DKK2 Mediates Osteolysis, Invasiveness, and Metastatic Spread in Ewing Sarcoma

Ewing sarcomas are undifferentiated bone or soft tissue tumors of enigmatic histogenesis mostly occurring in children and adolescents. Genetically, Ewing sarcoma are defined by EWS/ETS translocations (1, 2). These highly malignant tumors frequently arise in bones possibly descending from a mesenchymal stem cell in transition from an undifferentiated state to a more differentiated phenotype of the endothelial, neuroectodermal or, as hypothesized here, of the chondro-osseous lineage (2–4).

Although prognosis of patients with localized disease has markedly improved in past decades, metastatic disease—present in about 25% of patients with Ewing sarcoma at diagnosis—is usually associated with fatal outcome (5, 6). Especially the development of metastases in bones is a catastrophic event in the clinical course of patients with Ewing sarcoma (5, 7). Hence, there is an urgent need to understand the fundamental molecular mechanisms of Ewing sarcoma differentiation, invasion, and osteolytic tumor growth to possibly identify novel therapeutic strategies to prevent metastasis.

Previously, we described an Ewing sarcoma–specific expression signature comprising 37 genes that are highly upregulated or specifically expressed in Ewing sarcoma (8). One of them is dickkopf 2 (Dickkopf, Xenopus, Homolog 2) a member of the dickkopf family. Recently, DKK2 was reported as a key player in stem cell signaling networks due to its function as a Notch signaling target (9). However, its precise cellular function remains elusive. For instance, secreted DKK2 can function as either a Wnt agonist or an antagonist, depending on the cellular context and the expressed amount of its binding partner low-density lipoprotein receptor–related protein 6 (LRP6) and its cofactor Kremen 2 (10–13). In the context of the Wnt/β-catenin pathway, DKK2 may promote osteoclastogenesis by enhancing the expression of osteoclast differentiation factors, thus leading to tumor-associated osteolysis (14, 15). Furthermore, Li and colleagues showed that DKK2 has a role in terminal osteoblast differentiation into mineralized matrices through both canonical Wnt antagonism–dependent and -independent mechanisms (16). On the basis of these observations, we hypothesized that DKK2 is involved in Ewing sarcoma invasion, osteolysis, and chondrogenic/osteogenic differentiation potential.

In the present study, we analyzed the putative oncogenic function of DKK2 in Ewing sarcoma in vitro and in vivo. We show that DKK2 increases Ewing sarcoma proliferation, anchorage-independent colony formation, osteolysis, and invasiveness, the latter being likely mediated by MMP1. In addition, we show that DKK2 expression prevents neuronal but in parallel enhances chondrogenic and osteogenic differentiation capacity of Ewing sarcoma. Furthermore, DKK2 regulates the expression of different genes important for homing, invasion into bone and osteolysis, suggesting that DKK2 overexpression is a critical factor in Ewing sarcoma malignancy, in particular bone metastasis.

Ewing sarcoma lines (MHH-ES1, RD-ES, SK-ES1, SK-N-MC, and TC-71) and neuroblastoma lines (CHP126, MHH-NB11, SHSY5Y, and SIMA) were obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ). Ewing sarcoma line VH64 was kindly provided by Marc Hotfilder (Münster University, Münster, Germany); osteosarcoma cell lines (HOS, HOS-58, MG-63, MNNG, SaOS, SJSA01, U2OS, and ZK-58) by Jan Smida and Michaela Nathrath, Institute of Pathology and Radiation Biology (Neuherberg, Germany). A673 was purchased from ATCC (LGC Standards). SB-KMS-KS1 cells and SB-KMS-MJ1 are Ewing sarcoma cell lines that were established in our laboratory (17, 18). Retrovirus packaging cell line PT67 was obtained from Takara Bio Europe/Clontech. Cells were maintained in a humidified incubator at 37°C in 5% to 8% CO₂ atmosphere in RPMI-1640 or Dulbeccos' Modified Eagle's Media (DMEM; both Invitrogen) containing 10% heat-inactivated FBS (Biochrom) and 100 μg/mL gentamicin (Invitrogen). Cell lines were checked routinely for purity (e.g., EWS-FLI1 translocation product, surface antigen, or HLA-phenotype) and mycoplasma contamination.

For transient RNA interference (RNAi), cells were transfected with siRNAs as described previously (17). siRNA sequences are provided in the Supplementary Information.

