ECM, extracellular matrix; PCR, polymerase chain reaction; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; RT, reverse transcriptase.
Osteosarcoma is the most frequent malignant bone tumor with a poor survival rate for patients with metastasis. Previous studies have shown that beside other proteases, distinct sets of cathepsins are involved in the process of metastasis of different tumors. In this study we investigated the expression of cathepsin proteases in human osteosarcoma metastasis. First, the mRNA expression of 14 human cathepsins was studied in SAOS-2 osteosarcoma cells and the highly metastatic LM5 and LM7 sublines by reverse transcriptase (RT)-polymerase chain reaction (PCR). The expression of cathepsin D, K, and L mRNA was found upregulated and that of cathepsin F, H, and V downregulated in the highly metastatic LM5 and LM7 cells. A subgroup of the cathepsin proteases was further studied at the protein level by Western blot analysis of cell extracts. The expression of cathepsin B and H was decreased and that of cathepsin D, K, and L was increased in the highly metastatic cell lines as compared to the SAOS-2 cell line. Diagnostic relevance of cathepsin K expression in osteosarcoma was revealed upon correlation of survival and metastasis with immunohistochemical cathepsin K staining of biopsies collected from 92 patients prior to chemotherapy. Patients with metastatic high-grade osteosarcoma and low cathepsin K expression at diagnosis had a better prognosis than those with high expression. Thus, it appears that cathepsin K expression is of predictive prognostic value for patients with high-grade tumors and metastasis at diagnosis. © 2007 Wiley-Liss, Inc.
Osteosarcoma is the most frequent primary malignant tumor of bone typically affecting children and young adults. The overall prognosis for patients with localized osteosarcoma has improved significantly (about 65% survive 5 yr) because of progress in chemotherapy and refined surgical techniques. In contrast, for 10–20% of the patients who present with metastasis, with the lung being the most common site, the 5-yr survival rate is 20%, which plateaued over the last 15 yr 1, 2. Therefore, new treatment strategies are required to prevent metastasis to improve the survival of these patients.
The development of an invasive tumor with metastatic potential is a complex, multistep process that includes cell proliferation, angiogenesis, tissue remodeling, and invasion 3. The degradation of the surrounding extracellular matrix (ECM) is one of the first steps initiated by the release of proteases from different cell types, including tumor and stromal cells 4. The proteases participating in these processes belong to all known classes of proteases (metallo-, serine-, aspartyl-, and cysteine proteases). Proteolytic enzymes can act directly by degrading ECM or cell surface proteins or indirectly by activating cascades of proteolytic enzymes. Under physiological conditions these cascades are regulated in complex networks of activators, inhibitors, and receptors. During tumor progression and metastasis this protease network is unbalanced with the consequence of not only destroying the surrounding ECM, but also releasing active molecules stored in the matrix or generating active fragments of the matrix proteins, which promote cell proliferation, invasion, and angiogenesis 5, 6. In addition to matrix metalloproteinases (MMPs) and serine proteases, cathepsin proteases are also involved in the different processes of metastasis 7–9. For cathepsins, 11 cysteine proteases (cathepsins B, C, F, H, K, L, O, S, V, W, and X), 2 aspartyl proteases (cathepsins D and E), and 1 serine protease (cathepsin G) have been identified in human 10. In normal cells, they are required for protein turnover, processing of proteins in secretory granules, antigen presentation, or bone remodeling 11, 12. In addition, cathepsins have also functions outside the lysosomal compartment. They execute programmed cell death upon release into the cytosol, and when released from the cell, they degrade ECM 13, 14. Several cathepsins have been identified as markers of tumor progression and metastasis in different types of cancer 8, 15. Specifically, cathepsin B and cathepsin L were suggested to participate in metastasis of osteosarcoma tumors and are expressed in corresponding cell lines 16–19.
The present study aimed at the identification of individual cathepsins as markers for metastasis in osteosarcoma. For this analysis we took advantage of the human SAOS-2 osteosarcoma cell line with low metastatic potential and of its derivatives LM5 and LM7 with increased metastatic potential in vivo 20. Recently, this model system has been used to study the role of interleukin-12 and Fas/FasL in osteosarcoma lung metastasis in vivo 21, 22. Here, the expression of 14 cathepsins was studied at the mRNA and the protein levels in the SAOS-2 osteosarcoma cell line and the highly metastatic LM5 and LM7 sublines. The results confirmed the observations of previous reports that suggested a role of cathepsin L in metastasis of osteosarcoma. Importantly, the results also imply that cathepsin K may be of diagnostic relevance with respect to prognosis of patients with high-grade tumors and metastasis at diagnosis.
