HA plasmids (Fig. 5D). Furthermore, this interaction was further validated by immunoprecipitation. Ectopically expressed WSB1 and RhoGDI2 were coprecipitated from 293T cells overexpressing both proteins (Fig. 5E). Consistently, endogenous WSB1 also interacts with endogenous RhoGDI2 as confirmed by immunoprecipitation (Fig. 5F) and immunofluorescence assays (Supplementary Fig. S5D). Therefore, our data clearly demonstrate that WSB1 directly interacts with RhoGDI2. To confirm the role of WSB1 as an E3 ligase for RhoGDI2 polyubiquitination, we performed ubiquitination assays. Polyubiquitinated RhoGDI2 only accumulated in cells that both overexpressed WSB1 and were treated with MG132 (Fig. 5G). Collectively, our data suggest that hypoxia-driven WSB1 promotes the proteasome-dependent degradation of RhoGDI2.

RhoGDIs modulate the cycling of Rho GTPases between active GTP-bound and inactive GDP-bound states (23). Our data indicate that ectopic WSB1 expression is associated with a marked decrease in RhoGDI2; thus, the activation of Rho GTPases should be expected in WSB1-overexpressing cells. By specifically pulling down the active GTP-bound form of Rho GTPases, we detected a substantial level of active Rac1 protein, a member of the Rho GTPases, in the wild-type WSB1-overexpression group but not in the vector or ΔSOCS groups (Fig. 6A). Our findings indicate that WSB1-triggereed RhoGDI2 suppression activates Rho GTPases.

Rho GTPases are well known as regulator of actin cytoskeletal organization and cell motility. We therefore examined the effect of WSB1 overexpression on the status of actin filament organization and cell motility. As illustrated in Fig. 6B and C, WSB1 overexpression significantly enhanced the fluorescence intensity of polymerized actin (F-actin), which suggests abundant F-actin expression, compared with the vector or mutant WSB1 groups. In addition to F-actin expression, the appearance of membrane ruffles and the formation of lamellipodia were also observed in WSB1-overexpressing cells. Dynamic monitoring of the random motility of KHOS/NP cells revealed an increased displacement of WSB1-overexpressing cells from their origin sites compared with control cells (Fig. 6D and Supplementary Figs S6 and S7). These data clearly demonstrate that WSB1 increases the amount of F-actin, promotes the formation of lamellipodia, and consequently leads to enhanced cell motility and migration ability.

To determine whether RhoGDI2 overexpression could reverse WSB1-driven metastasis of osteosarcoma cells, we first established a stable cell line ectopically expressing RhoGDI2 (Fig. 7A). Notably, RhoGDI2 overexpression almost completely abrogated WSB1-promoted cell migration (Fig. 7B and C). Similar responses were also observed via immunofluorescence staining of F-actin, as RhoGDI2 overexpression decreased the WSB1-enhanced fluorescence intensity of F-actin and the formation of membrane ruffles (Fig. 7D and E). Moreover, RhoGDI2 overexpression blocked WSB1-enhanced wound-closure capability of KHOS/NP cells (Supplementary Fig. S8). These data suggest that WSB1-enhanced cell motility and migration can be completely abrogated by RhoGDI2 overexpression in vitro.

To test whether this effect could be reproduced in vivo, we generated an osteosarcoma pulmonary metastasis model using nude mice intravenously injected with the indicated cells. Consistent with our in vitro data, WSB1 overexpression markedly increased the number of pulmonary metastatic nodules (GFP positive). However, the additional ectopic expression of RhoGDI2 impaired the effect of WSB1 overexpression as demonstrated by the decrease in pulmonary metastatic nodules (Fig. 7F and G). Taken together, these results not only indicate that RhoGDI2-overexpression reverses WSB1-driven metastasis, but also further support a pivotal role for the WSB1–RhoGDI2 axis in the metastasis of osteosarcoma cells.

Pulmonary metastasis has been recognized as the main cause of fatal outcomes in osteosarcoma patients, but its molecular mechanism is rarely discussed (1,24). Accumulating evidence suggests that tumor progression requires adaptation to hypoxic microenvironments, and molecules and proteins induced by hypoxia have thus attracted extensive attention. As a major regulator of the cellular response to hypoxia, HIF1α has been previously detected in osteosarcoma specimens (11,12,14,25). In line with previous work (26), our data also confirm that hypoxia induces WSB1 expression in a HIF1α-dependent manner, which further supports the notion that high expression levels of the HIF1α–WSB1 axis are associated with poor prognosis in osteosarcoma patients. There are increasing reports connecting hypoxia-promoted cancer metastasis with poor prognosis. Nevertheless, most of these studies focus on angiogenesis, chemokine, and stromal–tumor cell interactions, with little evidence of a direct effect of hypoxia on cell migration (7,15). More recently, HIF1α has been shown to activate the transcription of the RhoA and Rho kinase 1 (ROCK1) genes, indicating the activation of cell motility (27). Our findings suggest that hypoxia-driven WSB1 can promote the proteasomal degradation of RhoGDI2, leading to cytoskeletal changes that underlie the invasive cancer cell phenotype. In this regard, our results elucidate a novel branch of HIF1α signaling that directly mediates cell motility.

Despite evidence that WSB1 is expressed in some types of cancer, its role in cancer progression is still controversial. A gene expression analysis of 37 neuroblastoma patients revealed that increased WSB1 copy number correlates with good prognosis (28). In contrast, reduced cell proliferation and enhanced resistance to apoptosis are observed in both pancreatic cancer cells and neuroblastoma cells after WSB1 overexpression (16,29). More recently, data from hepatocellular carcinoma cells further suggest that WSB1 plays a critical role in hypoxia-induced chemoresistance (26). Our study shows that WSB1 controls RhoGDI2 to enhance the activity of Rho proteins and promotes osteosarcoma cell migration. In addition to osteosarcoma cells, we also observed the enhancement of WSB1-mediated migration in non–small cell lung cancer cell line A549 and cervical cancer cell line HeLa (data not shown), implying that the function of WSB1 in promoting cell migration may not be cell line specific. To our knowledge, our study is the first to demonstrate that WSB1 promotes the motility and metastatic potential of cancer cells.