For stable silencing of DKK2 expression, oligonucleotides were designed corresponding to the most efficient siRNA used for transient RNAi (see Supplementary Information), and retroviral gene transfer was conducted as described previously (17).

Differential gene expression was analyzed by quantitative RT-PCR (qRT-PCR) using TaqMan Universal PCR Master Mix and fluorescence detection with an AB 7300 Real-Time PCR System (both Applied Biosystems) as described previously (17). A list of the assays used is provided in the Supplementary Information.

For DNA content, cells were fixed in ice-cold 70% ethanol at −20°C overnight and stained with propidium iodide (50 mg/mL) plus RNase (5 mg/mL) in PBS. The fluorescence of cells was measured using a FACSCalibur flow cytometer (BD Biosciences) and analyzed using CellQuest software (BD Biosciences).

Cell proliferation was measured with an impedance-based instrument system (xCELLigence, Roche/ACEA Biosciences) enabling label-free real time cell analysis. Briefly, 1 × 10⁴ to 3 × 10⁴ cells were seeded into 96-well plates with 200 μL media containing 10% FBS and allowed to grow up to 72 hours. Cellular impedance was measured periodically every 4 hours, and gene knockdown was monitored by qRT-PCR.

Cells were seeded in duplicate into a 35-mm plate at a density of 5 × 10³ cells per 1.5-mL methylcellulose-based media (R&D Systems) according to the manufacturer's instructions and cultured for 10 to 14 days at 37°C/5% CO₂ in a humidified atmosphere.

To study cell invasion, the BioCoat Angiogenesis System: Endothelial Cell invasion was used (BD Biosciences) according to the manufacturer's instructions as described previously (18).

Procedures were described previously (17). Following antibodies were used: anti-DKK2 (1:100; PAB3570, Abnova), anti-GAP43 (1:500; ab7462; Abcam), and horseradish peroxidase (HRP)-coupled bovine anti-rabbit IgG (1:500; sc-2370; Santa Cruz Biotechnology). Equal protein loading was controlled with rabbit polyclonal to HPRT antibodies (1:500; sc20975; Santa Cruz Biotechnology).

For neuron/astrocyte-like cell differentiation, cells were cultured as already described (17), and the differentiated cells were identified by staining with an antibody directed against GFAP (BD Biosciences) and visualized with a fluorescein isothiocyanate (FITC)-labeled goat anti-mouse F(ab')2 fragment (Jackson ImmunoResearch Laboratories) or with an antibody against the neuronal marker GAP43 (Abcam) using Western blot analysis.

Comparative gene expression analysis was conducted as previously described (17, 18) using human Gene ST arrays (Affymetrix, GSE36100). Gene set enrichment analysis (GSEA) and set-to-set pathway analyses of leading-edge genes were conducted with the GSEA tool (http://www.broad.mit.edu/gsea) in default parameters (19).

Immunodeficient Rag2^−/−γc^−/− mice on a BALB/c background were obtained from the Central Institute for Experimental Animals (Kawasaki, Japan) and maintained in our animal facility under pathogen-free conditions in accordance with the Institutional Guidelines and approval by local authorities. Immunodeficient NOD/SCID Balb/c mice were purchased by Charles River Laboratories International. Experiments were carried out in 6- to 16-week-old mice.

For the analysis of in vivo tumor growth and metastatic potential, 1.5 × 10⁶ to 3 × 10⁶ Ewing sarcoma cells and derivates were injected in a volume of 0.2 mL into immunodeficient Rag2^−/−γc^−/− or NOD/SCID Balb/c mice as described previously (17, 18). To examine bone invasiveness and osteolysis, mice were anesthetized with 500 mg/mL Novaminsulfon (Ratiopharm) and isoflurane (Abbott). A 30-gauge needle was introduced through the proximal tibia plateau, and 2 × 10⁵ Ewing sarcoma cells in a volume of 20 μL were injected into the medullary cavity. In all experiments, tumors and affected tissues were recovered and processed for histologic analyses. Intratibial tumor formation was monitored by X-ray radiography.

Murine organs were fixed in phosphate-buffered 4% formaldehyde and embedded in paraffin. Three- to 5-μm thick sections were stained with hematoxylin and eosin (H&E). The amount of osteoclasts was detected by tartrate-resistant acid phosphatase (TRAP) staining. All sections were reviewed and interpreted by 2 pathologists (J. Calzada-Wack and I. Esposito).