SAOS-2 cells were purchased from the American Type Culture Collection (Rockville, MD). LM5 and LM7 cells (metastatic sublines of SAOS-2 cells, 20) were kindly provided by Dr. E. Kleinerman (The University of Texas, M. D. Anderson Cancer Center, Houston, TX). The SAOS-2, LM5, LM7 cell lines were grown in Dulbecco's modified Eagle medium (DMEM) (4.5 g/L glucose) and Ham F12 medium (1:1) supplemented with 2 mM L-glutamine, 10% fetal calf serum (GIBCO, Basel, Switzerland) in a humidified atmosphere of 95% air and 5% CO2 at 37°C.
Total RNA was isolated from individual cell lines (5–10 × 106 cells) with TriReagent (Sigma–Aldrich, St. Louis, MO) as recommended by the supplier. The RNA was quantified by measuring optical density at 260 and 280 nm in a UV-spectrophotometer. The integrity of the RNA was assessed by standard agarose gel-electrophoresis.
cDNA was synthesized from 1 µg of total RNA with the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA) and random primers at 37°C for 2 h. The expression levels of individual cathepsin transcripts were analyzed by polymerase chain reaction (PCR) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) transcripts were used as a reference. PCR reactions of 50 µL final volume contained 1X PCR buffer, 0.4 µL of the reverse transcription reaction, 1.25 U Taq polymerase (Eppendorf, Hamburg, Germany), 200 µM of each dNTP, 1.5 mM Mg2+, and 0.2 µM of forward and reverse primers specific for the transcripts encoding individual genes (Table 1). The primer pairs were designed with the LightCycler Probe Design Software 2.0 (Roche Diagnostics, Rotkreuz, Switzerland). After initial denaturation at 94°C for 3 min, cDNAs representing GAPDH or individual cathepsin transcripts were amplified by 27 or 35 PCR-cycles (denaturation at 94°C for 40 s, primer pair-dependent annealing at 63–69°C for 40 s, and elongation at 72°C for 20 s), respectively. PCR was completed by a final elongation step at 72°C for 7 min. PCR products were analyzed by agarose gel electrophoresis.
|Gene||Primer sequence (5′->3′)||Product size (bp)|
|CathB F||ACA GGC CAT GTG AGC CAC CG||152|
|CathB R||CGC TTT CCA TTC CTG CGT CTC TGT CTT G|
|CathC F||ATG GGT ATG CTA GAA GCG AGA ATC CGT||174|
|CathC R||GCA AGC TTC TTC CAC CAG CCC AAA|
|CathD F||GGC CCA TTC CCG AGG TGC TCA A||162|
|CathD R||TCC AGC AAG CGA TGT CCA GCA GTT T|
|CathE F||TGG AGA ATA TGC TGT GGA GTG TGC CAA CCT TA||244|
|CathE R||ACG GTT ATT CCC ACG GTC AAA GAC TGA GTA AA|
|CathF F||AGC CCA AGT CAG CCT TCA CTC||205|
|CathF R||GCT TCT TCC TTT GAC TCA TAT GTC CGG TTA|
|CathG F||AGC CAT CCG CCA CCC TCA ATA TAA||212|
|CathG R||GCT GCA CCT CTC GGA GTG TAT CT|
|CathH F||TCA CAA TCT ATG ACG AGG AAG CGA||159|
|CathH R||GTA CTG CAT GGT TTA CTT TAT CTG GAG TT|
|CathK F||AGT GTG GTT CCT GTT GGG CTT T||175|
|CathK R||CGG TTC TTC TGC ACA TAT TGG AAG GC|
|CathL F||ATG AAT CCT ACA CTC ATC CTT GCT GCC TTT||156|
|CathL R||CCA CAC TGC TCT CCT CCA TCC TTC TT|
|CathO F||TGA GCT GGC AAG ATT ATC TGG GAG G||209|
|CathO R||AGA AAC GGA ATC TGC AAT ACC ACA AAC ATT AC|