Here, we show that WSB1 directly binds to RhoGDI2 and promotes the degradation of RhoGDI2. However, their binding domain is unknown. Previous studies reported that RhoGDI2 is composed of two domains: a flexible N-terminal domain of approximately 70 residues and a folded 132-residue C-terminal domain. The N-terminal domain has two incipient helical structures (residues 36-57 and 20-25), which adopts a helix-turn-helix structure in the complex and plays an essential role for the binding (30). Therefore, further studies are needed to confirm whether RhoGDI2′s N-terminal domain is also critical for WSB1 binding.

Metastasis is a complex process that leads to the dissemination of cancer cells from the primary tumor to distant organs. A crucial step is cytoskeletal reprogramming, which transforms rigid, immobile epithelial cells to motile, invasive cancer cells (31,32). The small Rho GTPase family members Rho, Rac, and Cdc42 play critical roles in cell migration by regulating the dynamic assembly of actin filaments (33). RhoGDIs inhibit Rho GTPases by direct interaction and sequester Rho proteins in the cytoplasm to restrain them from activation at the membrane (23, 34). In our study, we found that the degradation of RhoGDI2 in WSB1-overexpressing cells led to Rac1 activation, and subsequently increased both the expression of polymerized actin and the formation of membrane ruffles, consistent with previous reports that the preferential binding of RhoGDI2 to Rac1 is related to a distinctive regulation of Rho proteins by RhoGDI2 (35, 36).

To date, three members (RhoGDI1, 2, and 3) of the RhoGDI family have been identified (23). Of these proteins, RhoGDI1 is well characterized as a ubiquitously expressed member of this family, whereas the expression of RhoGDI2 and RhoGDI3 is tissue specific (23,34). Although RhoGDI2 is highly expressed particularly in lymphocytes, the widespread tissue distribution of its mRNA has been recently reported (37). The loss of RhoGDI2 is associated with multiple types of metastatic cancers (22,38-40), which is also consistent with our finding that the WSB1-triggered depletion of RhoGDI2 promotes the metastatic potential of osteosarcoma cells. Of interest, our data show that WSB1 selectively inhibits RhoGDI2 but not RhoGDI1. Although it remains to be determined how WSB1 carries out this distinctive regulation, a similar phenomenon has also been observed in rictor-triggered RhoGDI2 expression, as revealed by the increase in RhoGDI2 rather than RhoGDI1 in rictor-null MEFs (41,42). Because only the loss of RhoGDI2, but not RhoGDI1, has been reported to be associated with metastasis, and because the functional distinction between RhoGDI1 and RhoGDI2 remains elusive, these clues further suggest the existence of highly organized intracellular network that coordinates individual signaling to mediate metastasis.

Colonization by intravenously injected tumor cells is widely used as a model for detecting metastasis. This model primarily leads to the formation of metastatic nodules in the lungs. Given that the main metastasis site of osteosarcoma is lung, intravenous injection model is a reasonable assay to explore the role of WSB1 in osteosarcoma metastasis. In our case, WSB1 overexpression markedly increased the pulmonary metastatic nodules, whereas RhoGDI2 blocked WSB1-enhanced metastatic ability. However, the intravenous injection model usually depicts only the late phases of the invasion–metastasis cascade, such as cancer cell dissemination and organ colonization, and fails to represent the earlier stages of the metastatic process, such as local invasion and intravasation. Therefore, we also examined spontaneous metastasis using an orthotopic transplantation model, which involves a more comprehensive process. Our results further confirmed that WSB1-overexpressing osteosarcoma cells lead to more lung metastatic nodes than control cells.

In summary, our current study shows that hypoxia induces WSB1 expression, which leads to Rho pathway activation via RhoGDI2-depletion and activation of the GTP-bound form of Rho proteins and actin polymerization, and consequently promotes osteosarcoma cell motility, migration, and metastasis. These results not only reveal a novel physiologic function for WSB1, but also delineate, for the first time, a critical molecular mechanism wherein hypoxia induces a WSB1–RhoGDI2 signaling axis to trigger cancer cell motility. The discovery of WSB1 as a novel hypoxia-driven osteosarcoma metastasis-supporting protein provides new opportunities to prevent and treat osteosarcoma with metastases.

No potential conflicts of interest were disclosed.

Conception and design: J. Cao, Y. Wang, M. Ying, Q. He, B. Yang

Development of methodology: Y. Wang, R. Dong, G. Lin

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Cao, Y. Wang, R. Dong, G. Lin, N. Zhang, J. Wang, N. Lin, L. Ding

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Cao, Y. Wang, R. Dong, G. Lin, N. Zhang, J. Wang, L. Ding, M. Ying, Q. He

Writing, review, and/or revision of the manuscript: J. Cao, N. Zhang, Y. Gu, M. Ying, B. Yang

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): N. Lin, Y. Gu, Q. He

Study supervision: N. Lin, Q. He, B. Yang

This work was supported by grants from the National Natural Science Foundation of China (No. 81373440 to L. Ding; No. 81273535 to B. Yang), the China Postdoctoral Science Foundation (No. 2014M550330 to J. Cao), and the Fundamental Research Funds for the Central Universities (J. Cao, Q. He, and No. 2014XZZX004 to B. Yang).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.