Data are mean ± SEM as indicated. Differences were analyzed by unpaired 2-tailed Student's t test as indicated using Excel (Microsoft) or Prism 5 (GraphPad Software); P < 0.05 was considered statistically significant (*, P < 0.05; **, P < 0.005; ***, P < 0.0005).

In a previous microarray analysis, we identified DKK2 as 1 of 37 genes, which were strongly upregulated in Ewing sarcoma compared with normal tissues (Fig. 1A; ref. 8). To analyze specificity of DKK2 expression in Ewing sarcoma, we tested Ewing sarcoma cell lines against a series of different osteosarcoma and neuroblastoma cell lines. qRT-PCR revealed a high expression of DKK2 only in bone-associated tumors such as Ewing sarcoma and osteosarcomas but not in neuroblastomas and other small round blue cell tumors (Fig. 1B and Supplementary Fig. S1).

Although DKK2 seems upregulated in several pediatric bone tumors, we were curious whether DKK2 expression in Ewing sarcoma is dependent on the expression of the oncofusion protein EWS-FLI1. Transient knockdown of EWS-FLI1 by specific siRNA did not significantly affect DKK2 expression in 4 Ewing sarcoma cell lines, indicating DKK2 expression in Ewing sarcoma to be independent of EWS-FLI1 (Fig. 1C). Furthermore, constitutive suppression of DKK2 did not influence EWS-FLI1 mRNA levels (Fig. 1D).

To analyze the impact of DKK2 overexpression on the phenotype of Ewing sarcoma, we transiently and constitutively downregulated DKK2 in 3 different Ewing sarcoma cell lines (Fig. 2A and B). Using an xCELLigence instrument, we first analyzed the effect of DKK2 knockdown on the proliferation of Ewing sarcoma cell lines in vitro. Constitutive downregulation of DKK2 revealed a significant inhibition of proliferation in A673 and SK-N-MC cells (Fig. 2C), without significantly affecting cell-cycle progression (Fig. 2D). We next tested anchorage-independent growth of stable DKK2 infectants in methylcellulose. As shown in Fig. 2E, constitutive DKK2 knockdown clearly reduced colony formation in all 3 Ewing sarcoma cell lines in a dose-dependent manner (Fig. 2B).

We further examined whether suppression of DKK2 in Ewing sarcoma affects tumorigenicity and metastasis in vivo. We injected stable pSIREN^DKK2-infected A673 and SK-N-MC cells and respective controls subcutaneously into the inguinal region of immunodeficient mice and analyzed tumor growth. As shown in Fig. 3A, DKK2 knockdown clearly delayed local tumor growth in both Ewing sarcoma cell lines in a dose-dependent manner. Similarly, suppression of DKK2 significantly reduced the number of lung metastases, when cells were injected into the tail veins of Rag2^−/−γc^−/− mice (Fig. 3B). Furthermore, considering in addition liver and kidney metastases, pSIREN^DKK2 infectants revealed a strong reduction of overall metastatic potential by up to 95.8% (Table 1), suggesting a critical role of DKK2 in Ewing sarcoma growth and metastasis.

To identify possible DKK2 downstream targets that are presumably involved in metastasis, we transiently downregulated DKK2 in A673 and SK-N-MC cells (Fig. 2A) and compared their expression pattern with pSIREN^negsiRNA cells in a microarray analysis on human Gene ST arrays (Affymetrix, GSE36100). Considering a minimum linear fold change more than ±1.5, we identified 281 differentially regulated genes of which 207 genes were downregulated after DKK2 knockdown (Fig. 4A). Differential expression of 5 of these genes, which all have been reported to be involved in cellular invasiveness and migration, was confirmed by qRT-PCR (Fig. 4B). The only upregulated gene of these 5 genes is the key cell–cell adhesion molecule E-Cadherin (CDH1), which is a well-established antagonist of invasion and metastasis (20). The 4 downregulated genes are the integral cell membrane glycoprotein CD44, which has a postulated role in matrix adhesion and migration (21, 22), the cellular adhesion molecule 1 (ICAM1), which is important for endothelial transmigration into tissues (23), and the matrix metalloproteinases MMP1 and MMP7, both with crucial roles in migration and invasion (24–26). In this context, stably DKK2-silenced Ewing sarcoma cells showed significantly reduced invasiveness when subjected to Matrigel-covered Transwell assays (BioCoat Angiogenesis System; Fig. 4C). As previously reported, MMPs seems to be important for Ewing sarcoma invasiveness (18). Therefore, we examined the invasiveness of A673 and SK-N-MC cells after treatment with 3 or 6 nmol/L Batimastat (BB-94)—a specific MMP inhibitor. As expected, treatment with BB-94 prominently reduced the amount of cells crossing the Matrigel (Fig. 4D). We next transiently knocked down MMP1 or MMP7, each with 2 different specific siRNAs. As shown in Fig. 4E and F, only knockdown of MMP1, but not of MMP7, reduced invasiveness of Ewing sarcoma cells, suggesting that the reduced invasive potential of DKK2-silenced Ewing sarcoma cells is mediated—at least in part—via MMP1.