|CathS F||CAG TAC ATC ATT GAT AAC AAG GGC ATC GAC||160|
|CathS R||CCA CAG CTT CTT TCA GGA CAT CTT CTC|
|CathV F||GCC AGT GAA GAA TCA GAA ACA GTG TG||205|
|CathV R||GCC TCC GTT CTC CTT GAC ATA C|
|CathW F||GGC GCA TCA GTT TCT GGG ATT T||211|
|CathW R||CAG GCC ACC TTC TGG TAC TTC TT|
|CathX F||TCC TGT CCG TGC AGA ACG TCA TCG||156|
|CathX R||GGT TAA ACT TGT CAC ACT CCT GGT CCT TGG|
|GAPDH F||TGA ACG GGA AGC TCA CTG GCA TGG||202|
|GAPDH R||TGG GTG TCG CTG TTG AAG TCA GAG GAG A|
Cells were detached from Petri dishes with phosphate-buffered saline (pH 7.4) (PBS)/0.05% EDTA, collected by centrifugation, and snap-frozen in liquid nitrogen. Frozen cells were extracted with 125 mM Tris/HCl, pH 6.8, 20% (v/v) glycerol, and 4.6% (w/v) SDS at 15°C for 1 h and sonication with a Microsonifier (Heat Systems-Ultrasonics, Farmingdale, NY). Cell debris was removed by centrifugation at 12 000g and 15°C for 20 min.
Proteins in cell extracts were separated by SDS–PAGE and electrotransferred to nitrocellulose Hybond™ ECL™ membranes (Amersham Biosciences UK Limited, Buckinghamshire, UK) with a Trans-Blot® SD semi-dry transfer cell (Bio-Rad Laboratories, Hercules, CA) at 15 V for 60 min. Membranes were blocked with 50 mM Tris/HCl, pH 7.4, 150 mM NaCl, 0.1% Tween 20, and 5% (w/v) low fat milk (blocking buffer) at room temperature for 2 h. Blots were incubated at 4°C overnight with antibodies to individual cathepsins or GAPDH (reference for protein loading) diluted in blocking buffer. After washing four times at room temperature for 10 min with 50 mM Tris/HCl, pH 7.4, 150 mM NaCl, 0.1% (v/v) Tween20 (washing buffer), blots were incubated at room temperature for 1 h with horseradish peroxidase-conjugated secondary antibodies (diluted 1:5000 in blocking buffer) (Jackson ImmunoResearch Laboratories, West Grove, PA) and washed four times for 10 min at room temperature with washing buffer. The peroxidase-conjugated antibodies were visualized by chemiluminescence with the Immobilon chemiluminescence substrate (Millipore, Billerica, MA) using the VersaDoc™ Imaging System (Bio-Rad). Chemiluminescence was quantified with QuantityOne 1-D analysis software (Bio-Rad) and the expression of the individual cathepsins in SAOS-2 cells was set as 100%. Polyclonal antibodies to cathepsin B, cathepsin C, cathepsin D, cathepsin F, cathepsin H, and cathepsin L were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Polyclonal antibodies to cathepsin K were obtained from Calbiochem (San Diego, CA) (for Western blots) and United States Biological (Swampscott, MA) (for immunohistochemistry), and polyclonal antibodies to cathepsin V and GAPDH were from Abcam (Cambridge, UK). Cathepsin C, F, and V remained undetectable in extracts of osteosarcoma cell lines. However, Cathepsin V was recognized as a precursor protein in control COS-7 cells that were transfected with a corresponding expression plasmid. Plasmids for similar control expression experiments with cathepsin C and F were not available to us.