Subsequent GSEA of our microarray data identified 75 differentially regulated gene sets [normalized enrichment score (NES) < −1.54 or NES > 1.30], that belong to 1 of the 3 ontologies: molecular function, cellular component, or biologic processes (C5: GO gene sets, v3.0; ref. 27). Set-to-set analysis of leading-edge top ranked downregulated gene sets (C5_all, v3.0) revealed a strong overrepresentation of gene sets involved in anti-apoptotic pathways (Supplementary Fig. S2A–S2C) that possibly explains the slightly higher rate of dead cells after DKK2 knockdown in vitro and in vivo (Fig. 2D, sub-G₁ fraction; data not shown). Consistently, DKK2 knockdown increased the rates of apoptosis as measured by Annexin-V and 7-aminoactinomycin D (7-AAD) staining (Supplementary Fig. S2D). Moreover, the set-to-set analysis of the upregulated gene sets (C5_all, v3.0) revealed a strong overrepresentation of gene sets involved in neuronal differentiation and development (Fig. 5A and B). To verify this prediction, we first examined the overexpression of 3 leading-edge genes, which are all involved in neurogenesis, namely, glial fibrillary acidic protein (GFAP), nerve growth factor receptor (NGFR), and slit, drosophila, homolog of, 2 (SLIT2), using qRT-PCR (Fig. 5C). Subsequently, we induced neurogenic differentiation with 0.1 mmol/L BHA in 2% dimethyl sulfoxide (DMSO) in stable A673, SK-N-MC, and TC-71 DKK2 short hairpin RNA (shRNA) infectants. We observed that Ewing sarcoma cell lines were able to fully differentiate and express GFAP, a major intermediate filament protein of mature astrocytes (28), only after DKK2 suppression (Fig. 5D). Similarly growth-associated protein 43 (GAP43), a protein exclusively expressed on the synaptosomal membrane of nerve tissues (29), was significantly higher expressed in Ewing sarcoma cells after DKK2 knockdown, as shown in Fig. 5E, indicating that DKK2 inhibits the neuronal differentiation potential of Ewing sarcoma cells.

We next investigated whether DKK2 can influence their differentiation potential toward other tissues or lineages. Min and colleagues already showed that DKK2 promotes angiogenesis (30), therefore, we examined the endothelial differentiation capacity of Ewing sarcoma cells, but we did not detect any differences between cells with/without DKK2 knockdown in tube formation assays (Supplementary Fig. S3A and S3B). Moreover, given that DKK2 is also implicated in terminal osteoblast differentiation (14, 16) and osteoclastogenesis (14, 15), we subsequently analyzed the ability of Ewing sarcoma cells to differentiate along bone lineages. As shown in Supplementary Fig. S4, induction of chondrogenic and osteogenic differentiation with specific growth media was clearly impaired in DKK2 knockdown cells. Furthermore, we examined the differential expression of key players associated with bone colonization in stable DKK2 infectants. We chose chemokine, CXC motif, receptor 4 (CXCR4), parathyroid hormone-like hormone (PTHrP), RUNT-related transcription factor 2 (RUNX2), and transforming growth factor beta-1 (TGFβ1) for follow-up due to their involvement in preparing the premetastatic niche, homing, and invasion to bone, as well as their role in local and metastatic tumor growth in bones (26). We could show that the expression of these genes was significantly reduced in Ewing sarcoma cells with constitutive DKK2 knockdown compared with controls (Fig. 6A). Interestingly, 3 of the 5 genes are also involved in the osteolytic behavior of tumor cells (26), which seems important as Ewing sarcoma are well-known osteolytic tumors. To verify this observation, we further analyzed the expression of other genes associated with osteolysis, namely hypoxia-inducible factor 1, alpha subunit (HIF1α), jagged 1 (JAG1), interleukin 6 (IL6), and vascular endothelial growth factor receptor 1 (VEGF; ref. 26). As expected, knockdown of DKK2 significantly reduced the expression of these genes (Fig. 6B)