For our immunohistochemical analysis, the specimens of 92 patients were included. The patients had a mean age of 24.5 (range, 2–72) yr, 40 tumors were low-grade and 52 tumors were high-grade. The clinical classification of high-grade and low-grade osteosarcoma was performed according to criteria published by Unni 23. Thirty-five patients were classified into stage 1, 39 into stage 2, and 18 into stage 3. The median survival of these patients was 88 (range, 4–400) mo (mean 131.6 mo). At the latest follow-up, 53 patients were alive and 39 died of disease. These human osteosarcoma specimens were paraffin embedded and sections of 5 µm were placed on superfrost-charged slides. Immunohistochemistry was performed essentially as described 24, 25. Briefly, the slides were deparaffinized in xylene (2 × 5 min), successively soaked in absolute ethanol and a series of ethanol/water mixtures, and then rinsed with tap water. Endogenous peroxidase activity was blocked with 0.3% H2O2 in methanol (1:1, v/v). After a tap water rinse, sections were placed into 1 mM preheated (∼90°C) EDTA (pH 8) and then rinsed in tap water and placed in PBS. Sections were incubated for 5 min with DAKO Protein block (DAKO, Cytomation, Carpenteria, CA), and then incubated with the primary anti-cathepsin K antibody for 30 min. The DAKO mouse EnVision HRP System/DAB and the DAKO Autostainer were utilized for detection. Sections were counterstained with light hematoxylin and then mounted with a coverslip. Prostate cancer tissue served as a positive control. Basic routine hematoxylin and Eosin staining (modified Schmidt's hematoxylin) was performed for all specimens to ensure tissue quality. Cathepsin K staining was graded from 0 to 4 according to its intensity and area of staining. (0 = no staining, 1 = some (<25%) islands of staining, 2 = homogenous weak staining of 25–50% of the slide, 3 = intermediate intensity and/or islands of intense staining covering 50–75% of the slide, 4 = intense homogenous staining covering >75% of the slide).
The follow-up, presence, or absence as well as timing of metastasis of patients with low- and high-grade osteosarcomas were subjected to survival analysis. Kaplan–Meier, log-rank statistics was used for all survival curves. Multivariate regression analysis was performed with the Cox proportional hazard model. P-values ≤ 0.05 were considered as statistically significant. JMP 6.0.0. software package was used for analysis (Cary, NC).
The expression of 14 human cathepsin genes was investigated at the transcript level by reverse transcriptase (RT)-PCR in low metastatic SAOS-2 osteosarcoma cells and its highly metastatic LM5 and LM7 sublines. Cathepsin B and C were uniformly expressed in all three cell lines and cathepsin C-encoding mRNA was the most abundant cathepsin gene transcript (Figure 1). Cathepsin E, G, and X-encoding mRNAs remained undetectable in all three cell lines. Transcripts for cathepsin O, S, and W were expressed at low and comparable levels (Figure 1). The expression of cathepsin F, H, and V transcripts was highest in SAOS-2 cells and downregulated in the highly metastatic LM5 and LM7 cells (Figure 1). Cathepsin D, K, and L transcripts, on the other hand, were found upregulated in LM5 and LM7 cells when compared to the parental SAOS-2 cells.
The expression of selected cathepsins at the protein level was investigated by Western blot analysis of total cell extracts. Cathepsin C, F, and V proteins remained undetectable with the available antibodies (not shown). Precursor and mature forms of cathepsin B with molecular weights of 38 and 26 kDa, respectively, and of cathepsin D, 46 and 33 kDa in size, respectively, were recognized in SAOS-2, LM5, and LM7 cell lines (Figure 2A). The differences in protein expression in the three cell lines corresponded to those observed at the mRNA level. The expression of the mature form of cathepsin B in LM5 and LM7 cells was approximately 50% of that in SAOS-2 cells (Figure 2B). The expression of cathepsin D in LM5 and LM7 cells, on the other hand, was increased 2- to 2.5-fold when compared to SAOS-2 cells. Antibodies to cathepsin H recognized two precursor forms of 42 and 38 kDa, but mature cathepsin H, reported to be 24 kDa in size, remained undetectable (Figure 2A). Consistent with the findings at the mRNA level, the expression of the cathepsin H precursor forms was lower in LM5 and LM7 cells and approximately 50% of that found in SAOS-2 cells (Figure 2B). Cathepsin L was predominantly expressed in the mature biologically active form of 32 kDa (Figure 2A) and the expression was increased in LM5 as compared to SAOS-2 and LM7 cells and comparable to the results found of the transcript analysis. Cathepsin K, similar to cathepsin B and D, was also expressed in the precursor and the mature form. The expression levels of the mature form of 27 kDa were similar in SAOS-2 and LM5 and LM7 cells and lower than those of the 40 kDa precursor form (Figure 2A). But, interestingly, the expression of the precursor protein was found to be increased twofold in LM5 and LM7 cells as compared to SAOS-2 cells. In LM7 cells the high precursor levels contrasted with the low levels of mRNA, suggesting posttranslational accumulation of the enzyme precursor. The here observed upregulation in cells with high metastatic potential, the importance of cathepsin K in normal bone turnover and giant cell tumor 26, and the fact that the expression of cathepsin K, unlike that of cathepsin B and cathepsin L, has not been investigated in osteosarcoma tumors before, prompted us to carry out an immunohistochemical analysis in human osteosarcoma tumor samples.
Cathepsin K was immunohistochemically stained in 92 osteosarcoma tissue specimens, collected prior to any chemotherapeutic treatment of the patients. Representative immunostainings are shown in Figure 3. Table 2 summarizes tumor grade, metastasis, cathepsin K immunohistochemistry, and survival. Fifteen (29%) of 52 patients with high-grade osteosarcoma and only 3 (8%) of 40 patients with low-grade osteosarcoma developed metastasis. Eight (15%) of 52 high-grade osteosarcoma specimens and 1 (2.5%) of 40 low-grade tumor specimens stained intensively for cathepsin K. Eight (53%) out of 15 specimens from high-grade osteosarcoma patients with metastasis showed high cathepsin K staining. This was also observed in one out of three specimens from patients with low-grade osteosarcoma and metastasis. Interestingly, this was a secondary dedifferentiation of a parosteal osteosarcoma. Irrespective of the grading, 50% of patients with metastasis showed high cathepsin K staining compared to 19% of those without metastasis (P < 0.04).
* Staining intensity has been defined in Material and Methods.
# Survival of ≤3 yr has been considered as “poor”; survival of > 3 yr has been considered “good”.
To determine the relationship of cathepsin K staining and overall survival, the outcome of patients with high intensity staining was compared to that of patients with low intensity cathepsin K staining. These analyses indicated that high intensity cathepsin K staining is a significant indicator of poor survival (P < 0.0003). However, histological grade may represent a confounding variable. Therefore, we reduced our survival analysis to patients with high-grade lesions. In this selected group, we were unable to find a significantly poorer outcome for patients with high intensity cathepsin K staining (P < 0.36). Because metastasis has a detrimental impact on prognosis, we next analyzed the expression of cathepsin K in this context. We found that in patients with high-grade lesions and metastasis at the time of diagnosis the survival was significantly better with low cathepsin K tumor staining than in those patients with high cathepsin K staining, which was found for univariate, (but not multivariate) analysis (P < 0.03) (Figure 3). Thus, it appears that cathepsin K expression is of predictive prognostic value for patients with high-grade tumors and metastasis at diagnosis.
Proteolytic enzymes are important factors in the diverse processes of metastasis. Modified matrix components provide the tumor cells with a large variety of signals that modulate cell proliferation, migration, or apoptosis. In osteosarcoma, cathepsin B and L have been suggested to play important roles in metastasis 16, 19. Here, the role of the other known cathepsins in osteosarcoma and metastasis was investigated. The mRNA expression profile of 14 human cathepsins was compared in SAOS-2 osteosarcoma cells and their highly metastatic LM5 and LM7 sublines. Moreover, the expression of selected cathepsins at the protein level was investigated by Western blot analysis. Immunohistochemical cathepsin K staining in osteosarcoma tumor samples revealed diagnostic relevance.
Expression of cathepsin C, F, and V in osteosarcoma tumors or cell lines has never been reported before. Here, transcripts encoding the three cathepsins were detected in all three cell lines and those of cathepsin C were most abundant. On Western blots, cathepsin C, F, and V remained undetectable, presumably because of insufficient sensitivity of the antibodies that were available to us. Otherwise, discrepant findings with regard to mRNA and protein expression have been described for a cohort of genes in lung adenocarcinomas by Chen and coworkers 27. They proposed negative regulation feedback mechanism either at the mRNA or the protein level, or other up-to-now unknown regulatory mechanisms as an explanation for their observations.