On the basis of these results, we analyzed whether suppression of DKK2 has a role in the process of bone invasion and osteolysis in vivo. We injected constitutively DKK2 silenced A673 cells into the tibiae of immunodeficient Rag2^−/−γc^−/− mice and analyzed bone infiltration and destruction by X-ray radiography and histology. As shown in Fig. 6C–F, DKK2 knockdown significantly reduced bone infiltration and osteolysis. Only 25% of the mice injected with A673 pSIREN^DKK2 cells exhibited an infiltration of tumor cells into bone, whereas bone invasion was observed in 87.5% of the respective controls (Fig. 6D). The pSIREN^DKK2-mediated lesions were smaller and less destructive than the lesions caused by control cells, as measured by the amount of TRAP-positive osteoclasts in the tumor (Fig. 6E and F). In addition, pilot screening of human Ewing sarcoma samples suggested the level of DKK2 expression to be associated with invasive growth and survival (Supplementary Fig. S5), although a careful retrospective study here is necessary to obtain significant results.

In summary, these data show that DKK2 critically influences the differentiation process of Ewing sarcoma cells along the chondrogenic and osteogenic lineage and is crucial for Ewing sarcoma bone invasion and osteolysis in vitro and in vivo.

Ewing sarcoma are osteolytic bone tumors, characterized by early metastasis into lungs and bones (2). In the present study, we identified DKK2 as a highly overexpressed protein in Ewing sarcoma that seems critically involved in the malignant behavior of this disease, presumably, in part, via the activation of Wnt/β-catenin signaling (Supplementary Fig. S6). Our data support a role of DKK2 overexpression for the maintenance of Ewing sarcoma osteolysis, bone invasiveness, and metastatic spread, associated with a reprograming from neuronal to chondro-osseous differentiation. Furthermore, we showed that DKK2 is also important for anchorage-independent colony formation and proliferation of Ewing sarcoma cells in vitro and for tumorigenicity in vivo. High expression of DKK2 in Ewing sarcoma raises the question whether DKK2 expression is dependent on the oncogenic fusion protein EWS-FLI1, as reported by Miyagawa and colleagues (31). However, in line with the findings of Navarro and colleagues (32), we found that the expression of DKK2 is mostly independent of EWS-FLI1.

Our functional analyses suggest that in Ewing sarcoma, MMP1 but not MMP7 is responsible for the reduced invasiveness after DKK2 knockdown in vitro. MMP1 as well as MMP7 are known to be involved in bone invasion and bone metastasis (24–26), although this finding does not preclude other DKK2 driven factors to be relevant for invasiveness and metastasis. One possible alternative candidate is E-cadherin, a key cell–cell adhesion molecule, which is upregulated after DKK2 knockdown and which has a strong anti-invasive and antimetastatic role in tumor progression (20). In Ewing sarcoma, it mainly seems to suppress anoikis through activation of ERBB4; however, this gene is not enhanced after DKK2-mediated CDH1 increase (data not shown; ref. 33). Other genes associated with invasion and metastasis are suppressed after DKK2 knockdown such as CD44—an integral cell membrane glycoprotein. As CD44 promotes cell proliferation, invasion, and migration and further inhibits apoptosis, its downregulation after DKK2 knockdown may, in part, explain the increase of cell death observed in our study (21, 22, 34). In addition, invasion in Ewing sarcoma cells may also be promoted by the intercellular adhesion molecule 1 (ICAM1), which is important for the extravasation of circulating tumor cells during metastasis (23). Of note, 4 of our 5 validated DKK2 target genes are associated with bone colonization, bone invasion, and osseous differentiation, which is interesting in the pathophysiologic context of Ewing sarcoma that frequently arise in and metastasize into bones.