The expression of cathepsin B and H was found to be downregulated at the mRNA and protein levels in the metastatic LM5 and LM7 cell lines as compared to the parental SAOS-2 cells. Downregulation of cathepsin H is consistent with other cancers, like cervical cancer 28 and melanoma, where the expression of cathepsin H was inversely related to the invasive potential of the tumor 29. In colorectal cancer high carcinoembryonic antigen (CEA) and low cathepsin H expression correlated with poor prognosis 30. An upregulation of cathepsin H, on the other hand, was found in bladder carcinomas 31 and basal cell carcinomas 32. These differences in cathepsin H expression in different tumors imply specific functions of this protease depending on the tissue type and the specific processes during tumorigenesis and metastasis. In the osteosarcoma cell lines investigated here, cathepsin H was only recognized in its inactive precursor form, although cathepsin D, which has been postulated as an activating protease for cathepsin H 33, was abundantly expressed in all cell lines. Thus, another protease appears to be required for the activation of cathepsin H in the osteosarcoma cell lines studied here. Cathepsin B was downregulated in LM5 and LM7 osteosarcoma cells with increased metastatic potential. Conversely cathepsin B was found to be expressed or even upregulated in different tumors such as melanoma 29, 34, cervical cancer 28, mammary cancer 35, and dedifferentiated parosteal osteosarcoma 16. The observed differences in cathepsin B expression in osteosarcoma compared to other tumors prompted us to hypothesize that this protease is mainly involved in the extravasation of the cells from the primary tumor. In osteosarcoma metastasis proteases other than those in the primary tumor may be of biological importance.
Cathepsin D and K were upregulated at the protein level in the highly metastatic LM5 and LM7 cell lines. This is in line with a previous reported role of cathepsin D as a prognostic marker for poor outcome in squamous cell carcinoma 36, oral cancer 37, or breast cancer 12. In these tumors, cathepsin D stimulated cell proliferation, growth of micro-metastasis, and angiogenesis. The functional importance of the upregulation of cathepsin D in the osteosarcoma cells with high metastatic potential remains to be further investigated. Cathepsin L expression was demonstrated in metastatic bone tumors 19 and its suppression in human MNNG/HOS osteosarcoma cells by antisense oligonucleotides inhibited motility and invasion but not adhesion in vitro 18. This protease was also proposed as most potent in degrading structural proteins of the basement membrane including collagen and laminin 38. Moreover, cathepsin L has been shown to be involved in the activation of other proteases that are important in metastatic processes. One could argue that the expression level does not always correlate with the functional relevance and vice versa. Therefore, the increase of cathepsin L expression, even at a low level as shown in here, can lead to an amplification of the protein degradation process necessary for metastasis through the activation of other enzymes like uPA, MMPs, or other cathepsins.
Cathepsin K was found upregulated in the highly metastatic cell lines LM5 and LM7 both on mRNA and protein levels. In addition, based on our immunohistochemistry analysis cathepsin K appears to be an indicator for metastasis in osteosarcoma and, as a consequence, for poor outcome. Interestingly, high cathepsin K staining correlated with poor prognosis for osteosarcoma patients with metastasis at the time of diagnosis, suggesting a function of cathepsin K in one of the initial processes of metastasis such as degradation of collagen I at the site of the primary tumor in the bone. Previously, cathepsin K expression has been identified mainly in osteoclasts but also in osteoblasts 39, 40. Moreover, cathepsin K has been found to be expressed in prostate cancer 41 or giant cell tumor of bone 26. In osteosarcoma only a few markers such as cytochrome P450 (CYP3A4/5) 42, LDL receptor-related protein 5 (LRP5) 43, and WT1 25 were postulated to predict the clinical outcome of patients. CYP3A4/5 was suggested to be involved in growth regulation through metabolism of endogenous growth regulatory factors 42, and the overexpression of the Wnt coreceptor LRP5 seems to increase proliferation in osteosarcomas 43. However, none of these markers has proteolytic activities like cathepsin K, which is known to be an important protease for the degradation of bone matrix proteins, predominantly fibrillar type I collagen. Thus, the here observed increased expression of this protease in osteosarcoma tumor tissue may contribute to the processes of osteolysis during primary tumor progression of osteosarcomas and to bone metastasis.
In conclusion, the biological relevance of cathepsin B, D, H, and L in osteosarcoma remains to be established. Cathepsin K appears to be a useful prognostic marker for patients with high-grade osteosarcoma and metastasis.
We thank Bettina Langsam for excellent technical assistance. This study was supported in part by the Zurcher Krebsliga (Zurich, Switzerland), the WL & J Wolf Foundation (Zurich, Switzerland), the University of Zurich, and the Schweizerischer Verein Balgrist.