In recent years, some genes were identified that enabled bone metastasis of different tumor entities. Among them are secreted phosphoprotein 1 (OPN; ref. 35), a potent ligand of CD44 (22), cbp/p300-interacting transactivator, with glu/asp-rich c-terminal domain, 2 (CITED2; ref. 36), cAMP response element-binding protein 1 (CREB1; ref. 37), and secreted protein, acidic, cysteine-rich (SPARC; ref. 38). Interestingly, CD44 and OPN as well as the target genes of CITED2 (MMPs, TGFβ1, PTHrP), CREB1 (MMPs, PTHrP), and SPARC (VEGF) are also regulated by DKK2, as shown in this study. Moreover, we showed that DKK2 modulates the expression of additional genes important for osteolysis (CXCR4, HIF1α, JAG1, IL6, PTHrP, and VEGF). Furthermore, high CXCR4 expression seems associated with poor prognosis in patients with Ewing sarcoma (39). The transcription factor RUNX2 affects cancer cell invasion and osteolysis (26) and can bind to the oncogenic fusion protein EWS-FLI1 (40). This interaction is potentially fostered by a DKK2-dependent RUNX2 induction and thus promotes the differentiation of Ewing sarcoma cells (40). This could explain the increased neuronal differentiation of Ewing sarcoma cells after DKK2 knockdown showing that not only EWS-FLI1–driven histone methyltransferase EZH2 influences the differentiation potential in Ewing sarcoma (17). However, their chondrogenic and osteogenic differentiation capacity is reduced after DKK2 knockdown, which may be again mediated, at least in part, by RUNX2, as this protein promotes osteogenic and chondrogenic differentiation (26, 40, 41). Furthermore, we could show for the first time that reduced osteolysis after DKK2 knockdown in vivo is probably mediated by decreased expression of osteolytic genes, which function either directly via the stimulation of osteoclastic bone resorption (PTHrP, IL6, VEGF) or indirectly (HIF1α, JAG1) by increasing the ratio of receptor activator of NF-κB ligand (RANKL) to osteoprotegerin (OPG) to activate the maturation of osteoclast precursors (26, 42). These data in addition might be of clinical relevance as pilot data obtained from human Ewing sarcoma immunohistology indicate the level of DKK2 expression to be associated with tumor invasiveness and survival.

In synopsis, the current study provides evidence that DKK2 is critically involved in Ewing sarcoma pathology, especially in bone-associated tumor growth and metastasis. With DKK2, we could identify a first element for the mechanisms of bone invasiveness and osteolysis in Ewing sarcoma. DKK family members seem likely to be involved in other bone-infiltrating tumors including osteosarcoma as well as breast or prostate cancer (43–45), supporting a critical role of Wnt/β-catenin signaling for bone metastasis. Hence, DKK2 may constitute a promising new drug target in different bone cancers and bone metastasis that can possibly be used to prevent or at least to delay metastasis.

No potential conflicts of interest were disclosed.

Conception and design: S. Burdach, G.H.S. Richter

Development of methodology: K. Hauer, I. Esposito

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K. Hauer, J. Calzada-Wack, T.G.P. Grunewald, D. Baumhoer, S. Plehm

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): K. Hauer, J. Calzada-Wack, K. Steiger, T.G.P. Grunewald, D. Baumhoer, S. Plehm, T. Buch, O.P. da Costa, I. Esposito, S. Burdach, G.H.S. Richter

Writing, review, and/or revision of the manuscript: K. Hauer, J. Calzada-Wack, T.G.P. Grunewald, D. Baumhoer, O.P. da Costa, I. Esposito, S. Burdach, G.H.S. Richter

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): T. Buch, O.P. da Costa

Study supervision: T. Buch, S. Burdach, G.H.S. Richter

The authors thank Colette Berns for expert technical assistance and Uwe Thiel for support and critically reading of the manuscript. Immunodeficient Rag2^−/−γc^−/− mice were kindly provided by Mamoru Ito (Central Institute for Experimental Animals).

This work was supported by grants from Technische Universität München (KKF_B08–05 to T.G.P. Grunewald and A09–02 to G.H.S. Richter); the Wilhelm-Sander-Stiftung (2009.901.1 to G.H.S. Richter and S. Burdach); is part of the Translational Sarcoma Research Network supported by the BMBF (FK 01GM0870 and 01GM1104B to G.H.S. Richter and S. Burdach); and Deutsche Forschungsgemeinschaft (DFG_GR3728/1–1 to T.G.P. Grunewald, G.H.S. Richter, and S. Burdach; DFG_GR3728/2–1 to T.G.P. Grunewald). Work in I. Esposito laboratory is supported by the National Genome Research Network (NGFNplus, 01GS0850) and in the laboratory of T. Buch by the Deutsche Forschungsgemeinschaft (SFB-TR22).

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.